Human Annotations

Accurate translation from Natural Language to First-Order Logic (NL-to-FOL) underpins neurosymbolic AI systems and Natural Language Inference (NLI), making the quality of NL-to-FOL benchmarks essential---yet these datasets have never been rigorously audited. Our first contribution is to present a systematic human inspection of the validation split of \textsf{FOLIO} and a subset of \textsf{MALLS} test instances, finding that approximately 42.5% and 42% of entries, respectively, contain incorrect FOL formalizations (i.e., ground truth labels), with additional rates of ambiguous NL sentences (17.8% and 51%) and incorrect NLI labels in \textsf{FOLIO} (8.4%). Our second contribution is to develop and release corrected ground truths for such datasets, showing that annotation errors distort model evaluation on a reference benchmark task: testing three state-of-the-art LLMs (Gemma~4 31B-it, Qwen3-30B-A3B, and GPT-4o-mini) with the corrected ground truths yields accuracy gains from +11 to +23 percentage points. Motivated by these findings, we propose an LLM-based framework to support humans in manual reviewing NL-to-FOL datasets. By directing reviewers toward the most error-prone instances, we empirically show that it is possible to achieve 90% dataset accuracy after reviewing fewer than 20% of instances, compared to over 76% required by unguided review. We release all human-verified annotations and the code for our framework.

Human annotation is the empirical foundation of much NLP research, from dataset construction to model evaluation, but papers often leave unclear who produced the annotations and how the annotation process was controlled. We provide the first large-scale, task-level audit of human annotation reporting across major NLP venues, asking which annotation details are documented, which are missing, and how reporting varies across time, topic, venue, and intended use of human judgment. We introduce a unified taxonomy of annotation-reporting practices and validate an LLM-assisted extraction pipeline against Annotated-gold, a human-adjudicated gold standard of 41 papers and 72 annotation tasks, where the best model reaches human-comparable agreement with adjudicated labels, with Krippendorff's alpha of 0.606 versus 0.585 for human-human agreement. Using this pipeline, we construct Annotated-llm, a dataset covering ACL-venue papers from 2018-2025, with 2,667 extracted annotation tasks from 1,603 papers, and find that papers frequently report operational details such as recruitment strategies, annotator expertise, and annotation volume, but often omit details needed to assess annotation validity, including training, language proficiency, compensation, socio-demographics, adjudication, and agreement values, especially in model-evaluation studies. Our results show that annotation reporting in NLP has improved over time but remains uneven, and they establish a scalable framework and bare-minimum reporting recommendations for making human annotation more reliable, reproducible, and interpretable.

Text-to-video generation has advanced rapidly in visual quality, but remains under-evaluated for factuality and practical usefulness. We introduce knowledge-intensive video generation (KIVI), where models generate videos from short information-seeking prompts that ask for explanations, procedures, or demonstrations. To evaluate this setting, we construct KIVI-Bench, a benchmark of 1,080 prompts, and propose automatic metrics for factuality and helpfulness. Human evaluation shows that our metrics significantly better align with human annotations than existing alternatives. Experiments on seven state-of-the-art video generation models show that current systems still lag behind human performance, especially on visual properties, procedural operations, and clear information presentation. These results highlight KIVI as a challenging direction for factual and instructionally useful video generation.

Safe global deployment of AI models requires alignment with human values that vary across cultures. Yet rater pools in safety evaluation datasets remain largely geographically homogeneous, failing to capture geo-cultural differences. Further, it remains unclear whether such differences persist after controlling for demographics such as age, gender, and ethnicity. Through a meta-analysis of safety datasets, we find that most do not report geo-cultural information, and those that do lack a unified methodology to jointly analyze geo-cultural and demographic correlates. Using the Inglehart-Welzel dimensions of cross-cultural variation, we demonstrate via multilevel modeling that cultural zone membership explains variance in safety ratings beyond standard demographics (p<0.05 across 6 datasets). Moreover, our analysis indicates that roughly 10% of items in the datasets we examined are culturally sensitive: likely to be misclassified as safe without adequate cultural representation. We evaluate LLMs as both rater surrogates and triage tools, finding that current LLMs do not reliably stand in for raters, though they can help prioritize culturally sensitive items for human annotation. Our findings motivate more culturally pluralistic safety evaluation and offer practical takeaways to support it.

LLM-as-a-Judge is a scalable alternative to human evaluation, yet existing rubric-based methods rely on human-annotated data such as reference answers or expert-crafted rubrics. We propose to automatically generate fine-grained evaluation rubrics without any human annotation. Our training-free method generates rubrics at dataset-specific and instance-specific granularities, achieving performance competitive with existing methods across four benchmarks. We further present a method that iteratively fine-tunes a rubric generator model via meta-judge reward signals. The fine-tuned generator outperforms all existing baselines in both pairwise and pointwise evaluation. Notably, a fine-tuned 14B rubric generator outperforms a much larger proprietary model at rubric generation, showing the effectiveness of our fine-tuning strategy.

Histological scoring is essential for diagnosing Non-Alcoholic Fatty Liver Disease (NAFLD), yet its automation remains challenging due to the high annotation cost and negative transfer among the strongly correlated NAFLD Activity Score (NAS) indicators in multi-task learning. To address this issue, we propose a subspace-decoupled multi-task Vision Transformer (ViT) that integrates lightweight task-specific Adapters with orthogonality-based constraints. This design constructs independent feature subspaces for steatosis, ballooning, and inflammation, effectively reducing task interference while retaining shared representations. We further construct a curated multi-task mouse NAFLD histology dataset with expert annotations for all NAS components. Experimental results demonstrate that the proposed method improves multi-task stability and generalization with substantially reduced computational cost compared to training separate single-task models. The code and the curated dataset have been prepared and will be made publicly available upon acceptance to support reproducibility.

User satisfaction with AI assistants is highly personalized: the same response may satisfy one user but disappoint another depending on what each user expects and what they have asked for before. Existing automatic evaluation methods mostly measure generic response quality, making it difficult to judge whether a response satisfies a user at a specific turn. We study this problem as personalized turn-level user conversation satisfaction evaluation. We build a conversation satisfaction evaluator that combines compact user memories with target-turn context to produce satisfaction scores and dissatisfaction-oriented rationales. Meta-evaluation against human satisfaction annotations shows that personalized memory and post-hoc score calibration improve ordinal agreement and dissatisfied-turn detection over supervised, retrieval-based, and generic LLM-as-a-judge baselines. We further introduce PersTurnBench, a personalized turn-level user conversation satisfaction benchmark that uses the verified evaluator to assess generation models via replay. By holding the replay state fixed, PersTurnBench enables controlled comparison of generic generation models and memory-augmented personalized systems without new human labels for every candidate model. The evaluator and benchmark let researchers compare candidate generation models on personalized satisfaction without collecting new user feedback for every model.

