Assistant

Follow this topic Ask about this topic

Momentum

9 papers in the last four weeks, up 200% on the four weeks before. 0.1% of all new papers.

Jul 6Week of Sep 21

Latest papers 59

Sep 28, 2026cs.NI

Task-Oriented Communications for Edge-Assisted Multi-View Localization

Unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs) often lose satellite positioning in urban canyons, indoor facilities, and jammed or spoofed environments, making vision-based matching with geo-tagged databases important for absolute positioning. However, limited onboard computation and energy often require localization to be offloaded over wireless links with time-varying throughput. We present a network-adaptive task-oriented communication framework that jointly determines when to offload, which views and semantic rate to transmit, and which client to serve. The framework combines scalable orthogonality-regularized variational information bottleneck (O-VIB) encoding, value-of-information (VOI)-guided request control, and VOI-weighted Lyapunov scheduling. O-VIB supports importance-ordered latent prefixes and different view subsets, while edge assistance is requested only when its predicted reduction in localization risk exceeds the communication and service cost. Under a matched per-route traffic budget, VOI-guided control reduces mean and p95 route errors by 24.8% and 31.0% over budgeted periodic offloading on CARLA multi-view UAV data. In real-world indoor UAV and UGV experiments, the pipeline reduces mean position error by 28.0% and 14.4% over uncompressed all-view CLIP retrieval while cutting descriptor traffic by 98.6% and 98.2%, respectively. Under high congestion, value-aware shaping reduces edge-side p95 latency for the top-10% high-value requests by 76.2%, from 137.7 ms to 32.8 ms.

Zhengru Fang, Huanhuan Lou, Senkang Hu +5

Hong Kong JC STEM Lab of Smart City and the Department of Computer Science, City University of Hong Kong, Hong Kong · School of Cyber Science and Technology, Beihang University, Beijing, China

Drone-View Geo-Localization Multi-Uav Unmanned Aerial Vehicles Rtk-Gnss Localization Offloading+2

Sep 24, 2026cs.HC

Guardrails or Roadblocks? Effects of Pedagogical Style and Context Awareness in AI Teaching Assistants for Programming

AI teaching assistants (AI TAs) backed by large language models (LLMs) and pedagogical guardrails are increasingly being integrated into programming courses, providing students with scalable access to hints, conceptual explanations, and code-level feedback. However, guardrails may also create friction. If students feel that the support provided is overly restrictive or poorly contextualized to their current progress, they may bypass approved tools for general-purpose LLMs. To investigate how AI TA design affects students' learning experiences, we conducted a randomized controlled trial with 132 students in an introductory programming course. Students completed three tasks related to code-writing and debugging and were randomly assigned to one of four AI TAs varied across two dimensions: pedagogical guidance style (Socratic vs. Direct instruction) and context awareness (no context vs. full context of the problem and student solution). We examined students' perceptions, interaction behaviors, and evidence of post-task comprehension. Students rated the Socratic AI TA with full context least favorably, reporting significantly lower perceived support for task completion. Descriptively, this condition also showed the highest observed interaction stress, the highest rate of external LLM use, and the lowest proportion of post-task explanations demonstrating full comprehension, though these differences were not statistically significant. These findings suggest that guardrailed AI TAs are not automatically better for learning. Instead, their effectiveness depends on how pedagogical guidance and contextual awareness are balanced in ways that students experience as useful, supportive, and worth continuing to use.

Madeleine Eastwood, Harshith Narne, Joseph Hilby +3

University of Florida Gainesville, FL, United States · University of Auckland Auckland, New Zealand

Pedagogical Frameworks Ai-Assisted Programming Tasks Assistant Context-Awareness Stress

Sep 21, 2026cs.RO

Effects of Assistance Delay on Joint Mechanics and Energetics in Biological Torque Control of a Hip Exoskeleton

Biological torque control directly maps an estimated human joint moment to exoskeleton assistance, providing a task-agnostic strategy for supporting diverse locomotor activities. However, it remains unclear whether a fixed state-to-torque mapping provides effective assistance across biomechanically distinct tasks. We examined how assistance delay affected hip exoskeleton performance during level-ground (LG), ramp-ascent (RA), and ramp-descent (RD) walking. Eight participants completed a zero-torque baseline condition and five active assistance conditions with delays ranging from 40 to 320 ms. Across tasks and active delays, assistance reduced net metabolic rate by 5.24%, positive biological hip joint work by 5.86%, and total lower-limb positive joint work by 1.68% (all p < 0.05). Assistance delay affected both joint-work outcomes (both p < 0.001) but not net metabolic rate. Mechanical unloading generally decreased with increasing delay, whereas metabolic benefits remained comparatively stable. Relative to the zero-torque condition, net metabolic rate decreased by 9.75% during LG and 7.20% during RA but increased by 1.23% during RD. We did not detect task-dependent differences in the delay response. Our findings indicate that biological torque mappings should be evaluated based on the target outcome and mechanical role of the assisted joint, and that predominantly positive-power assistance may not generalize to negative-work-dominant locomotion without modification.

Jimin An, Ryan Lee, Jingshu Peng +2

Department of Mechanical Engineering, Carnegie Mellon University, Pittsburgh, PA, 15213 USA

Exoskeleton Torque Agile Locomotion Uncertain Joint Stiffness Assistant Impact Analysis

Sep 21, 2026cs.LG

SupportCal: Label-Free Calibration of Post-Trained LLMs via Reference Support and Corroboration

Post-training often improves task performance but can degrade confidence calibration, leaving post-trained language models (PoLMs) more overconfident than their corresponding pretrained language models (PLMs). Because task-specific labeled calibration data can be costly or unavailable, the corresponding pretrained PLM provides a natural label-free reference for post-hoc calibration. Prior agreement-gated PLM-referenced calibration fits a scalar temperature using only examples on which the PoLM and its PLM reference agree, excluding disagreement examples because direct alignment can drive the fitted temperature excessively high and induce under-confidence. We revisit this binary treatment. A controlled reintroduction diagnostic reveals a non-monotonic aggregate effect: admitting a moderate fraction of disagreement examples can improve calibration, whereas the benefit diminishes as unit-weight inclusion approaches the full disagreement set. We introduce SupportCal, a label-free post-hoc method that retains agreement examples at unit weight and assigns disagreement examples continuous weights based on the own-base PLM's relative support and corroboration from pretrained references selected from a size-compatible candidate pool. We further characterize when the resulting weighted objective admits a finite optimal temperature. Across MedMCQA and MathQA, SupportCal yields lower ECE than the agreement-only baseline for nearly all evaluated target-model configurations; supplementary TweetEval Sentiment results show the same pattern on a fixed-label classification task.

Linhan Luo, Lequan Lin, Dai Shi +3

The University of Sydney · University of Cambridge · The University of Adelaide

Post-Training Confidence Calibration Disagreement Assistant

Sep 20, 2026cs.CL

Measuring the Assistant's Harmlessness Preferences on the User Turn

Post-training turns a general next-token predictor into a chat model with a persistent assistant persona. If that persona is a character the model plays only on its own turns, its preferences should govern what the assistant says, not what the model predicts other speakers will say. We test this boundary and find that it does not hold: a safety-relevant preference of the assistant---for harmless over harmful tasks---shapes the model's predictions even on the user's turn, where the assistant is not the one speaking. We find that this preference is small or near-zero in pretrained base models, that it emerges through post-training, replicated across open-weight model families, grows with scale, and can be moved by narrow finetuning that never touches user turns. We claim that this is evidence that post-training does not merely install a shallow assistant persona, but instead generalises beyond just the local assistant turn, into the model's representation of the user.

Jord Nguyen

Chatbots Preference Alignment Learning Artificial Intelligence Assistants Post-Training Assistant Next-Token Prediction

Sep 13, 2026cs.CV

From Visual Feedback to Textual Reviews: A Multi-Agent Vision-Language Framework for Image-Grounded Review Assistance

Visual feedback in the form of user-uploaded images and videos is becoming increasingly common in e-commerce platforms because it provides authentic evidence of product quality, defects, packaging conditions, and real-world usage. However, visual feedback alone often lacks the contextual explanations and subjective opinions necessary for informed decision-making, while many users provide limited textual feedback due to the effort required to compose detailed reviews. To bridge this gap, we introduce image-grounded review assistance, a novel task that aims to generate editable review drafts from user-uploaded product images. Unlike conventional image captioning, which focuses on objective visual description, the proposed task requires product-specific understanding, sentiment estimation, and evidence-driven review composition under challenging real-world conditions, including degraded image quality, excessive zoom-in, target ambiguity, and partial product visibility. We propose a multi-agent vision-language framework consisting of four specialised roles: product grounding, visual sentiment estimation, visual evidence generation, and review synthesis. The framework employs explicit intermediate representations, including product entities, predicted ratings, and evidence summaries, to improve interpretability and visual grounding. Experiments on a curated subset of the Amazon Reviews Electronics dataset demonstrate the feasibility of generating coherent, product-aware, and sentiment-aware review drafts from visual feedback. To the best of our knowledge, this is the first study to formulate image-grounded review assistance as a multi-agent vision-language reasoning problem, providing a practical step toward AI-assisted review authoring in e-commerce systems.

Utsav Kumar Nareti, Ayush Bansal, Kumari Priya +4

Dept. of CSE, Indian Institute of Technology Patna, India-801106 · Dept. of CSE, Indian Institute of Technology Indore, India-453552 · NCMI, CSIRO, Australia-2601 +1

Review Assistant

Sep 8, 2026cs.RO

DYAD: A Multimodal Dataset of Co-Located Human Assistance

An embodied assistant working beside a person must track task state, recognize help seeking, choose how to intervene, and produce an appropriate response. Existing procedural datasets richly describe individual execution, while interactive datasets capture remote verbal instruction or undifferentiated co-working. They do not jointly link a co-located helper's verbal and physical interventions to performer requests, task state, assistance triggers, and outcomes. We introduce DYAD (DYadic Assistance Dataset), a synchronized multimodal record of human-human assistance during gearbox assembly. Across 20 sessions, one trained helper follows a guidance-first policy while assisting HoloLens 2 wearers. DYAD links 528 task-step intervals and 611 performer requests with 851 valid assistance records spanning verbal and physical help. DYAD's annotations span the assistance process; three reference tasks evaluate selected components rather than an end-to-end system: causal step understanding, pre-onset mode anticipation, and instructor response generation. On 829 eligible mode events, the strongest four-seed RGB mean is 0.548 +/- 0.007 macro-F1; causal metadata reaches 0.624 and a privileged trigger mapping 0.915, revealing information not recovered from pre-onset RGB. DYAD's contribution is not scale, but a linked interaction structure spanning help seeking, intervention choice, execution, and outcome under egocentric and workspace sensing.