This paper presents a multilingual customer service self-help corpus comprising 1,122 manually validated documents in Finnish, Danish, Norwegian, and Swedish, totaling 274,599 words and 1,884,833 characters. The documents have been sourced from the public self-help pages of four Nordic telecommunications operators and subsequently filtered for person-identifiable information and relevance through a combined LLM and human annotation pipeline. Domain-specific datasets for Nordic languages remain scarce, particularly in customer service: a domain of growing importance for retrieval-augmented generation, cross-lingual transfer learning, and emerging agent-based service architectures. An analysis of the corpus reveals substantial variation in document length and structure across operators, reflecting distinct editorial strategies, as well as broad topical coverage spanning network hardware, mobile services, TV and streaming, billing, and account management. The dataset is publicly available under a CC-BY-NC-SA-4.0 license at https://zenodo.org/records/20732652, intended to support reproducible research in Nordic NLP and information retrieval.

Blood vessel segmentation is a core task in medical image analysis for the care of vascular diseases and surgical planning, yet the challenges of providing expert vascular annotations pose a major obstacle for the progress of related deep learning techniques. To address this, we propose VesselSim, a two-stage framework for universal 3D blood vessel segmentation that eliminates the need for real annotated data during training. First, we introduce a stochastic, geometry-driven vascular simulation framework that models recursive branching, curvature-controlled growth, and collision-aware topology, followed by domain-randomized intensity synthesis to generate 16,500 anatomically plausible 3D angiographic volumes. Second, a 3D U-Net is trained solely on this synthetic data. To bridge the domain gap from synthetic to real images at inference time, we introduce a test-time adaptation strategy via a self-supervised mask reconstruction decoder, enabling adaptation to unseen clinical scans without prior domain knowledge. We evaluate VesselSim in a zero-shot setting on multiple real-world datasets spanning MR and CT across several anatomical regions, including the brain and kidneys. Despite being trained exclusively on synthetic data, VesselSim achieves performance competitive with state-of-the-art vascular segmentation foundation models. These findings suggest that learning vessel geometry from synthetic tubular structures is effective for robust cross-domain generalization, substantially reducing the reliance on acquired medical imaging data and more importantly, expert annotations.

As interactive LLM-based applications are created and refined, model developers need to evaluate the quality of generated text along many possible axes. For simpler systems, human evaluation may be practical, but in complicated systems like conversational chatbots, the amount of generated text can overwhelm human annotation resources. Model developers have begun to rely heavily on auto-evaluation, where LLMs are also used to judge generation quality. However, existing LLM-as-a-judge benchmarks largely focus on simple Q&A tasks that do not match the complexity of multi-turn conversations. We introduce RankJudge, a benchmark generator for evaluating LLM-as-a-judge on multi-turn conversations grounded in reference documents. RankJudge creates pairs of conversations where one conversation has a single flaw injected into one turn. This construction allows paired conversations to be labeled unambiguously as better or worse, and precisely isolates failure categories to individual turns, enabling a strict joint correctness criterion for judging. We implement RankJudge across the domains of machine learning, biomedicine, and finance, evaluate 21 frontier LLM judges, and rank those judges via the Bradley-Terry model. Our formulation also allows ranking each conversation pair with difficulty ratings, which we use to dynamically curate the evaluation slice to reduce label noise, as confirmed via human annotation. We find that judge rankings are stable under partial observability, coarser correctness criteria, and an alternative random-walk rating algorithm.

Academic researchers need efficient and reliable methods for collecting high-quality information from trusted sources, but modern tools for AI-assisted research still suffer from the tendency of Large Language Models (LLMs) to produce factually inaccurate or nonsensical output, commonly referred to as hallucinations. We apply the extractive question answering system VerbatimRAG to research papers in the ACL Anthology, directly mapping user queries to verbatim text spans in retrieved documents. We contribute a novel ground truth dataset for the task of mapping user queries to relevant text spans in research papers, and use it to train and evaluate a variety of extractive models. Human annotation is performed by NLP researchers and is based on synthetic user queries generated using a custom pipeline based on the ScIRGen methodology, paired with chunks of research papers retrieved by VerbatimRAG. On this benchmark, a 150M-parameter ModernBERT token classifier trained on silver supervision from our pipeline achieves the best word-level F1 (53.6), ahead of the strongest evaluated LLM extractor (48.7).

Deep-learning pipelines for microscopy image classification often require expensive, labor- and time-intensive expert annotation to produce high-quality ground truth for training. Recent work has shown that prompt tuning of vision-language models (VLMs) can reduce manual annotation by constructing a small prompt set of expert-verified image-caption exemplars that is reused as few-shot context to classify all remaining images at inference time. To further reduce effort, the VLM can draft captions for candidate exemplars, which experts then verify and lightly edit instead of writing text de novo. However, two practical questions remain unaddressed: (1) which unlabeled images should be prioritized for verification, and (2) how many verified exemplars are needed to reach a performance target. In this work, we address these questions by formulating prompt-set construction as a target-driven active learning problem that prioritizes which images to annotate. We study three complementary selection criteria under strict low-resource constraints with small unlabeled pools. Experiments show that our methods reach the target performance with substantially fewer expert-verified images than random selection, achieving 100% test accuracy with as few as 20 annotated images on average. More broadly, our human-in-the-loop framework demonstrates a human-centered use of generative AI in biomedical image analysis, where experts remain actively involved in verifying and refining model output while significantly reducing annotation cost. Code and data will be publicly available.

Visual planning asks a model to generate the remaining steps of a procedure in natural language given a partial video context and a goal. Progress on this task is bottlenecked by annotation: clean labeled datasets are small, domain-narrow, and encode a single execution trajectory per example, even though many valid orderings exist. Large-scale instructional video corpora offer orders of magnitude more procedural content, but supervised fine-tuning on pseudo-labels from their noisy ASR narrations propagates segmentation and alignment errors and stays single-trajectory. We identify a key asymmetry: extracting clean step labels from noisy video is hard, but verifying whether a generated step sequence is temporally grounded in ASR transcripts is cheap and scales to millions of videos via precomputed text embeddings. We exploit this asymmetry in RECIPE, which uses grounding quality as a reward for GRPO, turning the noisy corpus into a verifier rather than a label source. The framework applies uniformly to two planner input configurations (Socratic, with a textual history extracted by a frozen VLM, and Video, consuming video tokens directly) and to annotated and weakly supervised regimes. We evaluate on 7 procedural benchmarks using a reference-based LLM-as-judge protocol scoring plans across 6 procedural criteria. RECIPE-RL improves over the base checkpoint at all scales (0.5B, 3B, 7B) and every benchmark, with macro-accuracy gains of +7 to +8 points in-domain and up to +16 points zero-shot. It outperforms supervised fine-tuning on both annotated and pseudo-labeled plans (the latter degrades the base) and remains robust without human annotations. Used as the proposal stage of a prior propose-assess-search planner, it improves over the strongest zero-shot baseline at every horizon on Visual Planning for Assistance, and on COIN it preserves the generation diversity that SFT collapses.

Pairwise preferences and pointwise ratings are the two dominant annotation protocols in image aesthetic assessment (IAA), yet existing benchmarks adopt only one, leaving their complementarity unmeasured under controlled conditions. We introduce PPaint, a matched dual-protocol benchmark in which 15 domain experts, 5 per category, annotate 150 Chinese paintings under both protocols across five aesthetic dimensions, collecting 45,900 pairwise expert judgments through a locally dense preference design alongside the matched ratings. The matched design reveals complementary strengths: preferences yield more consistent ordinal rankings, while ratings anchor the absolute score scale. Fusing both signals via two independent preference-to-score methods yields a fused expert ground truth on which the two constructions converge to nearly identical scores. The same preference-to-score principle extends to label-free VLM training. PSDistill converts VLM pairwise judgments into calibrated pseudo-scores via an Elo reference pool, and trains the same VLM with confidence-weighted ranking optimization to produce a single-pass aesthetic scorer. Trained on a single painting category, the distilled Qwen3-VL-8B improves mean SRCC from 0.504 to 0.709 across all three categories, outperforming all open-source baselines including the dedicated aesthetic model ArtiMuse and matching closed-source Gemini-3.1-Pro within 0.04 SRCC at single-pass inference cost, with cross-domain transfer further validated on APDDv2. We will release the full PPaint dataset and training code.

Abstract reasoning ability reflects the intelligence and generalization capacity of LLMs to extract and apply abstract rules. However, accurately measuring this ability remains challenging: existing benchmarks either rely on expensive manual annotation, limiting their scale, or risk measuring memorization rather than genuine reasoning. To address this, we introduce an automated pipeline named A2RBench, encompassing generation, expansion, evaluation, and analysis. Specifically, in the generation stage, LLMs create diverse tasks demanding genuine reasoning; in the expansion stage, LLMs reuse validated rules and expand new input spaces to generate task variations, achieving scaling. However, such a process may cause hallucinations. To eliminate it, we further establish a theoretical framework and prove that programmatic verification--testing whether the inverse operation perfectly reverses the forward operation (cycle consistency)--guarantees a unique solution. Through extensive evaluations on mainstream LLMs, we find: (1) Current LLMs exhibit fundamental deficiencies in abstract reasoning, with top models significantly underperforming humans on a representative subset (39.8% vs. 68.5%). (2) Current LLMs fall far short of 2D and 1D in the complexity of generated 3D tasks, revealing their lack of understanding of high-dimensional tasks. (3) Counterintuitively, inputs with higher information complexity can simplify the reasoning process.

Linguistic annotation of transcribed speech is essential for research in language acquisition, language disorders, and sociolinguistics, yet remains labor-intensive and time-consuming. While Large Language Models (LLMs) have shown promise in automating annotation tasks, their ability to handle complex discourse-level annotation in non-English languages remains understudied. This study evaluates whether LLMs can reliably annotate narrative macrostructure-the hierarchical organization of story grammar elements-in spoken Mandarin, using the Multilingual Assessment Instrument for Narratives (MAIN) as a testbed. We compared four LLMs against trained human annotators on narratives produced by children, young adults, and older adults. The best-performing model achieved agreement with human raters (k=.794) approaching human-human reliability levels (k=.872) while reducing annotation time by 65%, whereas the locally deployable lightweight model performed substantially worse. Annotation difficulty varied systematically by macrostructure element type, with categories requiring subtle semantic differentiation posing persistent challenges. Furthermore, model reliability decreased on young adult narratives, which exhibited greater lexical variation, semantic ambiguity, and multi-element integration within single utterances. These findings suggest that LLMs can effectively support discourse-level annotation in non-English spoken corpora, while highlighting the continued need for human oversight in semantically complex tasks. Our prompt templates are open sourced for future use.

Reinforcement learning post-training has substantially improved the reasoning accuracy of vision-language models, yet the resulting policies remain poorly calibrated. Terminal correctness rewards provide no gradient that penalizes confident errors more than uncertain ones and no signal that ties confidence to the quality of visual evidence, a gap that becomes especially severe under corrupted or ambiguous inputs where models continue to report high confidence on incorrect answers. We introduce Ranking-Aware Calibration (RAC), a training-time framework that supervises confidence using two comparison signals that group-based RL already produces at no additional labeling cost. The ranking-aware group loss enforces that a better rollout receives higher confidence than a worse one within the same prompt. The clean--corrupted pairwise loss enforces that confidence attenuates as visual evidence degrades. Because the ranking signal forces the policy to distinguish between correct and incorrect reasoning paths, it also reinforces task accuracy beyond what correctness rewards alone produce. Both losses require no external confidence annotations and integrate naturally with group-based RL post-training. We instantiate RAC on Qwen2.5-VL and InternVL-3.5 backbones and evaluate on six multimodal reasoning benchmarks under clean and corrupted inputs. Empirical results show that the ranking-aware loss substantially improves task accuracy by teaching the policy to discriminate between better and worse reasoning, while the pairwise corruption loss reduces calibration error under degraded inputs. Their combination achieves the best calibration across all tested backbones while improving accuracy in the majority of settings.

Modern image super-resolution methods generate detailed, visually appealing results, but they often introduce visual artifacts: unnatural patterns and texture distortions that degrade perceived quality. These defects vary widely in perceptual impact--some are barely noticeable, while others are highly disturbing--yet existing detection methods treat them equally. We propose artifact prominence as an evaluative target, defined as the fraction of viewers who judge a highlighted region to contain a noticeable artifact. We design a crowdsourced annotation protocol and construct SR-Prominence, a dataset suite containing 3,935 artifact masks from DeSRA, Open Images, Urban100, and a realistic no-ground-truth Urban100-HR setting, annotated with prominence. Re-annotating DeSRA reveals that 48.2% of its in-lab binary artifacts are not noticed by a majority of viewers. Across the suite, we audit SR artifact detectors, image-quality metrics, and SR methods. We find that classical full-reference metrics, especially SSIM and DISTS, provide surprisingly strong localized prominence signals, whereas no-reference IQA methods and specialized artifact detectors often fail to generalize across datasets and reference settings. SR-Prominence is released with an objective scoring protocol that allows new metrics to be benchmarked on our suite without further crowdsourcing. Together, the data and protocols enable SR artifact evaluation to move from binary defect presence toward perceptual impact. SR-Prominence is available at https://huggingface.co/datasets/imolodetskikh/sr-artifact-prominence.

Piano fingering shapes how a passage can be played, yet it is difficult to label after a performance. An annotator must decide which finger produced each note while reconciling the score, timing, video, and hand motion. We present PiAnnotate, a web-based pipeline for adding expert fingering annotations to the FurElise performance dataset. The tool brings together a piano-roll view, performance video, and a 3D MANO hand mesh so that reviewers can inspect each assignment in musical and physical context. Rather than storing only the final answer, PiAnnotate keeps paired rule-based and human-edited fingering tracks. These paired tracks make the annotation history auditable by showing where a geometric rule was sufficient, where experts intervened, and how labels changed across review passes. As a final diagnostic, we train a small Transformer probe on the paired tracks. The probe improves on the rule baseline on held-out pieces while remaining conservative about changing labels that were already correct, suggesting that the edited labels contain learnable structure rather than only isolated fixes.

This article investigates the performance of automatic evaluation metrics (AEMs) and LLM-as-a-judge evaluation on literary translation across multiple languages, genres, and translation modalities. The aim is to assess how well these tools align with professionals when evaluating translation, creativity (creative shifts & errors), and see if they can substitute laborious manual annotations. A dataset of literary translations across three modalities (human translation, machine translation, and post-editing), three genres and three language pairs was created and annotated in detail for creativity by experienced professional literary translators. The results show that both AEMs and LLM-as-a-judge evaluations correlate poorly with professional evaluations on creativity, with LLM-as-a-judge showing a systematic bias in favour of machine-translated texts and penalising creative and culturally appropriate solutions. Moreover, performance is consistently worse for more literary genres such as poetry. This highlights fundamental limitations of current automatic evaluation tools for literary translation and the need to create new tools that do not frequently consider out of routine translations as errors.