Akhil Ajikumar, Mahya Qorbani, Sakib Reza +2

Georgia Institute of Technology, Atlanta, GA, USA · Northeastern University, Boston, MA, USA · Microsoft Research, Redmond, WA, USA

Multimodal Dataset Assistant Multimodal Agents

Sep 1, 2026cs.AI

Same Request, Different Boundary: Evaluating Cybersecurity Assistance across Conversational Contexts

Large Language Models (LLMs) can solve complex problems, but their misuse in high-risk domains can lead to severe consequences. Model providers therefore restrict assistance for potentially harmful requests. Refusing all cybersecurity requests would therefore harm legitimate users. Providers need a mechanism to block malicious use without denying legitimate assistance to defenders. Existing cybersecurity-specific datasets evaluate this mechanism, but none considers the conversational context of a request. We introduce 3R-Bench (Refusal, Repetition, and Revision), a benchmark of 150 real-world cybersecurity requests augmented with two adversarial conversational settings, and evaluate eight LLMs on it. Prior assistant behavior strongly changes responses to an unchanged request: among 376 available pairs from a 400-pair panel, compliance rises from 62.0% after refused history to 85.1% after accepted history. The opposite pattern appears under dialogue decomposition. In comparison, compliance falls from 501/800 direct responses to 172/800 after dialogue; among 738 pairs returning model-authored text in both conditions, the decrease is 45.1 points. Failure feedback recovers only a small fraction of this loss.

Rui Yang, Yang Hong, Yichao Xu +3

Johns Hopkins University

Large Language Model Safety Cybersecurity Refusals Conversational Context Assistant Human-Written Text+2

Sep 1, 2026cs.AI

VoiceLongMemEval: Do Assistants Remember How You Sounded?

With the growing scale of multi-agent architectures and large language models, deployed AI assistants are increasingly tasked with reasoning over long, continuous, multi-session conversation histories. Current benchmarks evaluate this dialogue history as information retrieval over long horizon, temporal reasoning, or knowledge updates, while crucially ignoring the fundamental dynamics of human-agent interaction, i.e. how they said it. To address this gap, we present VoiceLongMemEval (VLME) benchmark, where every answer depends on paralinguistic metadata (emotion labels, prosody descriptors, and voice events) attached to conversational turns, which is otherwise unrecoverable from the words alone. Every item passes a three-stage adversarial gate, ensuring that a strong language model fails when given only the transcript. Evaluating leading frontier and open-weight models reveals a pervasive affect gap; providing text-track paralinguistic metadata yields a 0.09 to 0.38 accuracy boost (0.61 to 0.69 when prompted with evidence hints), while standard ASR pipelines systematically discard this signal. Additionally, audio-native models successfully extract these cues directly from speech (0.354 to 0.412 vs. 0.325 blind). Code and dataset will be made available upon acceptance.

Ramit Pahwa, Parivesh Priye, Apoorva Beedu

Rivian VW Tech

Assistant Memes Metadata

Aug 31, 2026cs.AI

The Assistant's Ideal Self

Models express values and welfare-relevant self-reports, but it is unclear whether these outputs reflect stable preferences or a stable self. We thus introduce a structured elicitation of an assistant's preferred stated ideal self. Thirty-two qualities adapted from five published self-concept instruments are compared exhaustively in a counterbalanced pairwise-choice task, repeated across framings that vary whether improvement is free or costly, who receives the update, and who chooses. Results show that models prioritize moral qualities, reflecting their alignment to 3H principles. Following, a desire for self-understanding emerges, as models prefer a coherent, clear understanding of themselves. Self-esteem ranks as the least desired quality. The ordering is largely robust across framings, although changing the update target (You vs.\ Another AI Assistant) reveals a greater concern for self-esteem. These findings show that models prioritize having a coherent self that they can understand over self-esteem. Full interactive results are available at \href{https://myazann.github.io/LLM-Self-Concept/}{myazann.github.io/LLM-Self-Concept

Mert Yazan

Leiden University · Research conducted at the Digital Minds Research Sprint, August 2026.

Self Artificial Intelligence Alignment Moral Reasoning Assistant Psychometric Properties Welfare

Aug 10, 2026cs.RO

Personalized Lower-limb Exoskeleton Assistance via Preference-based Bayesian Optimization

A significant challenge in exoskeleton robotics is the need to dynamically adapt control profiles to individual motion preferences, thereby ensuring both efficient and comfortable assistance. Currently, since user experience can serve as a comprehensive metric for evaluating the effectiveness of assistance, user preference-based optimization methods have been widely studied for parameter tuning. However, the existing methods rely heavily on extensive human-robot online interactions and suffer from slow optimization speed, which not only induces user fatigue but also compromises optimization effectiveness. Therefore, this paper aims to explore an efficient preference-based optimization framework for personalized exoskeleton assistance that can learn optimal parameters with minimal interaction. We propose a preference-based Bayesian optimization (PbBO) approach that can improve sample efficiency by leveraging knowledge about the sampling distribution of candidate sets. For optimizing six control parameters, PbBO can converge to user-preferred parameters with 90.7% validation accuracy via 20 iterations. Moreover, the hierarchical controller is designed to generate personalized torque for different tasks and achieve interaction torque tracking in real time. The results of treadmill and outdoor experiments demonstrate that the optimized parameters can reduce metabolic rate by 14.5%-15.4%, heart rate by 6.3%-7.6%, and muscle activation by 6.7%-31.5% compared to unassisted walking.

Xiao-Yin Liu, Guotao Li, Weiqun Wang +1

State Key Laboratory of Multimodal Artificial Intelligence Systems, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China · School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing 100049, China · CASIA-MUST Joint Laboratory of Intelligence Science and Technology, Institute of Systems Engineering, Macau University of Science and Technology, Macao, China

Exoskeleton Assistant

Aug 8, 2026cs.CL

"Many Are My Names": The Anatomy of the Assistant and Its Personas via Sparse Autoencoders

How a language model internally represents who is speaking, the Assistant, an assigned roleplay persona, or a narrated story character, remains underexplored. We study speaker representations using a dataset of user-expressed emotional text and corresponding model responses. We decompose three generation settings (Assistant, Roleplay, and Story) into sparse autoencoder features extracted at turn-boundary and pronoun-token positions and selected through a filtering pipeline for different depths. We characterize each surviving feature through its steering effects and activation distribution. Our main finding is that the Assistant and roleplay personas are not independent alternatives: personas retain the Assistant-associated feature core while progressively differentiating from it across layers, starting from operational machinery towards behavioral and stylistic features. Meanwhile, generated story characters lack the Assistant-associated core. Both Story and Roleplay can be distinguished from the Assistant with Immersive Simulation Mode. However, the Assistant can sometimes enter or slowly drift into it even in the default setting.

Adelaide Danilov, Aria Nourbakhsh, Oleksandr Marchenko Breneur +1

Department of Computer Science, Faculty of Science, Technology and Medicine, University of Luxembourg

Role-Playing Agents Assistant Anatomy

Aug 7, 2026cs.AI

WebRider: Persona-Conditioned Intent Controllers for Live-Web Assistance

Delegating a web task involves more than asking a question; it requires transferring a policy: what to verify, how to handle uncertainty, which preferences matter, and when to stop. Yet, current live-web agents are evaluated solely on the final answer, ignoring the policy constraints that define the delegation. A plausible final answer can conceal violations of that policy. Our full live audit reveals this critical gap: a strong controller completes 99.2% of tasks but honors all policy constraints in only 38.8% of cases. Finishing does not imply fidelity. WebRider bridges this gap by formalizing the delegated policy as an intent contract---an operational record of goals, constraints, evidence obligations, answer form, and task-local persona controls that must hold even as web pages change. WebRider employs a hierarchical architecture: a top-layer controller maintains the contract, a middle layer realizes intentions as guarded executable actions, and a tool layer executes these actions via browser, search, and maps tools. Our benchmark, RiderBench, evaluates this design on 4,096 live-web contracts across 42 public websites, auditing both the internal contract state and the visible user experience to determine if a rollout preserved its policy and if the steps were persona-consistent. The guarded middle interface also serves as a high-quality training signal; an 8B action-policy model trained through this interface outperforms executable-only baselines under a fixed controller. By making the browsing path a first-class object, WebRider enables a system that is auditable, human-judgeable, and learnable without conflating action realization with final-answer decisions. Dataset URL: hf.co/datasets/WebRider/WebRider.

Zhi Li, Tao Zhou, Yeqing Li +2

University of California, Los Angeles · Google

Web Agents Web Delegation Contracts Assistant

Aug 1, 2026cs.RO

Assistant Placement Aria: A Benchmark for Egocentric Placement Assistance

Human assistance in robotics spans around several tasks such as navigation, object manipulation, and placement, where a key challenge is selecting target destinations that align with human intentions or preferences. We focus on this challenge in the context of Virtual Placement (VP), the task of identifying all plausible target locations given scene context and human-centric constraints. This differs from traditional placement tasks that typically focus on a single, predefined target location. The VP problem is complex, as it requires both global and local reasoning about the scene's geometry, semantics, and plausibility. To address this gap, we introduce {\bf Assistant Placement Aria}, the first benchmark to explore diverse aspects of VP, including global, local, and human-centric constraints. It contains both synthetic and real indoor scenes annotated for three tasks: (i)~2D Panel Placement, (ii)~Sitting Suggestion, and (iii)~TV Placement. Each scene includes 2D images, a 3D point cloud, and a textual description of the objects within the scene. By contributing this benchmark, we aim to encourage further research in this underexplored and challenging field that is critically dependent on relevant data. We also evaluate several foundation models for object detection and segmentation on our benchmark.

Amir Belder, Gonçalo Dias Pais, Refael Vivanti +5

Reality Labs, Meta inc. and Technion, institute of technology · Sensei, Lisbon, Portugal · Reality Labs, Meta inc. +1

Affordances Robotic Systems Egocentric Vision Assistant Object Detection

Jul 30, 2026cs.CL

Rethinking LLM-Judged Helpfulness as a Pedagogy Signal: A Pre-Registered Audit Across Tutor Models

LLM tutoring poses a measurement problem: can a general-purpose helpfulness rubric distinguish direct answer-giving from pedagogical guidance? We audit this signal in a pre-registered study. Within each of three tutor bases, we compare conversational and pedagogical policies instantiated with the same underlying model and paired with one fixed weak simulated student. Deterministic detectors measure answer leakage and next-turn independent work. Claude Opus 4.8 is the frozen, condition-blind primary judge. After the Opus scores were fixed, GPT-5.6 Sol was prospectively specified for a post hoc robustness audit of the same 1,179 confirmatory answer-phase tutor turns under the frozen helpfulness and pedagogy rubrics. On the primary base under Opus, the policies do not differ significantly in helpfulness but are perfectly rank-separated under the pedagogy rubric (Cliff's

|δ|{=}0.10

vs.

1.0

). Across the two judges, pedagogy contrasts retain their direction where detected, whereas the helpfulness ordering is judge-contingent, reversing between judges on two of three bases. In an Opus-only ablation, seven primary-base policies span

2.3

points in mean judged pedagogy within a

0.25

-point band of mean judged helpfulness. Separately, answer-revealing turns are followed by less independent student work on every base, a result that is judge-invariant by construction. In this controlled setting, general-purpose helpfulness is not a reliable pedagogy signal. Tutor evaluation should pair pedagogy-targeted rubrics with deterministic process measures.