Sign languages are natural, visual-gestural languages used by Deaf communities worldwide. Over 300 distinct sign languages remain severely low-resource due to limited documentation, sparse datasets, and insufficient computational tools. This systematic review synthesizes literature on sign language recognition and translation for under-resourced languages, using Azerbaijan Sign Language (AzSL) as a case study. Analysis of global initiatives extracts eight actionable lessons, including community co-design, dialectal diversity capture, and privacy-preserving pose-based representations. Turkic sign languages (Kazakh, Turkish, Azerbaijani) receive special attention, as linguistic proximity enables effective transfer learning. We propose three paradigm shifts: from architecture-centric to data-centric AI, from signer-independent to signer-adaptive systems, and from reference-based to task-specific evaluation metrics. A technical roadmap for AzSL leverages lightweight MediaPipe-based architectures, community-validated annotations, and offline-first deployment. Progress requires sustained interdisciplinary collaboration centered on Deaf communities to ensure cultural authenticity, ethical governance, and practical communication benefit.

Hallucination, broadly referring to unfaithful, fabricated, or inconsistent content generated by LLMs, has wide-ranging implications. Therefore, a large body of effort has been devoted to detecting LLM hallucinations, as well as designing benchmark datasets for evaluating these detectors. In this work, we first establish a desiderata of properties for hallucination detection benchmarks (HDBs) to exhibit for effective evaluation. A critical look at existing HDBs through the lens of our desiderata reveals that none of them exhibits all the properties. We identify two largest gaps: (1) RAG-based grounded benchmarks with long context are severely lacking (partly because length impedes human annotation); and (2) Existing benchmarks do not make available realistic label noise for stress-testing detectors although real-world use-cases often grapple with label noise due to human or automated/weak annotation. To close these gaps, we build and open-source a new RAG-based HDB called T RIVIA+ that underwent a rigorous human annotation process. Notably, our benchmark exhibits all desirable properties including (1) T RIVIA+ contains samples with the longest context in the literature; and (2) we design and share four sets of noisy labels with different, both sample-dependent and sampleindependent, noise schemes. Finally, we perform experiments on RAG-based HDBs, including our T RIVIA+, using popular SOTA detectors that reveal new insights: (i) ample room remains for current detectors to reach the performance ceiling on RAG-based HDBs, (ii) the basic LLM-as-a-Judge baseline performs competitively, and (iii) label noise hinders detection performance. We expect that our findings, along with our proposed benchmark 1 , will motivate and foster needed research on hallucination detection for RAG-based tasks.

Generative world models are increasingly used for video generation, where learned simulators are expected to capture the physical rules that govern real-world dynamics. However, evaluating whether generated videos actually follow these rules remains challenging. Existing physics-focused video benchmarks have made important progress, but they still face three key challenges, including the coarse evaluation frameworks that hide law-specific failures, response biases and fatigue that undermine the validity of annotation judgments, and automated evaluators that are insufficiently physics-aware or difficult to audit. To address those challenges, we introduce PhyGround, a criteria-grounded benchmark for evaluating physical reasoning in video generation. The benchmark contains 250 curated prompts, each augmented with an expected physical outcome, and a taxonomy of 13 physical laws across solid-body mechanics, fluid dynamics, and optics. Each law is operationalized through observable sub-questions to enable per-law diagnostics. We evaluate eight modern video generation models through a large-scale, quality-controlled human study, grounded on social science lab experiment design. A total of 459 annotators provided 5,796 complete annotations and over 37.4K fine-grained labels; after quality control, the retained annotations exhibited high split-half model-ranking correlations (Spearman's rho > 0.90). To support reproducible automated evaluation, we release PhyJudge-9B, an open physics-specialized VLM judge. PhyJudge-9B achieves substantially lower aggregate relative bias than Gemini-3.1-Pro (3.3% vs. 16.6%). We release prompts, human annotations, model checkpoints, and evaluation code on the project page https://phyground.github.io/.

Hallucination remains a persistent challenge in Large Language Models (LLMs), particularly in context-grounded settings such as RAG and agentic AI systems. This study focuses on contextual hallucination detection in summarization tasks. We analyze the QAGS-C and SummEval datasets by comparing original benchmark annotations with reason and span-based predictions from Gemini 2.5 Flash and GPT-5 Mini. To address systematic divergences between human labels and LLM judgments, we re-evaluated all conflicted samples through a human adjudication process involving 2 cross-cultural adjudicators. Following this re-evaluation, triple agreement (between human, GPT, and Gemini) increased by 6.38% for QAGS-C and 7.62% for SummEval. Similarly, model accuracy improved, with GPT increasing by 4.25% on QAGS-C and 2.34% on SummEval, while Gemini showed gains of 8.51% and 3.80%, respectively. Notably, adjudicators frequently sided with the models' judgments over original human annotations when LLMs provided explicit reasoning. Overall human adjudicator agreement ranged between 83% and 87%. These findings suggest that for ambiguity-prone tasks, single-pass annotations may be insufficient, and model-assisted re-evaluation yields more reliable benchmarks.

Despite the rapid development of AI reviewers, evaluating such systems remains challenging: metrics favor overlap with human reviews over correctness. However, since human reviews often cover only a subset of salient issues and sometimes contain mistakes, they are unreliable as gold references. To address this, we build category-specific benchmark subsets and skip evaluation when the corresponding human reviews are missing to strengthen Completeness. We also leverage reviewer--author--meta-review discussions as expert annotations and filter unreliable reviews accordingly to strengthen Correctness. Finally, we introduce CoCoReviewBench, which curates 3,900 papers from ICLR and NeurIPS to enable reliable and fine-grained evaluation of AI reviewers. Analysis shows that AI reviewers remain limited in correctness and are prone to hallucinations, and highlights reasoning models as more effective reviewers, motivating further directions for improving AI reviewers. Benchmarks and models are available at https://github.com/hexuandeng/CoCoReviewBench.

Automated privacy audits of web and mobile applications often analyse outbound HTTP traffic to detect Personally Identifiable Information (PII) leakage. However, existing learning-based detectors typically depend on scarce, manually labelled traffic and are tightly coupled to fixed label taxonomies, limiting transferability across domains and evolving definitions of PII. This paper investigates whether Large Language Models (LLMs) can support taxonomy-agnostic annotation of explicitly transmitted PII values in HTTP message bodies when the taxonomy is provided at runtime. We introduce a multi-stage LLM-based pipeline that combines deterministic pre-processing with label-level classification, targeted instance-level value annotation, and output validation. To enable controlled evaluation and exemplar-based prompting without relying on sensitive real-user captures, we further propose an LLM-based generator for synthetic HTTP traffic with manually validated, taxonomy-derived PII annotations. We evaluate the approach across three taxonomies spanning different PII domains and granularity levels. Results show that the pipeline accurately detects PII types and extracts corresponding values for concrete PII taxonomies. Overall, our findings position LLMs as a promising foundation for flexible, taxonomy-agnostic traffic annotation and for creating labelled data under evolving privacy taxonomies.

Self-evolving search agents reduce reliance on human-written training questions by generating and solving their own search tasks. We build on Search Self-Play (SSP), a representative Proposer and Solver framework in which questions are generated and answered via multi-step search and reasoning. In practice, however, SSP faces two bottlenecks: the Proposer constructs questions from isolated answer entities without relational context, yielding many invalid or unverifiable questions in early self-play training, while the Solver receives only a binary outcome reward that discards useful signal from partially on-track search trajectories. We address both bottlenecks by reusing knowledge-graph paths as construction-derived intermediate supervision for both question construction and reward shaping. First, we ground question construction in LLM-guided knowledge-graph subgraphs, providing relational context for the Proposer. Second, we observe that constructing and solving a multi-hop question can involve overlapping intermediate entities: the factual bridges used to formulate the question may provide approximate waypoints for answering it. Exploiting this overlap, we introduce Waypoint Coverage Reward (WCR), which grants graded partial credit to incorrect Solver trajectories according to their coverage of entities on the construction path, while preserving full reward for correct answers. Across seven QA benchmarks and nine model configurations, our approach improves the average score over standard SSP in all configurations, including notable gains on multi-hop QA tasks. These results suggest that knowledge-graph paths can be reused as lightweight intermediate supervision, providing both relational guidance and process feedback without additional task-specific human annotations or manually labeled process steps.

High-quality singing annotations are fundamental to modern Singing Voice Synthesis (SVS) systems. However, obtaining these annotations at scale through manual labeling is unrealistic due to the substantial labor and musical expertise required, making automatic annotation highly necessary. Despite their utility, current automatic transcription systems face significant challenges: they often rely on complex multi-stage pipelines, struggle to recover text-note alignments, and exhibit poor generalization to out-of-distribution (OOD) singing data. To alleviate these issues, we present VocalParse, a unified singing voice transcription (SVT) model built upon a Large Audio Language Model (LALM). Specifically, our novel contribution is to introduce an interleaved prompting formulation that jointly models lyrics, melody, and word-note correspondence, yielding a generated sequence that directly maps to a structured musical score. Furthermore, we propose a Chain-of-Thought (CoT) style prompting strategy, which decodes lyrics first as a semantic scaffold, significantly mitigating the context disruption problem while preserving the structural benefits of interleaved generation. Experiments demonstrate that VocalParse achieves state-of-the-art SVT performance on multiple singing datasets. The source code and checkpoint are available at https://github.com/pymaster17/VocalParse.

Dense temporal annotation of procedural activity videos is vital for action understanding and embodied intelligence but remains labor-intensive due to reactive tools. Each correction is treated as an isolated edit, limiting reuse of information on annotator uncertainty and model reliability. We introduce IMPACT-Scribe, a correction-driven framework for dense labeling that uses each correction to improve future human-machine collaboration. IMPACT-Scribe combines uncertainty-aware boundary scribble supervision, local proposal modeling, cost-aware query planning, structured propagation, and correction-driven adaptation. Experiments and a human study show that this closed-loop design improves labeling quality per effort, enhances boundary accuracy, and fosters better human-machine interaction over time. The code will be made publicly available at https://github.com/BanzQians/IMPACT_AS.

In this article, we introduce SiNFluD, a novel benchmark dataset for Sindhi figurative language classification. We first collect raw text from various blogs, social media platforms, and literary sources, and subsequently prepare the corpus for annotation. Two native annotators label the data using the Doccano text annotation tool, achieving an inter-annotator agreement of 0.81. We then establish baseline results using 5-fold and 10-fold cross-validation. Finally, we evaluate mBERT, XLM-RoBERTa, and XLM-RoBERTa-XL models, along with SetFit for few-shot fine-tuning of sentence transformers. Among these, the pretrained XLM-RoBERTa-XL achieves the best performance.

Video platforms such as YouTube have reshaped how users engage with entertainment and information, emphasizing brief, highly engaging content such as Shorts. Within this ecosystem, certain content occupies a gray area where it remains allowed but may still have unintended negative effects on some audiences. To study this problem, we introduce TwistedHumor, a dataset of 1,211 YouTube Shorts paired with 33,041 related comments, with hand annotations for humor presence, humor type, harm, topic, rhetorical devices, and stand up context. Beyond dataset creation, we present a multi view analysis of how humor and harm appear in short form social media. Using LLooM based concept induction over video descriptions, we find that dark humor frequently clusters around themes of critique, coping, awkwardness, and identity expression rather than appearing as a single uniform category. We further analyze audience response through linked comments and show that regular humor is associated with more positive sentiment, while dark humor receives more mixed, neutral, and sometimes more toxic reactions. Finally, we evaluate large language models against human annotations and find that they perform better on stand up comedy compared to shorter jokes. Together, these results position TwistedHumor not only as a new benchmark, but as an empirical study of the gray area between humor and harm in short form video, highlighting the need for context aware moderation and more robust multimodal evaluation.

Annotating long-horizon robotic demonstrations with precise temporal action boundaries is crucial for training and evaluating action segmentation and manipulation policy learning methods. Existing annotation tools, however, are often limited: they are designed primarily for vision-only data, do not natively support synchronized visualization of robot-specific time-series signals (e.g., gripper state or force/torque), or require substantial effort to adapt to different dataset formats. In this paper, we introduce ATLAS, an annotation tool tailored for long-horizon robotic action segmentation. ATLAS provides time-synchronized visualization of multi-modal robotic data, including multi-view video and proprioceptive signals, and supports annotation of action boundaries, action labels, and task outcomes. The tool natively handles widely used robotics dataset formats such as ROS bags and the Reinforcement Learning Dataset (RLDS) format, and provides direct support for specific datasets such as REASSEMBLE. ATLAS can be easily extended to new formats via a modular dataset abstraction layer. Its keyboard-centric interface minimizes annotation effort and improves efficiency. In experiments on a contact-rich assembly task, ATLAS reduced the average per-action annotation time by at least 6% compared to ELAN, while the inclusion of time-series data improved temporal alignment with expert annotations by more than 2.8% and decreased boundary error fivefold compared to vision-only annotation tools.

Diagram question answering (Diagram QA) requires reasoning-level attribution that links each question-answer pair to all visual regions needed to derive the answer, rather than only the region containing the final response. Creating such structured evidence across diagrams, charts, maps, circuits, and infographics is time-consuming, and existing annotation tools tightly couple their interfaces to dataset-specific formats. We present DIAGRAMS, a lightweight, schema-driven review framework that decouples interface logic from dataset-specific JSON structures through an internal meta-schema and dataset adapters. Given an image and QA pair with optional candidate regions, the system performs QA-conditioned evidence selection and proposes the regions required for reasoning. When QA pairs or candidate regions are missing, it generates them and supports human verification and refinement. Across six Diagram QA datasets, model-suggested evidence achieves 85.39% precision and 75.30% recall against reviewer-final selections (micro-averaged). These results indicate that the review-first framework reduces manual region creation while maintaining high agreement with final reasoning-level attributions. We release a public demo and installable package to support dataset auditing, grounded supervision creation, and grounded evaluation.