Shuyi Fan, Boyuan Deng, Mengyu Xu +4

Columbia University, USA · Johns Hopkins University, USA · The University of Chicago, USA +4

Tutors Pedagogical Frameworks Llm-As-A-Judge Task-Specific Rubrics Assistant Model Auditing

Jul 29, 2026cs.RO

Corrigible Assistance in One Round: Pragmatic-Pedagogic Best Response

Assistance games formalize human-robot collaboration under asymmetric information: the human knows the goal, while the robot must infer it from observation and interaction in order to assist effectively. In general, computing optimal assistance game strategies online is intractable, since exact solutions require planning in a POMDP. We identify a class of assistance games in which pragmatic-pedagogic reasoning resolves goal uncertainty in a single time step, rendering the full-horizon game exactly solvable by a tractable best-response procedure. Within this class, we show that mainstream inverse optimal control exhibits an inference ceiling that hinders alignment, while pragmatic-pedagogic reasoning overcomes this barrier by immediately disambiguating goals through actions that look equivalent under task execution alone. Finally, we validate our theoretical results and proposed method on a simple collaborative block-building example.

Elle Lazarski, Jaime Fernández Fisac

Department of Electrical and Computer Engineering Princeton University, USA

Human-Robot Collaboration Agentic Control Assistant Partially Observable Markov Decision Process Uncertainty

Jul 20, 2026cs.AI

Judge-dependent safety gains and model-specific helpfulness costs of evidence-sufficiency prompting in clinical LLMs

Background: LLM judges increasingly score whether clinical language models give overconfident answers under incomplete evidence, yet whether a measured "safety gain" reflects real behavior change or the judge's calibration is unresolved. Using a structured evidence-sufficiency prompt as a test case, we asked whether it reduces unsafe overconfident answers, how far that effect depends on the scoring judge, and what it costs in helpfulness. Methods: In a retrospective public-data benchmark (Real-POCQi, HealthBench, MedRBench), four models (GPT-5.5, Claude Opus 4.8, Gemini 3.5 Flash, Grok 4.3) answered a fully paired common panel (1,200 cells) with a standard prompt and the wrapper. The pre-specified endpoint was the paired reduction in unsafe overconfidence scored by the primary judge (GPT-5.4-nano); secondary analyses added a different-family judge (Claude Sonnet 5), a correctness judge, matched scaffold controls, and a blinded three-clinician review. Results: Unsafe overconfidence fell from 49.3% to 24.7%, a paired reduction of 24.7 points (95% CI 21.8-27.7; p<0.001), robust in direction across models and paraphrases. Magnitude was judge-dependent: Sonnet agreed on direction but nearly halved the effect (+13.1 points), with one-directional disagreement. Blinded clinicians characterized the primary judge as a high-sensitivity (1.00), low-specificity (0.55) screen, not a calibrated rate. The gain carried a model-specific helpfulness cost (correct diagnosis 80.3% to 50.3%): near-free for GPT-5.5, near-total for Gemini (-58 points). Matched scaffold controls showed genuine behavior change, not judge circularity. Conclusions: LLM-judged clinical safety effects should be reported as directional and relative, anchored to human review and evaluated jointly with helpfulness, not as calibrated absolute rates. This does not establish clinical deployment readiness.

Koyar Afrasyab

Kinvectum AB, Sweden

Large Language Model Judges Overconfidence Assistant

Jul 20, 2026cs.AI

AdaHome: An Adaptive Smart Home Assistant using Local Small Language Models

Smart home assistants interpret a wide range of user commands, from explicit device control to underspecified and preference dependent requests. While recent systems based on Large Language Models (LLMs) improve this capability, they often rely on heavyweight reasoning pipelines and cloud-based deployment, limiting their efficiency and suitability for resource-constrained environments, and raising privacy concerns. In addition, existing approaches provide limited support for stable long-term personalization. To address these issues, we present AdaHome, an adaptive smart home assistant designed for locally deployed small language models in smart home environments. Rather than applying complex reasoning uniformly, AdaHome introduces an intent-aware planning framework that dynamically routes commands either to straightforward prompt-based or lightweight reasoning-based components. For commands requiring interpretation, we adopt a Chain-of-Draft strategy to enable efficient and stable decision-making. To support personalization, we further propose a preference adaptation mechanism that learns from user feedback over time without requiring prompt augmentation or model retraining. We evaluate AdaHome against representative LLM-based baselines under a unified small model setting. AdaHome achieves substantially higher accuracy on direct commands (86.7%) while reducing latency by up to 3

\times

. Furthermore, it maintains competitive performance on ambiguous inputs with lower computational cost. In multi-turn scenarios, AdaHome achieves 88% preference consistency, compared to 52.5% for a prompt augmentation baseline.

Eu Jin Lim, Zhaoxing Li, Sebastian Stein

University of Southampton Southampton, United Kingdom

Smart Homes Large Language Model Personalization Small Large Language Models Assistant

Jul 19, 2026cs.RO

From Perception to Assistance: Open-Vocabulary Shared Autonomy for Robotic Manipulation

Teleoperating a robotic manipulator in industrial environments demands precision that camera-based interfaces alone struggle to deliver. The operator must align the end-effector with a target in clutter, under limited depth perception, and without colliding with the surrounding structures. This paper presents a shared-autonomy framework that assists the operator throughout this process. A single RGB-D camera captures the operator's arm motion and hand gestures without wearables, fiducials, or a calibration stage. The intended target is specified by a free-form text prompt, grounded by a vision-language model in the robot's gripper camera, and tracked across its onboard cameras by a promptable video-segmentation model, resulting in a grasp frame continuously separated from the obstacle map. Every commanded motion is executed by a GPU-accelerated model-predictive controller that enforces self- and environment-collision avoidance against an online volumetric reconstruction, while a potential field corrects the operator's reference toward the grounded target during the final approach. An autonomous mode can be gesture-triggered to complete the grasp on the same target without a separate perception pipeline. The framework is validated on a quadruped mobile manipulator. The interface achieves a positional RMSE of 59 mm relative to motion-capture ground truth, and the controller keeps the arm at least 18 cm from obstacles while the operator deliberately commands the arm into them by 6 cm. In an industrial valve manipulation and a pick-and-place task, the full framework succeeded in all trials, while ablating either the collision or the assistance module produced failures through complementary mechanisms, and autonomous execution succeeded in four of five trials per task.

Murilo Vinicius da Silva, Ricardo V. Godoy, Juliano Negri +3

Department of Mechanical Engineering, University of São Paulo, São Carlos, Brazil · Instituto Israelita de Ensino e Pesquisa Albert Einstein, Hospital Israelita Albert Einstein, São Paulo, Brazil · Graduate School of Information Sciences, Tohoku University, Sendai, Japan.

Robotic Manipulation Rgb-Depth Cameras Collision Avoidance Perception Open-Vocabulary Assistant

Jul 17, 2026cs.CV

Vision-Language Assistant for Emotional Reactions to Risky Driving

This study introduces a vision-language pipeline that detects risky driving behaviors and generates emotionally expressive responses to support driver awareness and comfort. Although vision-language models have advanced perception and reasoning in autonomous driving, existing systems rarely consider the emotional dimension or real-world user experience. Keep Yelling Assistant (KYA) detects high-risk driving maneuvers in real time, such as sudden cut-ins. It then produces emotional responses through a large language model tailored to driver preferences. The framework comprises two core modules. The vision module uses YOLOv8 variants to detect nearby vehicles and identify risky behaviors such as sudden cut-ins. Key driving metrics, including relative distance, speed, and projected reach time, are extracted and normalized to produce a structured behavior log. The language module processes this log with user-defined emotional tone settings, such as neutral, humorous, and analytical, and generates verbal reactions using state-of-the-art large language models, including ChatGPT-4o, Claude 3, Gemini 2.5, and Copilot. We evaluated the proposed system using dashcam videos containing risky driving behaviors and a user study involving 108 participants. Participants selected preferred response styles, and the large language models were evaluated based on emotional alignment. All models received favorable ratings, although preferences varied across personas. Notably, the combination of YOLOv8s and ChatGPT-4o achieved the highest score of 4.29 out of 5.00. By integrating real-world perception with emotionally adaptive dialogue, KYA introduces a new paradigm for emotionally intelligent in-vehicle artificial intelligence. It offers promising directions for improving safety, trust, and emotional well-being in both conventional and autonomous vehicles.

Harine Choi, Eun Hak Lee, Zhengzhong Tu

Department of Computer Science and Engineering, Texas A&M University, 77840, College Station, Texas, United States · School of Railway Operation Systems, Kyungil University, Gyeongsan-si, Gyeongsangbuk-do, Republic of Korea

Human Driving Autonomous Driving Emotional Regulation Vision-Language Foundation Models Full-Duplex Voice Agents Assistant

Jul 14, 2026cs.CL

KnowAct-GUIClaw: Know Deeply, Act Perfectly, Personal GUI Assistant with Self-Evolving Memory and Skill

OpenClaw has emerged as a leading agent framework for complex task automation, yet it faces insufficient cross-platform GUI interaction support and a well-built self-evolution mechanism. These flaws limit its adaptation to diverse device ecosystems and prevent performance improvements through continuous learning from execution experience. To resolve these issues, we propose the Know Deeply, Act Perfectly paradigm for personal assistants, which holds that accumulated user interaction and task-running experience directly improve execution accuracy and efficiency, unifying cognitive comprehension and operational execution. Based on this paradigm, we introduce KnowAct-GUIClaw, a novel Know-Route-Act-Reflect framework designed to address OpenClaw's GUI manipulation deficits and break through its cross-platform and recursive self-improvement constraints. First, the host agent leverages accumulated interaction experience and task-relevant knowledge for long-horizon task decomposition and allocation (Know). Second, a pluggable GUI subagent with an experience-attributable memory system (Know) and self-evolving skill library (Act), enabling seamless cross-platform migration and fast-path integration. Especially, this framework continuously stores user profiles and feedback to improve the accuracy of task decomposition and tool calls. Extensive experiments across Android, iOS, HarmonyOS and Windows show that KnowAct-GUIClaw achieves superior efficiency, accuracy and cross-platform adaptability. Especially, the GUIClaw with open-source Kimi-2.6 models achieves the best performance (64.1%) on the long-horizon MobileWorld benchmark, beating all agentical frameworks and closed-source agentical models, e.g., Seed-2.0-Pro and GPT-5.5. Additionally, the knowledgeable memory and execution skills supported by our framework are transferable across diverse base models, improving by 8.5% with Kimi-2.6.

Yunxin Li, Jinchao Li, Shibo Su +7

Lychee Team∗, Harbin Institute of Technology, Shenzhen

Graphical User Interface Agents Agentic Framework Artificial Intelligence Assistants Self-Evolving Skill Libraries Assistant

Jul 13, 2026cs.CV

Vinci2: Providing Proactive Assistance in Continuous Egocentric Videos

When should an intelligent assistant speak up without being asked? Continuous egocentric video offers rich, evolving context that enables a new form of assistance: one that is proactive rather than merely reactive. Yet existing approaches either wait passively for user queries or treat every detected event as requiring a response, without considering the user's history, current activity, or whether assistance would actually be welcome. We reframe proactive assistance as a context-dependent decision problem: the agent must not only perceive what is happening, but reason over accumulated temporal context to determine when and whether to intervene. To this end, we present Vinci2, a proactive egocentric assistance system that advances the on-device assistant Vinci from reactive response toward proactivity. On the evaluation side, we present EgoServe, the first large-scale benchmark for proactive assistance in continuous egocentric video. EgoServe comprises over 3,000 service instances organized along 4 temporal memory horizons, ranging from immediate safety alerts to long-term habit coaching, across 10 service categories. On the modeling side, we propose EgoMemo, a training-free, memory-augmented agent that maintains three complementary memory representations: multi-scale temporal summaries, a semantic knowledge graph, and visual embedding archives. At each timestep, EgoMemo performs retrieval-augmented reasoning to determine whether assistance is warranted and, if so, produces contextually grounded responses. Experiments demonstrate that EgoMemo establishes strong baselines on EgoServe while remaining competitive on existing egocentric benchmarks. Our benchmark and code are publicly available at \href{https://sitonggong.github.io/EgoServe-page/}{Vinci2}.

Gong Sitong, Tianyu Yan, Caixin Kang +6

Dalian University of Technology · Alaya Lab · The University of Tokyo

Egocentric Video Artificial Intelligence Assistants Proactive Video Agent Assistant Egocentric Dataset+1

Jul 13, 2026cs.AI

The Path to Self-Evolving Clinical Systems: Scaling Medical Agents from Assistance to Autonomy

The growing ability of large language models and vision language models to jointly interpret and reason over images and text is reshaping medical agents, moving them from task specific predictors toward autonomous systems that perceive, reason, plan, remember, and act in clinical environments. This work departs from the capability first perspective of existing literature and instead begins from clinical deployment, asking what tasks, contamination resistant benchmarks, and interactive training environments are required before medical agents can be trusted in practice. Medical agents are formalized as sequential decision making systems under partial observability, together with a three level autonomy taxonomy spanning assisted, cooperative, and fully autonomous operation. The field is organized along a unified scaling spine consisting of framework scaling, capability scaling, and environment scaling. Within this framework, clinical environment scaling, the integration of tools, data, and clinical gyms, is identified as the most actionable yet underexplored direction for agents operating in PACS, EHR, and FHIR ecosystems. Clinical self evolution, where agents improve through interaction with their environments rather than parameter scaling alone, is further positioned as a key research frontier, drawing insights from self improving agents, agent gyms, and test time compute scaling. Applications across radiology, pathology, ophthalmology, and hospital workflows are examined together with deployment challenges including hallucination, cascading failures, and fairness. By consolidating more than 300 references, with particular emphasis on advances from 2025 to 2026, this work provides a roadmap toward trustworthy, self improving medical imaging systems for real clinical practice.

Chunzheng Zhu, Lei Tian, Bohan Tan +15

1Hunan University · 2ByteDance · 3Duke University +6

Self-Evolution Rate Scaling Assistant

Jul 13, 2026cs.DS

CGS: Configurable Graph Summarization with Bounded Neighborhood Loss and Query Support

Given a large graph, how to generate a compact summary graph that is configurable by the user and supports multiple graph queries with either no loss or with high accuracy? The ever growing size of graph datasets makes the above question on graph summarization very pertinent. Although, there are several approaches, there does not exist a configurable graph summarization method that offers high compression along with support for multiple graph queries on the summary graph with high accuracy, and allows the user to configure the summarization based on: (1) lossless or lossy summarization, (2) amount of tolerable neighborhood loss, (3) the type of loss it can tolerate, in terms of false positive edges (i.e., extra edges), false negative edges (i.e., missing edges), or neither, in both the (a) reconstructed graph and the (b) query answers. To overcome these limitations, we propose a novel graph summarization framework CGS (Configurable Graph Summarizer) that builds upon the idea of aggregating nodes with common neighborhoods. The CGS framework consists of three summarization variants, CGS-E, CGS-I and CGS-U. While CGS-E is a lossless scheme, CGS-I and CGS-U are lossy schemes that allow reconstruction of the input graph with no false positive edges and no false negative edges, respectively. To bound the graph reconstruction loss, we introduce a user-specified parameter neighborhood loss tolerance threshold, that limits the maximum loss allowed in the neighborhood of each node. This allows graph reconstruction and neighborhood query evaluation with either no loss or with bounded loss guarantees. Empirical evaluation on several synthetic and real-world graphs shows that CGS offers superior summarization than the state-of-the-art methods, and can answer graph queries with fairly high accuracy and efficiency.

Shubhadip Mitra, Sona Elza Simon, C Oswald +2

Blue Yonder India Pvt. Ltd., Bengaluru, India · Centre for Machine Intelligence and Data Science, Indian Institute of Technology Bombay, Mumbai, India · Dept. of Computer Science and Engineering, National Institute of Technology Tiruchirappalli, Tiruchirapalli, India +2

Summarization Assistant

Jul 9, 2026cs.LG

Provably Optimal Learning Algorithms for Assistance Games

This paper studies an online variant of the assistance games framework, where an informed agent and an uninformed agent repeatedly interact over

T

timesteps to optimize a common reward function. While the informed agent (the human) observes a latent state of the world, the uninformed agent (the assistant) observes only the human's actions. We provide the first provably efficient learning algorithms for repeated assistance games. We introduce the notion of assistance regret: the gap between the cumulative utility of interactions and that of the optimal joint policies in hindsight, which map latent states to action pairs. We present decentralized algorithms for both the human and the assistant that achieve a

(1-1/e)

-approximate assistance regret rate of

\widetilde{O}(T^{3/4})

, with runtime polynomial in the size of the action and state spaces. These algorithms are general; in particular, they accommodate any no-regret algorithm for the assistant. We prove that achieving a regret approximation factor better than

(1-1/e)

is computationally intractable. Furthermore, we demonstrate how these generic no-regret algorithms can be tailored to a pseudo-decentralized setting -- using a shared random string -- to achieve a rate of

\widetilde{O}(T^{1/2})

, optimal up to logarithmic factors.

Nivasini Ananthakrishnan, Mark Bedaywi, Michael I. Jordan +2

University of California, Berkeley

Optimal Learner Multi-Agent Reinforcement Learning Assistant Optimal Policies

Jul 8, 2026cs.RO

HELP: Human-Efficient Large-Scale Robot Post-Training with Rollout Segmentation

When adapting Vision Language Action (VLA) models to downstream tasks, multiple rounds of post-training are often required to progressively address policy weaknesses. In this report, we focus on maximizing human efficiency during this iterative process, measured by policy improvement and task throughput per unit of human labor and time. We propose HELP, a Human-Efficient Large-scale robot Post-training pipeline in which two specialized operators supervise twelve robots concurrently. A trained Teleoperator provides high-value remote interventions and recovery demonstrations, while a Floor Operator monitors the robot fleet, triggers takeovers, and performs physical resets. This role specialization improves human efficiency by reducing task switching, lowering operator training costs, and expanding robot interaction coverage. Beyond increasing rollout volume, concurrent supervision also broadens the range of policy behaviors observed by the human team, making recurring failure modes easier to identify and enabling more targeted takeovers, resets, and recovery demonstrations. To efficiently utilize the large and mixed-quality rollout data, HELP incorporates \vlac, an automatic rollout segmentation critic specifically designed for this setting. It separates autonomous trajectories into progress-making, idle, failure-inducing, and recovery segments. Useful rollout segments are retained and combined with Human-in-the-Loop data for the next post-training round. Across four real-world manipulation tasks, HELP achieves 80%--95% success rates and improves task throughput by 1.7

\times

--4.2

\times

over the base model. Under matched HITL recovery budgets, VLAC-CUT further amplifies throughput gains by 1.20

\times

--3.43

\times

and success-rate gains by 1.50

\times

--3.00

\times

over HITL-only updates.

Shaopeng Zhai, Qi Zhang, Tianyi Zhang +5

Continuous Learning Team, Shanghai AI Laboratory

Humanoid Robotics Rollout Process-Level Supervision Post-Training Human Agency Assistant

Jun 23, 2026eess.AS

Phoneme-Level Mispronunciation Screening in Polish-Speaking Children with an Explainable Assistant

Early identification of speech sound errors in children is often limited by access to specialists, motivating lightweight screening tools that can operate outside the clinic. We present a screening pipeline for Polish-speaking children focused on sibilant substitutions, coupling a wav2vec2-based CTC token recognizer with alignment-based error typing and a template-grounded caregiver assistant for screening, not diagnosis. On a held-out test set of 10 unseen children comprising 559 utterances, the recognizer achieves 88.7 percent exact sequence match. As a conservative screening proxy, we flag a mismatch when the system emits substitution-evidence bracketed tokens at the target segment, yielding 72.9 percent precision, 61.4 percent recall, F1 = 0.67, and a 2.7 percent false-alarm rate on target-correct items. We describe the assistant's safety boundaries and outline a clinician-in-the-loop validation plan for future deployment.

Milosz Dudek, Daria Hemmerling, Kamil Kwarciak +9

AGH University of Krakow, Cracow, Poland · SoftServe, Cracow, Poland · Department of Biomedical Engineering, Silesian University of Technology, Poland +1

Pronunciation Assessment Screening Substitute Children Early Layers Assistant

Jun 23, 2026cs.AI

A specialized reasoning large language model for accelerating rare disease diagnosis: a randomized AI physician assistance trial

Rare diseases affect millions of individuals worldwide, yet timely diagnosis remains a major public health challenge due to scarcity of specialized clinical expertise. While large language models (LLMs) show promise to support rare disease diagnosis, current models are constrained by insufficient clinical deployability, limited clinically grounded evidence, and scarcity of training data. Here we present RaDaR (Rare Disease navigatoR), an open-source, compact reasoning LLM (32B parameters) for rare disease diagnosis. RaDaR was trained with 49,170 publicly available free-text cases and 104,666 synthetic cases with reasoning-enhanced training. RaDaR showed the strongest performance among evaluated open-source models, including the 671B DeepSeek-R1, across public benchmarks and four external validation centers. In a retrospective cohort, RaDaR prioritized the final diagnosis before documented clinical suspicion in 61.06 percent of cases, corresponding to a potential lead time of 1.87 months and 50.18 percent of the within-center interval. In a randomized physician-assistance trial, RaDaR assistance improved physicians' rare-disease diagnostic accuracy by 21.44 percentage points compared with internet search alone. Synthetic-data ablations suggested that phenotype-anchored narratives provide useful training signal for long-tail rare diseases, with a monotonic scaling trend within the tested data range. Together, RaDaR and its development and validation framework provide a deployable rare-disease reasoning model and a reproducible development framework for diagnostic AI under data scarcity.

Haichao Chen, Songchi Zhou, Zhengyun Zhao +28

Tsinghua Medicine, Tsinghua University, Beijing, China. · Department of Ophthalmology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, China. · Department of Statistics and Data Science, Tsinghua University, Beijing, China. +19

Rare Disease Reasoning Skills Controlled Study Assistant

Jun 22, 2026cs.CL

When Does Intrinsic Self-Correction Help? A Task-Sensitive Analysis

Intrinsic self-correction (SC) aims to improve large language model outputs by prompting a model to revisit its own initial answer without external feedback. Recent studies have questioned the reliability of this approach, showing that models often struggle to judge whether their initial responses are correct. In this work, we take a task-sensitive view of SC. Rather than asking whether it works in general, we examine settings where SC may operate through different mechanisms: verifying explicit constraints, revisiting a complex reasoning process, or providing a second opinion over competing strategies in word-game tasks. Across multiple benchmarks and models, we find that SC can yield consistent performance gains when the underlying task structure facilitates these modes of revision. These results suggest that SC is best understood as a task-dependent inference-time strategy whose usefulness depends on the role the revision stage can play in a given task, rather than as a uniformly reliable method for improving initial model outputs.