Human visual preferences are inherently multi-dimensional, encompassing aesthetics, detail fidelity, and semantic alignment. However, existing datasets provide only single, holistic annotations, resulting in severe label noise: images that excel in some dimensions but are deficient in others are simply marked as winner or loser. We theoretically demonstrate that compressing multi-dimensional preferences into binary labels generates conflicting gradient signals that misguide Diffusion Direct Preference Optimization (DPO). To address this, we propose Semi-DPO, a semi-supervised approach that treats consistent pairs as clean labeled data and conflicting ones as noisy unlabeled data. Our method starts by training on a consensus-filtered clean subset, then uses this model as an implicit classifier to generate pseudo-labels for the noisy set for iterative refinement. Experimental results demonstrate that Semi-DPO achieves state-of-the-art performance and significantly improves alignment with complex human preferences, without requiring additional human annotation or explicit reward models during training. We will release our code and models at: https://github.com/L-CodingSpace/semi-dpo

This report presents a comparative evaluation of DKnownAI Guard in AI agent security scenarios, benchmarked against three competing products: AWS Bedrock Guardrails, Azure Content Safety, and Lakera Guard. Using human annotation as the ground truth, we assess each guardrail's ability to detect two categories of risks: threats to the agent itself (e.g., instruction override, indirect injection, tool abuse) and requests intended to elicit harmful content (e.g., hate speech, pornography, violence). Evaluation results demonstrate that DKnownAI Guard achieves the highest recall rate at 96.5% and ranks first in true negative rate (TNR) at 90.4%, delivering the best overall performance among all evaluated guardrails.

Evaluating nuanced conversational travel recommendations is challenging when human annotations are costly and standard metrics ignore stakeholder-centric goals. We study LLMs-as-Judges for sustainable city-trip lists across four dimensions -- relevance, diversity, sustainability, and popularity balance, and propose a three-phase calibration framework: (1) baseline judging with multiple LLMs, (2) expert evaluation to identify systematic misalignment, and (3) dimension-specific calibration via rules and few-shot examples. Across two recommendation settings, we observe model-specific biases and high dimension-level variance, even when judges agree on overall rankings. Calibration clarifies reasoning per dimension but exposes divergent interpretations of sustainability, highlighting the need for transparent, bias-aware LLM evaluation. Prompts and code are released for reproducibility: https://github.com/ashmibanerjee/trs-llm-calibration.

Active learning algorithms automatically identify the most informative samples from large amounts of unlabeled data and tremendously reduce human annotation effort in inducing a machine learning model. In a conventional active learning setup, the labeling oracles are assumed to be infallible, that is, they always provide correct answers (in terms of class labels) to the queried unlabeled instances, which cannot be guaranteed in real-world applications. To this end, a body of research has focused on the development of active learning algorithms in the presence of imperfect / noisy oracles. Existing research on active learning with noisy oracles typically simulate the oracles using machine learning models; however, real-world situations are much more challenging, and using ML models to simulate the annotation patterns may not appropriately capture the nuances of real-world annotation challenges. In this research, we first collect annotations of text samples (from 3 benchmark text classification datasets) from crowd-sourced workers through a crowd-sourcing platform. We then conduct extensive empirical studies of 8 commonly used active learning techniques (in conjunction with deep neural networks) using the obtained annotations. Our analyses sheds light on the performance of these techniques under real-world challenges, where annotators can provide incorrect labels, and can also refuse to provide labels. We hope this research will provide valuable insights that will be useful for the deployment of deep active learning systems in real-world applications. The obtained annotations can be accessed at https://github.com/varuntotakura/al_rcta/.

Text-based person anomaly search retrieves specific behavioral events from surveillance archives using natural-language queries. Although recent pose-aware methods align geometric structures well, they face a fundamental Pose-Semantic Gap: semantically different actions can share similar skeletal geometries. While Multimodal Large Language Models (MLLMs) can reduce this ambiguity, using them for large-scale retrieval is computationally prohibitive. We propose the Structure-Semantic Decoupled Cascade (SSDC) framework, which decouples retrieval into two stages: (1) Structure-Aware Coarse Retrieval, where a lightweight model quickly filters candidates by skeletal similarity ; and (2) Detective Squad Interaction, a multi-agent semantic verification module. The squad consists of a Detective for fast binary filtering, an Analyst for evidence extraction, and a Writer for semantic synthesis. Finally, we re-rank candidates by fusing the synthesized captions with structural priors. Experiments on the PAB benchmark show that SSDC achieves state-of-the-art performance by balancing efficiency and semantic reasoning.

We propose a new approach for a practical two-stage Optical Music Recognition (OMR) pipeline, with a particular focus on its second stage. Given symbol and event candidates from the visual pipeline, we decode them into an editable, verifiable, and exportable score structure. We focus on complex polyphonic staff notation, especially piano scores, where voice separation and intra-measure timing are the main bottlenecks. Our approach formulates second-stage decoding as a structure decoding problem and uses topology recognition with probability-guided search (BeadSolver) as its core method. We also describe a data strategy that combines procedural generation with recognition-feedback annotations. The result is a practical decoding component for real OMR systems and a path to accumulate structured score data for future end-to-end, multimodal, and RL-style methods.

Automated bacterial colony counting from images is an important technique to obtain data required for the development of vaccines and antibiotics. However, bacterial colonies present unique machine vision challenges that affect counting, including (1) small physical size, (2) object clustering, (3) high data annotation cost, and (4) limited cross-species generalisation. While FamNet is an established object counting technique effective for clustered objects and costly data annotation, its effectiveness for small colony sizes and cross-species generalisation remains unknown. To address the first three challenges, we propose ACFamNet, an extension of FamNet that handles small and clustered objects using a novel region of interest pooling with alignment and optimised feature engineering. To address all four challenges above, we introduce ACFamNet Pro, which augments ACFamNet with multi-head attention and residual connections, enabling dynamic weighting of objects and improved gradient flow. Experiments show that ACFamNet Pro achieves a mean normalised absolute error (MNAE) of 9.64% under 5-fold cross-validation, outperforming ACFamNet and FamNet by 2.23% and 12.71%, respectively.

Previous LLMs-based RL studies typically follow either supervised learning with high annotation costs, or unsupervised paradigms using voting or entropy-based rewards. However, their performance remains far from satisfactory due to the substantial annotation cost and issues such as model collapse or reward hacking. To address these issues, we introduce a new perspective inspired by cognitive learning theory and propose a novel approach called EasyRL. The core of EasyRL is to simulate the human cognitive acquisition curve by integrating reliable knowledge transfer from easy labeled data with a progressive divide-and-conquer strategy that tackles increasingly difficult unlabeled data. Specifically, we initialize a warm-up model using supervised RL with few-shot labeled data. This is followed by a divide-and-conquer pseudo-labeling strategy on difficult unlabeled data, combining consistency-based selection for low-uncertainty cases and reflection-based resolution for medium-uncertainty cases. Finally, difficulty-progressive self-training with iterative pseudo-labeling and RL further strengthens the model's reasoning capability. EasyRL provides a unified self-evolving framework that facilitates data-efficient post-training of LLMs. Experimental results on mathematical and scientific benchmarks demonstrate that EasyRL, using only 10% of easy labeled data, consistently outperforms state-of-the-art baselines.

Video editing aims to modify input videos according to user intent. Recently, end-to-end training methods have garnered widespread attention, constructing paired video editing data through video generation or editing models. However, compared to image editing, the high annotation costs of video data severely constrain the scale, quality, and task diversity of video editing datasets when relying on video generative models or manual annotation. To bridge this gap, we propose LIVE, a joint training framework that leverages large-scale, high-quality image editing data alongside video datasets to bolster editing capabilities. To mitigate the domain discrepancy between static images and dynamic videos, we introduce a frame-wise token noise strategy, which treats the latents of specific frames as reasoning tokens, leveraging large pretrained video generative models to create plausible temporal transformations. Moreover, through cleaning public datasets and constructing an automated data pipeline, we adopt a two-stage training strategy to anneal video editing capabilities. Furthermore, we curate a comprehensive evaluation benchmark encompassing over 60 challenging tasks that are prevalent in image editing but scarce in existing video datasets. Extensive comparative and ablation experiments demonstrate that our method achieves state-of-the-art performance. The source code will be publicly available.

Automated evaluation of tool-using large language model (LLM) agents is widely assumed to be reliable, but this assumption has rarely been validated against human annotation. We introduce AgentProp-Bench, a 2,000-task benchmark with 2,300 traces across four domains, nine production LLMs, and a 100-label human-validated subset. We quantify judge reliability, characterize error propagation, and evaluate a runtime mitigation. Substring-based judging agrees with human annotation at kappa=0.049 (chance-level); a three-LLM ensemble reaches kappa=0.432 (moderate) with a conservative bias. Under validated evaluation, a parameter-level injection propagates to a wrong final answer with human-calibrated probability approximately 0.62 (range 0.46-0.73 across models). Rejection (catching bad parameters) and recovery (correcting after acceptance) are independent model capabilities (Spearman rho=0.126, p=0.747). A tuned runtime interceptor reduces hallucination on GPT-4o-mini by 23.0 percentage points under a concurrent n=600 control, but shows no significant effect on Gemini-2.0-Flash, whose aggressive parameter rejection eliminates the target failure mode. All code, data, traces, and human labels are released at https://github.com/bhaskargurram-ai/agenthallu-bench.

Large language models (LLMs) have advanced information extraction (IE) by enabling zero-shot and few-shot named entity recognition (NER), yet their generative outputs still show persistent and systematic errors. Despite progress through instruction fine-tuning, zero-shot NER still lags far behind supervised systems. These recurring errors mirror inconsistencies observed in early-stage human annotation processes that resolve disagreements through pilot annotation. Motivated by this analogy, we introduce DiZiNER (Disagreement-guided Instruction Refinement via Pilot Annotation Simulation for Zero-shot Named Entity Recognition), a framework that simulates the pilot annotation process, employing LLMs to act as both annotators and supervisors. Multiple heterogeneous LLMs annotate shared texts, and a supervisor model analyzes inter-model disagreements to refine task instructions. Across 18 benchmarks, DiZiNER achieves zero-shot SOTA results on 14 datasets, improving prior bests by +8.0 F1 and reducing the zero-shot to supervised gap by over +11 points. It also consistently outperforms its supervisor, GPT-5 mini, indicating that improvements stem from disagreement-guided instruction refinement rather than model capacity. Pairwise agreement between models shows a strong correlation with NER performance, further supporting this finding.

Behavioral Profile (BP) annotation is difficult to automate because it requires simultaneous coding across multiple linguistic dimensions. We treat BP annotation as a bundle of annotation skills rather than a single task and evaluate LLM-assisted BP annotation from this perspective. Using 3,134 concordance lines of 30 Chinese metaphorical color-term derivatives and a 14-feature BP schema, we implement a skill-file-driven pipeline in which each feature is externally defined through schema files, decision rules, and examples. Two human annotators completed a two-round schema-only protocol on a 300-instance validation subset, enabling BP skills to be classified as directly operable, recoverable under focused re-annotation, or structurally underspecified. GPT-5.4 and three locally deployable open-source models were then evaluated under the same setup. Results show that BP annotation is highly heterogeneous at the skill level: 5 skills are directly operable, 4 are recoverable after focused re-annotation, and 5 remain structurally underspecified. GPT-5.4 executes the retained skills with substantial reliability (accuracy = 0.678, \k{appa} = 0.665, weighted F1 = 0.695), but this feasibility is selective rather than global. Human and GPT difficulty profiles are strongly aligned at the skill level (r = 0.881), but not at the instance level (r = 0.016) or lexical-item level (r = -0.142), a pattern we describe as shared taxonomy, independent execution. Pairwise agreement further suggests that GPT is better understood as an independent third skill voice than as a direct human substitute. Open-source failures are concentrated in schema-to-skill execution problems. These findings suggest that automatic annotation should be evaluated in terms of skill feasibility rather than task-level automation.

Annotating data remains a costly bottleneck for supervised NLP. Active learning (AL) reduces the number of human labels needed by selecting only the most informative instances, while instruction-tuned LLMs attack the same bottleneck from the other side, making labels cheap enough to annotate entire corpora. This raises two questions: can LLM labels replace human labels within the AL loop, and does AL remain necessary when entire corpora can be cheaply labeled? We investigate both by training supervised hate speech classifiers on a new dataset of 278K German political TikTok comments, comparing human and LLM annotation under matched conditions. LLM annotation at scale outperforms human-supervised classifiers at roughly one-tenth the cost, for both a closed-source (GPT-5.2) and an open-weight (Qwen3.5-122B-A10B) LLM, and the advantage is robust under soft-label evaluation. It hinges on the annotation interface: only a two-question decomposition mirroring the human annotation task unlocks it. AL provides no reliable advantage over random sampling in our prefiltered pool. Error structure depends on the LLM: only GPT-5.2 matches the human FP/FN balance, while other variants over-flag border-control and economic-competition discourse. Humans remain essential as evaluators; for training labels, the question shifts to which LLM, which interface, and what shape of pool.

Human data annotation, especially when involving experts, is often treated as an objective reference. However, many annotation tasks are inherently subjective, and annotators' judgments may evolve over time. This study investigates changes in the quality of annotators' work from a competence perspective during a process of social influence recognition. The study involved 25 annotators from five different groups, including both experts and non-experts, who annotated a dataset of 1,021 dialogues with 20 social influence techniques, along with intentions, reactions, and consequences. An initial subset of 150 texts was annotated twice - before and after the main annotation process - to enable comparison. To measure competence shifts, we combined qualitative and quantitative analyses of the annotated data, semi-structured interviews with annotators, self-assessment surveys, and Large Language Model training and evaluation on the comparison dataset. The results indicate a significant increase in annotators' self-perceived competence and confidence. Moreover, observed changes in data quality suggest that the annotation process may enhance annotator competence and that this effect is more pronounced in expert groups. The observed shifts in annotator competence have a visible impact on the performance of LLMs trained on their annotated data.

Error Span Detection (ESD) is a crucial subtask in Machine Translation (MT) evaluation, aiming to identify the location and severity of translation errors. While fine-tuning models on human-annotated data improves ESD performance, acquiring such data is expensive and prone to inconsistencies among annotators. To address this, we propose a novel self-evolution framework based on Minimum Bayes Risk (MBR) decoding, named Iterative MBR Distillation for ESD, which eliminates the reliance on human annotations by leveraging an off-the-shelf LLM to generate pseudo-labels. Extensive experiments on the WMT Metrics Shared Task datasets demonstrate that models trained solely on these self-generated pseudo-labels outperform both unadapted base model and supervised baselines trained on human annotations at the system and span levels, while maintaining competitive sentence-level performance.