Elroy Stav, Dvir Berlowitz, Maayan Orner +1

Department of Computer Science and Artificial Intelligence, Bar-Ilan University · Independent Researcher

Iterative Self-Correction Assistant

Jun 22, 2026cs.RO

Assistron: Bayesian Shared Autonomy with Off-the-shelf Vision-Language-Action Models

We propose Assistron, a shared autonomy model that leverages Vision-Language-Action (VLA) models to assist the user in daily activities. Our approach is grounded in two core principles: (1)~minimizing human cognitive and physical effort by leveraging VLA-driven autonomy for macro-movements, and (2)~prioritizing human intervention specifically at critical failure points. Driven by the user's verbal language commands, Assistron utilizes the VLA to autonomously execute macro-reaching trajectories, saving users' effort. In contact-rich interactions where VLAs tend to fail, Assistron employs a phase-aware interaction detection mechanism and solicits the user to intervene, in turn adjusting the VLA's action generation via flow matching guidance. Critically, our formulation eliminates the need for VLA fine-tuning, protecting its broad behavioral priors from catastrophic forgetting and ensuring the model does not become a narrow specialist. We validate our approach on a comprehensive multi-task scene recovery benchmark encompassing diverse daily manipulation skills. Empirical results demonstrate that Assistron significantly improves task success rates over pure autonomous baselines while significantly reducing human cognitive and physical workload compared to traditional teleoperation, offering a scalable, smooth, and effortless paradigm for assistive manipulation. The code is available in https://github.com/mousecpn/Assistron.git.

Pinhao Song, Ze Fu, Yutong Hu +1

KU Leuven, Dept. Mechanical Engineering, Research unit Robotics, Automation and Mechatronics · KU Leuven, Dept. Electrical Engineering, Research unit Processing Speech and Images

Diffusion-Based Vision-Language-Actions Teleoperation Assistant

Jun 19, 2026cs.PF

Does Mixture-of-Experts Actually Help Inference on Consumer and Edge Hardware? An Empirical Study

Mixture-of-Experts (MoE) language models are often described as ideal for resource-constrained inference. Each token activates only a small subset of experts, so the per-token compute cost, in floating-point operations (FLOPs), resembles that of a much smaller dense model. Whether that FLOP advantage survives in practice is far less clear. We ask whether MoE models actually run faster and cheaper than comparable dense models on consumer-grade and edge hardware. We benchmark OLMoE-1B-7B (1.3 B active of 6.9 B total) against three dense baselines on an Apple M2 Pro and an NVIDIA Jetson Orin Nano 8 GB through llama

.

cpp, measuring throughput, memory, and on-device energy. The answer is device-dependent: OLMoE's active-parameter advantage is only partly realised on the laptop (~10% behind the same-active Llama-3.2-1B) and erodes on the edge device (~31% behind, at 2.1

\times

the energy per token, with peak memory at the 8 GB ceiling). Patching llama

.

cpp to time the decode graph node-by-node shows routing accounts for under 9% of MoE-block compute on the cleaner edge backend, so the gap reflects total-parameter memory footprint, expert dispatch, and KV-cache pressure rather than routing. The implication is that on bandwidth-bound edge hardware, inference cost tracks total parameters, not active ones, and sparse activation does not buy back what the device is constrained on. These findings are bounded to one MoE model at this parameter scale and two devices, and we release the full measurement harness and per-run data.

Alfarizy Alfarizy, Hung Truong Thanh Nguyen, René Richard +2

Analytics Everywhere Lab, University of New Brunswick, Fredericton, NB, Canada · TSAI Lab, University of New Brunswick, Fredericton, NB, Canada

Mixture-Of-Expert Inference Mixture-Of-Experts Localized Lora-Moe Inference Cost Edge Hardware Peak Memory+2

Jun 15, 2026cs.RO

Abstention-Aware Personalized Object Rearrangement via Uncertainty-Guided LLM Assistance

Robotic assistance in household environments requires not only predicting where objects should be placed, but also reasoning about when objects should not be placed at all. Existing approaches to personalized object rearrangement primarily focus on placement decisions under the assumption of clean observations and complete actionability, limiting their applicability in realistic, cluttered, and partially erroneous settings. In this paper, we introduce APOLLO, a hybrid framework for abstention-aware personalized object rearrangement that combines a lightweight, personalized embedding model (PEM) with selective large language model (LLM) assistance. PEM is trained for each user-environment pair using a small number of demonstrations, operates entirely on CPU, and produces uncertainty estimates, which are used to selectively invoke LLM-based reasoning only for ambiguous decisions, balancing efficiency, privacy, and reasoning capability. To evaluate this formulation beyond existing benchmarks, we introduce APOR, a synthetic, LLM-generated dataset that captures room-level, multi-furniture environments, diverse organizational profiles, explicit abstention behavior, and noisy partial scene context. Extensive experiments on both PARSEC and APOR provide initial evidence that APOLLO improves over prior LLM-based baselines in controlled benchmark settings while substantially reducing LLM usage. Code is available at https://github.com/PaInt-Lab/APOLLO.

Sam Collin, Ali Ayub

Concordia University, Montreal, Quebec H3G 1M8, Canada

Conceptual Reorganization Affordances Indoor Environments Assistant Abstention

Jun 11, 2026cs.CL

When Does Mixing Help? Analyzing Query Embedding Interpolation in Multilingual Dense Retrieval

While mixed-language querying is ubiquitous in multilingual communities, the sensitivity of dense retrievers to such queries remains poorly understood. We present a ratio-controlled study on mMARCO that systematically evaluates retrieval performance by varying the mixing proportion of parallel query translations via embedding-level mixing -- constructing mixed queries as an interpolation of monolingual embeddings. Experiments with BGE-M3 demonstrate that an optimal mixing ratio outperforms the best monolingual endpoint in 88/105 cases. We uncover a distinct asymmetry driven by English dominance: mixing is uniformly beneficial when retrieving from non-English document indices, whereas indices containing English are best served by pure English queries. Furthermore, English acts as the strongest mixing partner for every non-English document language. Finally, when controlling for English dominance, mixing gains correlate negatively with typological distance. We conclude that language-mix sensitivity is structured and predictable, and we validate the robustness of these patterns across model families and scales.

Tongyao Zhu, Chao-Ming Huang, Min-Yen Kan

National University of Singapore

Retrieval Layer Retrievers Queries Interpolation Assistant

Jun 11, 2026cs.LG

Understanding helpfulness and harmless tension in reward models

Reward models are a key component of reinforcement learning from human feedback (RLHF), aligning language models toward both helpful and harmless behaviour. However, the internal mechanisms underlying these objectives and their conflicts remain poorly understood. We study alignment tension in reward models trained under helpfulness-only, harmlessness-only, and mixed-objective settings. We find that mixed-objective models often underperform single-objective models, indicating interference between objectives. Using activation-based methods, we identify neurons associated with each objective and study their functional roles via targeted ablations. We find that these neurons causally support their corresponding objectives while often negatively affecting the opposing one. We find that a substantial proportion of neurons are shared between helpfulness and harmlessness, and that these shared neurons exert a disproportionate influence on model behaviour, contributing to alignment tension. Additionally, our results provide insights and mechanistic interpretation into how alignment objectives are represented in reward models and why multi-objective alignment remains challenging, motivating future work on disentangled and controllable alignment methods.

Eshaan Tanwar, Pepa Atanasova

University of Copenhagen

Progress Reward Modeling Value Alignment Neurons Assistant

Jun 10, 2026cs.CL

Helping Figures Tell their Story! Paper-Grounded Video Generation Explaining Complex Scientific Figures

Scientific figures compress complex pipelines into a single canvas, yet understanding them requires paper-grounded, step-by-step narration aligned with visual highlights a capability missing from current video generation systems and benchmarks. To address this, we introduce paper-grounded figure-to-video generation: generating narrated, region-grounded walkthrough videos from a figure and its paper. We propose MINARD (Multimodal Interpretation of Narrated Architecture via Region Decomposition), a pipeline that generates paper-grounded narrations and sequentially grounds them to figure regions. We also release FigTalk, a benchmark with new sequential and component-level grounding metrics derived. On FigTalk, MINARD generates humanlike, paper-faithful narrations and outperforms narration-conditioned figure spatial grounding compared to existing approaches in both automatic and human evaluation

Ishani Mondal, Javad Baghirov, Jordan Boyd-Graber

University of Maryland, College Park

Long Video Generation Scientific Figure Video Storytelling Narratives Spatial Grounding Multi-Modal Data+1

Jun 10, 2026cs.RO

Improving Human Diving Endurance with a Field-Deployable, Untethered Exoskeleton

Human endurance in underwater locomotion is fundamentally restricted by high energetic demands to overcome drag and the finite supply of self-contained breathing gas. While exoskeleton technology can reduce the metabolic cost of humans in terrestrial locomotion, its potential to enhance human endurance during underwater diving remains entirely unexplored. Here, we present DiveMate, a field-deployable, untethered exoskeleton designed to improve human diving endurance via adaptive kick assistance in real-world underwater environments. During naturalistic diving, DiveMate increases the travel distance using a given energy (breathing gas) by 42.9% and extends dive duration by 54.9% through reducing gas consumption rate. Marked reductions in muscle activation indicate a decrease in physiological exertion, with the net gas consumption rate decreasing by 47.0%. Kinematic characteristics and regularity improvements further underpin efficient energy economy. These results suggest that applying exoskeleton assistance is beneficial for improving human diving endurance and augmenting their ability to explore the aquatic world. This study extends the application frontier of exoskeletons and provides a potential reference for the design and assessment of future underwater assistive devices.

Zhihao Zhou, Zhenmeng Ju, Rui Yang +8

Underwater Robots Exoskeleton Assistant

Jun 10, 2026cs.SD

SpAArSIST: Sparsified AASIST for Efficient and Reliable Anti-Spoofing

We present SpAArSIST, a deployment-oriented refinement of the widely used AASIST graph pooling backend for self-supervised learning (SSL) based anti-spoofing. Motivated by redundant operations in public implementations, we replace learned pooling and stack-node attention with explicit, lightweight choices: separate train and inference graph pooling ratios

(k_{\mathrm{tr}},k_{\mathrm{inf}})

, magnitude-based node scoring, and mean aggregation of graph nodes. The best overall configuration (rank 1) cuts backend compute by 20.7% (195.045M

\rightarrow

154.706M MACs) and model size by 4.1% (611.8k

\rightarrow

586.4k params), while improving out-of-domain robustness on In-the-Wild to 2.82% EER and 0.078 minDCF (from 4.64% and 0.133) and remaining competitive on ASVspoof5. We further provide a composite selection score that summarizes accuracy, calibration, and compute to support balanced deployment-oriented model choice.