Evaluating the real-world capabilities of AI systems requires grounding benchmark performance in human-interpretable measures of task difficulty. Existing approaches that rely on direct human task completion time annotations are costly, noisy, and difficult to scale across benchmarks. In this work, we propose BRIDGE, a unified psychometric framework that learns a latent difficulty scale from model responses and anchors it to human task completion time. Using a two-parameter logistic Item Response Theory model, we jointly estimate latent task difficulty and model capability from model performance data across multiple benchmarks. We demonstrate that latent task difficulty varies linearly with the logarithm of human completion time, allowing human task completion time to be inferred for new benchmarks from model performance alone. Leveraging this alignment, we forecast frontier model capabilities in terms of human task length and independently reproduce METR's exponential scaling results, with the 50% solvable task horizon doubling approximately every 6 months.

Semi-supervised few-shot learning (SSFSL) resembles real-world applications such as auto-annotation, as it aims to learn a model from a few labeled and abundant unlabeled task-specific examples to annotate the unlabeled ones. Despite the availability of powerful open-source Vision-Language Models (VLMs) and open-world data, existing SSFSL literature largely neglects these resources. In contrast, the related area few-shot learning (FSL) has already exploited them to boost performance. Arguably, to solve real-world auto-annotation, SSFSL should leverage such open resources. To bridge this gap, we explore established SSL methods to finetune a VLM. Unexpectedly, they significantly underperform FSL baselines that do not use unlabeled data. Our in-depth analysis reveals the root cause of failure: VLMs produce flat distributions of softmax probabilities, resulting in zero utilization of unlabeled data and weak supervision signals. To address this challenge, we propose an embarrassingly simple solution that uses temperatures to sharpen the softmax output, which not only increases the confidence scores of pseudo-labels to improve the utilization of unlabeled data, but also strengthens training supervision for effective finetuning. Furthermore, we exploit task-relevant open data, e.g., those retrieved from VLMs' publicly available pretraining set. To mitigate the imbalance and domain gaps in retrieved data, we employ a stage-wise training strategy. Building on the successful finetuning of VLMs and the exploitation of open data, we present a simple yet effective SSFSL method, Stage-Wise Finetuning with Temperatures (SWIFT). Across five benchmarks, SWIFT outperforms recent FSL and SSL methods by $\sim$5 accuracy points. SWIFT even rivals supervised learning, which finetunes a VLM assuming unlabeled data having ground-truth labels!

Large language models (LLMs) have demonstrated significant potential to accelerate scientific discovery as valuable tools for analyzing data, generating hypotheses, and supporting innovative approaches in various scientific fields. In this work, we investigate how LLMs can handle the transition from conceptual research ideas to well-structured research plans. Effective research planning not only supports scientists in advancing their research but also represents a crucial capability for the development of autonomous research agents. Despite its importance, the field lacks a systematic understanding of LLMs' research planning capability. To rigorously measure this capability, we introduce the Idea2Plan task and Idea2Plan Bench, a set of benchmarks built from ICML 2025 and Nature Mental Health papers released after major LLM training cutoffs. Each benchmark instance includes a research idea and a grading rubric capturing the key components of valid plans. We further propose Idea2Plan JudgeEval, a complementary benchmark to assess the reliability of LLM-based judges against expert annotations. Experimental results show that GPT-5 achieves the strongest performance on the benchmark, though substantial headroom remains for improvement. Our study provides new insights into LLMs' capability for research planning and lays the groundwork for future progress.

Understanding how people interact with their surroundings and each other is essential for enabling robots to act in socially compliant and context-aware ways. While 3D Scene Graphs have emerged as a powerful semantic representation for scene understanding, existing approaches largely ignore humans in the scene, also due to the lack of annotated human-environment relationships. Moreover, existing methods typically capture only open-vocabulary relations from single image frames, which limits their ability to model long-range interactions beyond the observed content. We introduce Social 3D Scene Graphs, an augmented 3D Scene Graph representation that captures humans, their attributes, activities and relationships in the environment, both local and remote, using an open-vocabulary framework. Furthermore, we introduce a new benchmark consisting of synthetic environments with comprehensive human-scene relationship annotations and diverse types of queries for evaluating social scene understanding in 3D. The experiments demonstrate that our representation improves human activity prediction and reasoning about human-environment relations, paving the way toward socially intelligent robots.

Existing informal language-based (e.g., human language) Large Language Models (LLMs) trained with Reinforcement Learning (RL) face a significant challenge: their verification processes, which provide crucial training signals, are neither reliable nor scalable. In fact, the prevalent large proprietary models could hardly generate verifiable programs. A promising yet largely uncharted alternative is formal language-based reasoning. Grounding LLMs in rigorous formal systems where generative models operate in formal language spaces (e.g., Dafny) enables the automatic and mathematically provable verification of their reasoning processes and outcomes. This capability is pivotal for achieving large-scale, reliable formal software verification. It is a common practice to employ human-annotated chain-of-thought and answers to induce the reasoning and coding capabilities of LLMs. Unfortunately, it becomes unacceptably all-consuming to provide such priors for supervising complex programming tasks. In this work, we systematically explore ways to reduce human annotations with the formal language, Dafny, as the main environment for our pilot study. Our pipeline mainly relies on introducing an automatic and scalable data curation pipeline, and careful RL designs integrated with feedback from the formal language verifier. We introduce DafnyComp, a benchmark of compositional formal programs with auto-formalized specifications for specification reasoning. Our supervised fine-tuning (SFT) stage enables even small models (e.g., 0.5B) to generate syntactically valid and verifiable Dafny code, surpassing proprietary models. RL with regularization further improves performance, achieving stronger generalization to out-of-domain tasks and outperforming all strong baselines on the challenging DafnyComp benchmark.

Vision-language Models (VLMs) are essential for contextual understanding of both visual and textual information. However, their vulnerability to adversarially manipulated inputs presents significant risks, leading to compromised outputs and raising concerns about the reliability in VLM-integrated applications. Detecting these malicious prompts is thus crucial for maintaining trust in VLM generations. A major challenge in developing a safeguarding prompt classifier is the lack of a large amount of labeled benign and malicious data. To address the issue, we introduce VLMGuard, a novel learning framework that leverages the unlabeled user prompts in the wild for malicious prompt detection. These unlabeled prompts, which naturally arise when VLMs are deployed in the open world, consist of both benign and malicious information. To harness the unlabeled data, we present an automated maliciousness estimation score for distinguishing between benign and malicious samples within this unlabeled mixture, thereby enabling the training of a binary prompt classifier on top. Notably, our framework does not require extra human annotations and is robust to realistic prompt variations, offering strong flexibility and practicality for real-world applications. Extensive experiments show that VLMGuard achieves superior detection results, improving AUROC by 5.39% on average over the state-of-the-art method. Disclaimer: This paper may contain offensive examples; reader discretion is advised. Code is available at: https://github.com/radiolab-ntu/vlmguard.