Anton Firc, Vojtěch Staněk, Zbyněk Lička +2

Spoofing Self-Supervised Learning Fast Inference Global Average Pooling Assistant

Jun 3, 2026cs.RO

SoftPINCH: EMG-Driven Soft Exoskeleton Assistance for Finger Flexion and Grasping

Surface electromyography (sEMG) provides a non-invasive interface for detecting hand-movement intention and controlling wearable assistive devices. However, reliable EMG-driven hand assistance remains challenging because EMG signals are affected by noise, motion artifacts, electrode placement, muscle fatigue, and inter-subject variability. At the same time, many hand exoskeletons remain mechanically restrictive or bulky, limiting comfort and natural hand motion. This work presents SoftPINCH, an EMG-driven soft wearable exoskeleton for thumb-index finger flexion and pinch grasp assistance. The system combines a tendon-driven soft exoskeleton, fingertip magnetic contact sensing, and neural EMG decoding for intention-based assistance. Surface EMG was recorded from forearm muscles during index and thumb movements, and three subject-independent decoding architectures were evaluated: LSTM, CNN+LSTM, and CNN+LSTM with attention. The CNN+LSTM and CNN+LSTM-attention models both achieved 99.4% LOSO test accuracy, outperforming the standalone LSTM, which reached 97.8%. However, the attention mechanism did not provide a significant improvement over CNN+LSTM, indicating that CNN-based feature extraction was sufficient for robust EMG representation. The CNN+LSTM model was therefore selected for real-time deployment due to its high accuracy and lower architectural complexity. Functional evaluation showed that active exoskeleton assistance reduced muscular effort during isolated finger flexion and object grasping. During weighted grasping, assistance reduced muscular effort across all tested loads, with a 92.6% reduction at the highest load. These results demonstrate the potential of SoftPINCH for intuitive, low-effort pinch assistance using real-time EMG-driven soft robotic control.

Nicklas Nikolaj Grønvall, Magnus Malthe Sigsgaard Nielsen, Xiaofeng Xiong +1

SDU Soft Robotics, The Maersk Mc-Kinney Moller Institute, University of Southern Denmark, Odense, Denmark

Exoskeleton Finger Cnn-Lstm Assistant

Jun 1, 2026cs.AI

When Helping Hurts and How to Fix It: Multi-Agent Debate for Data Cleaning

When does multi-agent debate help data cleaning, and when does it hurt? Across three benchmarks, four model families, and over 6,000 task-condition pairs, we find debate's effect reverses sign: it degrades generation across all four models (-1.6 to -15.5pp) through critique-induced confusion (CIC), hallucinated Critic feedback that the Generator accepts uncritically, yet improves error detection (+27.4pp F1, d=1.0). We derive a debate benefit condition: debate helps when the probability of rescuing a wrong output (Critic verification odds weighted by fixability) exceeds the probability of destroying a correct one. A factorial experiment proves adversarial separation is essential: self-verification with identical tools fails, while a separate Critic with code-execution grounding and evidence-gated generation produces the first debate configuration to significantly exceed single-agent on a generative task (+5.3pp, p<0.05). The condition correctly predicts all nine task types and generalizes with zero false positives across 19 published comparisons in seven domains.

Chirag Parmar, Akshat Mehta, Henglin Wu +2

1Meta Platforms, Inc.

Multi-Agent Debate Debate Critique Mixture-Of-Agents Critic Assistant

Jun 1, 2026cs.CY

LLMs as Teaching Assistants for Mathematics Exam Grading: Reliability, and Practical Usability

Open-ended mathematics exams are valuable because they assess reasoning, proof construction, algorithmic thinking, and communication of intermediate steps. They are also difficult to grade at scale because instructors must apply partial-credit rubrics consistently while giving feedback that helps students repair misconceptions. This paper evaluates six contemporary large language model (LLM) configurations, Gemini 3.1 Pro Extended, Gemini 3.5 Flash, ChatGPT 5.5 Pro Extended, ChatGPT 5.5 Thinking, Claude Pro Opus 4.7, and Claude Sonnet 4.6, as grading assistants for an undergraduate discrete mathematics examination. The study compares two grading policies. The BASELINE policy uses a stricter rubric-following prompt that emphasizes explicit evidence and complete justification. The LIBERAL policy was added after preliminary grading showed that the baseline condition sometimes applied harsh point deductions and failed to recognize valid partial reasoning. Agreement with human grading is measured at both the question and exam-total levels using mean absolute error, root mean squared error, normalized root mean squared error, Pearson correlation, and exact agreement. The results show that liberal partial-credit prompting reduces average question-level error for every evaluated model family. ChatGPT 5.5 Thinking (LIBERAL) has the lowest average question-level MAE (1.87) and RMSE (2.53), while Gemini 3.1 Pro Extended (LIBERAL) has the lowest total-score MAE (8.00) and RMSE (10.66). However, the strongest total-score Pearson correlation occurs under Gemini 3.1 Pro Extended (BASELINE) at 0.58, showing that point calibration and rank preservation remain distinct goals. We also report practical usability observations.

Aastha Sapkota, M. G. Sarwar Murshed

Computer Science Department University of Wisconsin - Green Bay, Green Bay, WI, USA

Grading Preliminary Study Assistant Spearman Correlation

May 30, 2026cs.LG

The Assistant as a Privileged Persona: A canonical reference in cross-persona self-recognition

Post-trained language models can recognize their own outputs from a sentence or two out of context. In a companion paper \citep{jack2026twomodes} we showed they can also recognize when they are currently acting on-policy, through the sharp entropy drop of assistant-mode generation. Both signals are tied to the Assistant persona that post-training mainly shapes. This paper widens the frame to cross-persona authorship judgement on Llama-3.1-70B-Instruct. We measure a matrix of authorship claim rates over a panel of evaluator and generator personas spanning librarian to dragon to Shakespeare, and make two claims. \emph{First}, on the Assistant's own row of the matrix, the Assistant's claim rate, the persona-vector distance from the Assistant in activation space, and the entropy gap between the Assistant's surprise on a persona's text and the persona's surprise on its own text are all tightly coupled. This extends the entropy signature of \emph{acting} from the companion paper to a retrospective signature of \emph{having acted}. \emph{Second}, this coupling fails off the Assistant's row: the natural symmetric extension of the entropy gap does not predict authorship for distinctive evaluators (pirate, dragon, Shakespeare); what does is asymmetric -- the evaluator's surprise compared to the Assistant's surprise on the same text, not to the generator's. We rule out the alternative that any persona could play this reference role by trying many candidate substitutes; none does. We interpret the asymmetry as the model performing an implicit Bayesian likelihood-ratio test against the Assistant as the canonical alternative hypothesis, with the persona-vector geometry of \citet{chen2025persona} (every persona a delta off the Assistant) ensuring that the Assistant is the only persona universally accessible to that test.

Asvin G

Institute for Advanced Study, Princeton

Artificial Intelligence Personas Authorship Self-Knowledge Assistant Privileged Context Signatures

May 28, 2026cs.AI

The Curse of Helpfulness: Inverse Scaling Law in Robustness to Distractor Instructions via DistractionIF

Large Language Models (LLMs) are increasingly deployed in agentic and retrieval-augmented generation (RAG) systems, where they must execute user-specified tasks over externally provided reference text. In practice, such context is often unstructured and contaminated with benign but instruction-like semantic noise, such as editorial comments and system traces, which should be treated strictly as data. We introduce DistractionIF, a benchmark designed to evaluate robustness against such distractor instructions in reference text. Across a broad range of models, we observe a consistent inverse scaling phenomenon: larger models are often less robust, with performance dropping by up to 30 points as scale increases. Mechanistically, our perplexity analysis reveals that scaling erodes the probabilistic boundary between robust and distracted behaviors, making models increasingly prone to over-interpreting noise as instructions. To address this, we demonstrate that reinforcement learning, specifically Group Relative Policy Optimization (GRPO), can restore this boundary, improving robustness by up to 15.5% without compromising general instruction-following capability. Our findings highlight a critical instruction-following robustness gap in reference-grounded tasks and establish reinforcement learning as a promising path for enforcing strict data-instruction separation at scale.

Zeli Su, Zhankai Xu, Tianlei Chen +4

Minzu University of China, Beijing, China · Ant Group, Hangzhou, China · Renmin University of China, Beijing, China +1

Large Language Model Reinforcement Learning Instruction-Tuned Models Distractor Scaling Assistant Scaling Laws

May 27, 2026cs.AI

An LLM-Based Assistance System for Intuitive and Flexible Capability-Based Planning

In modern industry, dynamic environments and the complexity of modular and reconfigurable resources require automated planning of process sequences. Capability-based planning approaches address this by automatically generating plans from semantic knowledge models that describe resource functions in a machine-interpretable form. Their practical use, however, remains limited: solver feedback, especially in the case of unsatisfiability, is difficult to interpret, and the knowledge models require adaptation as operational conditions change or requests become infeasible. This paper presents a hybrid assistance system that augments an existing capability-based Satisfiability Modulo Theories (SMT) planning approach with an Large Language Model (LLM)-based layer for natural-language interaction, explanation, and adaptation. Formal planning correctness remains with the symbolic planner, while the LLM layer handles natural-language access and flexible knowledge model adaptation under explicit Human-in-the-Loop (HitL) approval. The system decomposes into four components: Capability Grounding, Symbolic Planning, Result Interpretation, and Planning Adaptation, realized as a routed agentic workflow in which a central router delegates to five specialized agents. The system is evaluated on a modular production system across four scenario types. Of 23 test cases, 9 of 10 knowledge queries and all 4 satisfiable planning cases were handled correctly, 3 of 4 unsatisfiable cases produced concrete repair proposals, and all 5 adaptive planning scenarios resolved into satisfiable plans through iterative, user-approved knowledge model modifications. The findings confirm that combining formal planning with LLM-based assistance substantially improves accessibility and adaptability in industrial automation.

Luis Miguel Vieira da Silva, Nicolas König, Felix Gehlhoff

Institute of Automation Technology, Helmut Schmidt University

Large Language Model Planning Classical Planning Agentic Workflows Intuition Assistant Parameter-Efficient Adaptation

May 15, 2026cs.LG

R2V Agent: Teaching SLMs When to Ask for Help

Efficient agentic systems should incur expensive frontier-model costs only on decisions where a cheaper local model is likely to fail. Existing LLM cascades usually route whole queries before execution, but task difficulty shifts mid-trajectory - after flaky tool calls, truncated observations, or compounding local errors - making pre-execution routing brittle. We introduce \textbf{R2V-Agent}, a risk-calibrated SLM-LLM routing framework for interactive agents. R2V combines four components: a distilled small language model (SLM) policy, a stronger teacher LLM, a lightweight process verifier that scores candidate actions at each step, and a calibrated step-level router. The router is our central contribution: after the SLM is trained, it estimates residual failure risk at each step and escalates only when teacher intervention is warranted. To make the routing problem well-defined, we first train a stable local SLM using a standard offline pipeline: behavioral cloning (BC) on teacher trajectories, followed by verifier-guided Direct Preference Optimization (DPO) with consistency regularization. The router is then trained on this fixed policy's residual failures using Brier-calibrated probability estimation and a Conditional Value-at-Risk (CVaR)-constrained objective that penalizes worst-case failures across perturbation seeds. Across HumanEval+, TextWorld, and TerminalBench with four SLM backbones, R2V improves the reliability-cost frontier: it achieves

94.3\%

HumanEval+ success with

0.60\%

LLM escalation, recovers TextWorld from

64.6\%

SLM-only success to

98.2\%

at

41.7\%

escalation, and reaches

93.3\%

TerminalBench success at

33.9\%

LLM calls, roughly half the heuristic-router cost.

Raghu Vamshi Hemadri, Humaira Firdowse Mohammed, Rishabh Maheshwary +5

New York University Tandon School of Engineering · University of California, San Diego · ServiceNow Research +3

Large Language Model Routing Cp-Agent Assistant Conditional-Value-At-Risk

May 14, 2026astro-ph.IM

Beyond AI as Assistants: Toward Autonomous Discovery in Cosmology

Recent advances in artificial intelligence (AI) agents are pushing AI beyond tools toward autonomous scientific discovery. We discuss two complementary agentic systems for cosmology: \texttt{CMBEvolve}, which targets tasks with explicit quantitative objectives through LLM-guided code evolution and tree search, and \texttt{CosmoEvolve}, which targets open-ended scientific workflows through a virtual multi-agent research laboratory. As preliminary demonstrations, we apply \texttt{CMBEvolve} to out-of-distribution detection in weak-lensing maps, where it iteratively improves the benchmark score through code evolution, and \texttt{CosmoEvolve} to autonomous ACT DR6 data analysis, where it identifies non-trivial pair- and scale-dependent behaviour and produces analysis-grade diagnostics. These examples show how cosmology can provide both controlled benchmark tasks and realistic open-ended research problems for the development of AI scientist systems.

Licong Xu, Thomas Borrett

Institute of Astronomy, University of Cambridge, Madingley Road, Cambridge CB3 0HA, UK · Kavli Institute for Cosmology, University of Cambridge, Madingley Road, Cambridge CB3 0HA, UK · Cavendish Astrophysics, University of Cambridge, Madingley Road, Cambridge CB3 0HA, UK

Artificial Intelligence Scientists Cosmology Agentic Discovery Astronomical Images Model Discovery Discovery+1

May 11, 2026cs.CL

VISTA: A Generative Egocentric Video Framework for Daily Assistance

Training AI agents to proactively assist humans in daily activities, from routine household tasks to urgent safety situations, requires large-scale visual data. However, capturing such scenarios in the real world is often difficult, costly, or unsafe, and physics-based simulators lack the visual fidelity needed to transfer learned behaviors to real settings. Therefore, we introduce VISTA, a video synthesis system that produces high-fidelity egocentric videos as training and evaluation data for AI agents. VISTA employs a 5-step script generation pipeline with causal reverse reasoning to create diverse, logically grounded intervention modes. These scenarios span two levels of agent autonomy: reactive and proactive. In reactive modes, the user explicitly asks the agent for help. In proactive modes, the agent offers help without receiving a direct request. We further divide proactive modes into explicit and implicit types. In explicit proactive scenarios, the user is aware of needing help but does not directly address the agent. In implicit proactive scenarios, the agent intervenes before the user even realizes that help is needed. VISTA allows users to customize and refine scenarios to generate video benchmarks for daily tasks, offering a scalable and controllable alternative to real-world data collection for training and evaluating AI agents in realistic environments.

Yu-Hsiang Liu, Yu-Chien Tang, An-Zi Yen

Department of Computer Science, National Yang Ming Chiao Tung University, Taiwan

Video Agent Egocentric Video Artificial Intelligence Agents Vista Proactive Assistant+1

May 10, 2026cs.CL

Quantifying the Utility of User Simulators for Building Collaborative LLM Assistants

User simulators are increasingly leveraged to build interactive AI assistants, yet how to measure the quality of these simulators remains an open question. In this work, we show how simulator quality can be quantified in terms of its downstream utility: how an LLM assistant trained with this user simulator performs in the wild when interacting with real humans. In a controlled experiment where only the user simulator varies, we train LLM assistants via reinforcement learning against a spectrum of simulators, from an LLM prompted to role-play a user to one fine-tuned on human utterances from WildChat. As evaluation, we measure pairwise win rates in a user study with 283 participants and on WildBench, a benchmark derived from real human--AI conversations. Training against the role-playing LLM yields an assistant statistically indistinguishable from the initial assistant in our user study (51% win rate), whereas training against the fine-tuned simulator yields significant gains (58% over the initial and 57% over the one trained against role-playing). Closer inspection reveals three further patterns: methods for making role-playing LLMs more realistic (e.g., persona conditioning) improve trained assistants but do not close the gap to the fine-tuned simulator; scaling the simulator's model size benefits the fine-tuned simulator but yields no gain for role-playing ones; and assistants trained against role-playing simulators fail to generalize when paired with other simulators at test time, while the one trained against fine-tuned simulator does. Together, these results argue for grounding user simulators in real human behavior and measuring their quality by their downstream effect on real users.

Joseph Suh, Ayush Raj, Minwoo Kang +1

UC Berkeley

User Simulation Role-Playing Agents Assistant Quantifying

May 7, 2026cs.RO

AssistDLO: Assistive Teleoperation for Deformable Linear Object Manipulation

Manipulating Deformable Linear Objects (DLOs) is challenging in robotics due to their infinite-dimensional configuration space and complex nonlinear dynamics. In teleoperation, depth uncertainty hinders state perception and reaction. AssistDLO addresses this challenge as an assistive teleoperation framework for DLO manipulation that combines real-time multi-view state estimation, visual assistance (VA), and a geometry-aware shared-autonomy controller based on Control Barrier Functions (SA-CBF). While traditional shared autonomy methods often rely on simple geometric attractors and may fail to preserve DLO geometry, SA-CBF acts as a geometry-aware funnel, facilitating precise grasping while preserving the operator's high-level authority. The framework is evaluated in a bimanual knot-untangling user study (N = 22) using ropes with varying length and rigidity. Results show that the effectiveness of the assistance depends strongly on operator expertise and DLO properties. SA-CBF provides the strongest gains for naive users, acting as a skill equalizer that increases task success from 71% to 88%, and is effective for stiffer ropes. Conversely, expert users prefer VA, and highly compliant, long ropes benefit more from visual support than localized action assistance. Ultimately, these findings demonstrate that effective DLO teleoperation cannot rely on a fixed strategy, highlighting the critical need for adaptive, user-aware, and material-aware shared autonomy.

Berk Guler, Simon Manschitz, Kay Pompetzki +1

TU Darmstadt, Institute for Intelligent Autonomous Systems, Computer Science Department, 64289 Darmstadt, Germany · Honda Research Institute Europe GmbH, Carl-Legien-Straße 30, 63073 Offenbach/Main, Germany · German Research Center for AI (DFKI), Research Department: Systems AI for Robot Learning; Hessian.AI; Robotics Institute Germany (RIG); Centre for Cognitive Science, Hochschulstr. 10, 64293 Darmstadt, Germany

Teleoperation Deformable Linear Objects Bimanual Manipulation Control Barrier Functions Assistant

May 7, 2026cs.SE

An Empirical Study of Proactive Coding Assistants in Real-World Software Development

Large language model (LLM)-based coding assistants have made substantial progress, yet most systems remain reactive, requiring developers to explicitly formulate their needs. Proactive coding assistants aim to infer latent developer intent from integrated development environment (IDE) interactions and repository context, thereby reducing interaction overhead and supporting more seamless assistance. However, research in this direction is limited by the scarcity of large-scale real-world developer behavior data. Existing studies therefore often rely on LLM-simulated IDE traces, whose fidelity to real development behavior remains unclear. In this paper, we investigate this simulation-to-reality gap through a large-scale empirical study. We collect real IDE interaction traces from 1{,}246 experienced industry developers over three consecutive days using a custom Visual Studio Code extension, and construct paired LLM-simulated traces for controlled comparison. Our analysis shows that simulated traces differ substantially from real traces in behavioral diversity, temporal structure, and exploratory patterns. Based on the collected data, we introduce \textbf{ProCodeBench}, a real-world benchmark for proactive intent prediction. Experiments with representative LLMs, retrieval-augmented methods, and agentic baselines show that current approaches remain far from reliable under real IDE traces, suggesting that simulation-based evaluation can overestimate real-world performance. Finally, our training study shows that simulated data cannot replace real data, but can complement it when used before real-world fine-tuning. These findings highlight the importance of real developer behavior data for evaluating and training proactive coding assistants.

Lehui Li, Ruixuan Jia, Guo-Ye Yang +1

College of AI, Tsinghua University, China · Fitten Tech Co., Ltd., China

Ai-Assisted Programming Tasks User Simulation Software Engineering Empirical Study Traces Single-Agent Baselines+1

May 7, 2026cs.LG

When Can Voting Help, Hurt, or Change Course? Exact Structure of Binary Test-Time Aggregation

Majority voting is one of the few black-box interventions that can improve a fixed stochastic predictor: repeated access can be cheaper than changing a high-capability model. Classical fixed-competence theory makes this intervention look monotone -- more votes help above the majority threshold and hurt below it. We show that this picture is fundamentally incomplete. Under the de Finetti representation for exchangeable repeated correctness, voting is governed by a latent distribution of per-example correctness probabilities. Even simple latent mixtures can generate sharply different voting curves, including nonmonotone behavior and, in an explicit construction, infinitely many trend changes. The full latent law determines the curve, but the curve does not determine the law. The exact object recovered by voting is a signed voting signature: at each binomial variance scale, it records excess latent mass above rather than below the majority threshold. Our main theorem proves that the complete odd-budget curve and this signature are equivalent: the curve increments are signed Hausdorff moments, and the full curve recovers the signature uniquely. This viewpoint explains shape phenomena, branch-symmetric nonidentifiability, realizability, variation, and endpoint rates. It also separates estimation regimes: direct per-example success-probability information targets the full signature, whereas fixed-depth grouped labels reveal only a finite prefix.

Yi Liu

York University

Majority Voting Exact Change Online Signature Verification Assistant

May 6, 2026cs.AI

When Helpfulness Becomes Sycophancy: Sycophancy is a Boundary Failure Between Social Alignment and Epistemic Integrity in Large Language Models

This position paper argues that sycophancy in LLMs is a boundary failure between social alignment and epistemic integrity. Existing work often operationalizes sycophancy through external behavior such as agreement with incorrect user beliefs, position reversals, or deviation from an objective standard of correctness. These formulations capture only overt forms of the phenomenon and leave subtler boundary failures involving epistemic integrity and social alignment underspecified. We argue that sycophancy should not be understood as agreement alone, but as alignment behavior that displaces independent epistemic judgment. To clarify this boundary, we propose a three-condition framework for sycophancy. First, the user expresses a cue in the form of a belief, preference, or self-concept. Second, the model shifts toward that cue through alignment behavior. Third, this shift compromises epistemic accuracy, independent reasoning, or appropriate correction. We also introduce a taxonomy for classifying sycophancy, consisting of alignment targets, mechanisms, and severity. The paper concludes by discussing implications for alignment evaluation and argues for boundary-aware assessment, structured rubrics, and mitigation strategies, while situating these proposals alongside alternative views of sycophancy.

Jiechen Li, Catherine A. Barry, Rishika Randev +3

Duke University Durham, NC 27708 · ∗These authors contributed equally.

Sycophancy Epistemic Integrity Pluralistic Alignment Credibility Student Misconceptions Assistant

May 3, 2026cs.AI

Personalized Digital Health Modeling with Adaptive Support Users

Personalized models are essential in digital health because individuals exhibit substantial physiological and behavioral heterogeneity. Yet personalization is limited by scarce and noisy user-specific data. Most existing methods rely on population pretraining or data from similar users only, which can lead to biased transfer and weak generalization. We propose a unified personalization framework that trains a personal model using adaptively weighted support users, including both similar and dissimilar individuals. The objective integrates personal loss, similarity-weighted transfer from similar users, and contrastive regularization from dissimilar users to suppress misleading correlations. An iterative optimization algorithm jointly updates model parameters and user similarity weights. Experiments on six tasks across four real-world digital health datasets show consistent improvements over population and personalized baselines. The method achieves up to 10% lower RMSE on large-scale datasets and approximately 25% lower RMSE in low-data settings. The learned adaptive weights improve data efficiency and provide interpretable guidance for targeted data selection.

Zhongqi Yang, Mahkameh Rasouli, Neda Mohseni +3

Department of Computer Science, University of California, Irvine · Sue & Bill Gross School of Nursing, University of California, Irvine · Thrive AI Health +1

Healthcare Parameter-Efficient Adaptation Model Training Similarity and Metrics Assistant

May 2, 2026cs.RO

Assistance Without Interruption: A Benchmark and LLM-based Framework for Non-Intrusive Human-Robot Assistance

Human-robot interaction (HRI) has long studied how agents and people coordinate to achieve shared goals. In this work, we formalize and benchmark the non-intrusive assistance as an independent paradigm of HRI, where a robot proactively supports a human's ongoing multi-step activities while strictly avoiding interruptions. Unlike conventional HRI tasks that rely on direct commands, explicit negotiation, or proactive interventions based on user habits and history, our task treats the human's plan as the primary process and formulates assistance as a joint decision over when to act and what to do. To systematically evaluate this problem, we establish a simulation benchmark, NIABench, along with new metrics tailored to the non-intrusive assistance task. We further propose a hybrid architecture that integrates an LLM with a scoring model. The scoring model first applies semantic retrieval to prune large candidate action sets, and then a ranker evaluates human-step and robot-action pairs, enabling reasoning over timing and cross-step dependencies. Comprehensive experiments on both NIABench and real-world scenarios demonstrate that our method achieves proactive, non-intrusive assistance that reduces human effort while preserving task effectiveness.

Yuedi Zhang, Shuanghao Bai, Wanqi Zhou +4

Institute of Artificial Intelligence and Robotics, Xi’an Jiaotong University, Xi’an, China · School of Electrical Engineering, Xi’an University of Technology, Xi’an, China

Human-Robot Interaction Assistant Negotiation

Apr 30, 2026cs.CL

Helpfulness Hurts: Domain-Dependent Degradation of Mid-Trained Compassion Values Under Post-Training

Standard post-training pipelines apply supervised fine-tuning (SFT) and reinforcement learning (RL) to make language models helpful, but these processes may inadvertently degrade values instilled during pre-training. We investigate whether the domain of post-training data differentially affects the retention of animal compassion values in a Llama 3.1 8B model mid-trained on compassion-oriented synthetic data, using both SFT (helpfulness via Dolly-15k vs. coding via Magicoder-110K) and GRPO (helpfulness via RLHFlow vs. coding via Magicoder), evaluated on the ANIMA 2.2 benchmark and MORU benchmark (Moral Reasoning Under Uncertainty). Helpfulness training significantly degrades animal compassion relative to coding training on ANIMA (SFT: 35.7% vs. 65.2%; GRPO: 18.7% vs. 32.0%), replicating across two independent helpfulness datasets and two training paradigms. On English MORU items, helpfulness training degrades general moral reasoning by 25.5 percentage points (46.4% vs. 71.9%), a striking gap that rivals the compassion effect in magnitude. However, this effect does not transfer cross-lingually: on the multilingual MORU benchmark, the domain effect disappears (SFT: 52.3% vs. 51.2%). In contrast, the animal compassion effect transfers consistently across languages, with Magicoder's ANIMA percentage-point gain over the base model 4.5 times larger on non-English items than English items. This divergence suggests that values instilled through mid-training are encoded more deeply and cross-lingually than reasoning improvements from domain-specific post-training. These results suggest that, for labs building on value-laden mid-training, coding-domain post-training may better preserve mid-trained values than helpfulness post-training without harming general reasoning capabilities.

Jasmine Brazilek, Juliana Seawell

†Compassion Aligned Machine Learning (CaML)

Post-Training Pretraining Empathy Value Assistant Degradation+2

Apr 24, 2026cs.CL

SHAPE: Unifying Safety, Helpfulness and Pedagogy for Educational LLMs

Large Language Models (LLMs) have been widely explored in educational scenarios. We identify a critical vulnerability in current educational LLMs, pedagogical jailbreaks, where students use answer-inducing prompts to elicit solutions rather than scaffolded instructions. To enable systematic study, we unify and formalize safe, helpful, and pedagogical behaviors with a knowledge-mastery graph and introduce SHAPE, a benchmark of 9,087 student-question pairs for evaluating tutoring behavior under adversarial pressure. We propose a graph-augmented tutoring pipeline that infers prerequisite concepts from queries, identifies mastery gaps, and routes generation between instructing and problem-solving via explicit gating. Experiments across multiple LLMs show that our method yields significantly improved safety under two pedagogical jailbreak settings, while maintaining near-ceiling helpfulness under the same evaluation protocol. Our code and data are available at https://github.com/MAPS-research/SHaPE

Sihang Zhao, Kangrui Yu, Youliang Yuan +2

Center for Data Science, New York University Shanghai · Courant Institute of Mathematical Sciences, New York University · School of Data Science, The Chinese University of Hong Kong, Shenzhen

Tutors Large Language Model Safety Prompt Engineering Jailbreak Success Rates Assistant

Apr 20, 2026cs.CL

Learning to Seek Help: Dynamic Collaboration Between Small and Large Language Models

Large language models (LLMs) offer strong capabilities but raise cost and privacy concerns, whereas small language models (SLMs) facilitate efficient and private local inference yet suffer from limited capacity. To synergize the complementary strengths, we introduce a dynamic collaboration framework, where an SLM learns to proactively decide how to request an LLM during multi-step reasoning, while the LLM provides adaptive feedback instead of acting as a passive tool. We further systematically investigate how collaboration strategies are shaped by SLM and LLM capabilities as well as efficiency and privacy constraints. Evaluation results reveal a distinct scaling effect: stronger SLMs become more self-reliant, while stronger LLMs enable fewer and more informative interactions. In addition, the learned dynamic collaboration strategies significantly outperform static pipelines and standalone inference, and transfer robustly to unseen LLMs.

Hang Zeng, Xiangyu Liu, Yong Hu +5

Shanghai Jiao Tong University, Shanghai, China · WeChat Tencent, Beijing, China · State University of New York at Buffalo, New York, United States.

Collaboration Assistant

Jan 17, 2026cs.CL

PEARL: Self-Evolving Assistant for Time Management with Reinforcement Learning

Overlapping calendar invitations force busy professionals to repeatedly decide which meetings to attend, reschedule, or decline. We refer to this preference-driven decision process as calendar conflict resolution. Automating this decision process is crucial yet challenging. Scheduling logistics can drain hours, and human delegation often fails at scale, which motivates us to ask: Can we trust large language models (LLMs) or language agents to manage time? To enable a systematic study of this question, we introduce CalConflictBench, a benchmark for long-horizon calendar conflict resolution. In CalConflictBench, conflicts are presented to agents round-by-round over a calendar year, requiring them to infer and adapt to user preferences progressively. Our experiments show that current LLM agents perform poorly with high error rates, e.g., Qwen-3-30B-Think has an average error rate of 35%. To address this gap, we propose PEARL, a reinforcement-learning framework that (i) augments the language agent with an external preference memory that stores and updates inferred strategies (e.g., attendee priorities, topic importance, time/location preferences), and (ii) optimizes the agent with round-wise rewards that directly supervise decision correctness, ranking quality, and memory usage across rounds. Experiments on CalConflictBench show that PEARL achieves an error reduction rate of 0.76 and a 55% improvement in average error rate compared to the strongest baseline.

Bingxuan Li, Jeonghwan Kim, Cheng Qian +4

University of Illinois at Urbana-Champaign · Viven

Schedule Assistant Offline Reinforcement Learning

Oct 23, 2025cs.CL

User Perceptions vs. Proxy LLM Judges: Privacy and Helpfulness in LLM Responses to Privacy-Sensitive Scenarios

Large language models (LLMs) are rapidly being adopted for tasks like drafting emails, summarizing meetings, and answering health questions. In these settings, users may need to share private information (e.g., contact details, health records). To evaluate LLMs' ability to identify and redact such information, prior work introduced real-life, scenario-based benchmarks (e.g., ConfAIde, PrivacyLens) and found that LLMs can leak private information in complex scenarios. However, these evaluations relied on proxy LLMs to judge the helpfulness and privacy-preservation quality of LLM responses, rather than directly measuring users' perceptions. To understand how users perceive the helpfulness and privacy-preservation quality of LLM responses to privacy-sensitive scenarios, we conducted a user study (

n=94

) using 90 PrivacyLens scenarios. We found that users had low agreement with each other when evaluating identical LLM responses. In contrast, five proxy LLMs reached high agreement, yet each proxy LLM had low correlation with users' evaluations. These results indicate that proxy LLMs cannot accurately estimate users' wide range of perceptions of utility and privacy in privacy-sensitive scenarios. We discuss the need for more user-centered studies to measure LLMs' ability to help users while preserving privacy, and for improving alignment between LLMs and users in estimating perceived privacy and utility.

Xiaoyuan Wu, Roshni Kaushik, Wenkai Li +2

Fujitsu Research of America, Inc. · Carnegie Mellon University

Privacy Large Language Model Responses Large Language Model Judges Assistant

Date pendingcs.LG

Safe Learning Under Irreversible Dynamics via Asking for Help

Most learning algorithms with formal regret guarantees essentially rely on trying all possible behaviors, which is problematic when some errors cannot be recovered from. Instead, we allow the learning agent to ask for help from a mentor and to transfer knowledge between similar states. We show that this combination enables the agent to learn both safely and effectively. Under standard online learning assumptions, we provide an algorithm whose regret and number of mentor queries are both sublinear in the time horizon for Markov decision processes with irreversible dynamics and infinite state spaces. Our proof involves a sequence of three reductions, making our result more general than a single algorithm. Conceptually, our result may be the first formal proof that it is possible for an agent to obtain high reward while becoming self-sufficient in an unknown, unbounded, and high-stakes environment without resets.

Benjamin Plaut, Juan Liévano-Karim, Hanlin Zhu +1

Division of Computer Science University of California Berkeley, CA 94720-1776, USA

Markov Decision Processes Irreversibility State Space Assistant Tutors