Lock-In

Follow this topic Ask about this topic

Momentum

3 papers in the last four weeks, with none the four weeks before. 0.0% of all new papers.

Jul 13Week of Sep 28

Latest papers 30

All topics

RSI-Master: Structuring Experiments to Guide Autonomous Model Improvement
Sep 28, 2026Yaxin Du, Xiyuan Yang, Zhifan Zhou +10Recursive Self-Improvement Artificial Intelligence Research
Shanghai Jiao Tong University · Carnegie Mellon University · University of Waterloo
LOCKR: A Hidden-State Trajectory-Guided Planner for Detecting and Repairing Stable-but-Wrong Lock-In in Diffusion Language Models
Sep 23, 2026Guoshenghui Zhao, Tan Yu, Weijie ZhaoDiffusion Language Models Lock-In
Rochester Institute of Technology · NVIDIA Corporation
Canonical locks that encode part-whole hierarchies
Sep 22, 2026Rajat Modi, Yogesh Singh RawatRepresentation Learning Hierarchical
Department of Computer Science, University of Central Florida, United States.
ZeroLock: Concurrent Memory-Efficient LLM Training via Modular Update Decoupling
Aug 8, 2026Wentao Dai, Xuanran Li, Yuxiang Zhang +2Lock-In Chunk
Department of Computer Science and Engineering, Southern University of Science and Technology, China · Department of Mathematics, Southern University of Science and Technology, China · School of Computing, Montclair State University, New Jersey, USA
Evidence Lock Before Commitment: A Frozen Interface Degrades LLM-as-Judge Evaluation
Aug 5, 2026Divyansh SinghCommitment Lock-In
University of Florida
RL-Lock: Reinforcement Learning for Generating Interlocking Assemblies
Aug 3, 2026Xuyang Ma, Chaewoon Kim, Haonan Zhang +3Assembly Lock-In
Singapore University of Technology and Design · The Hong Kong University of Science and Technology · Beijing Normal-Hong Kong Baptist University
Scaling, Lock-In, and Proxy Compliance: A Political Economy of Responsible AI
Jul 30, 2026Florian A. D. Burnat, Brittany I. DavidsonAccountability Artificial Intelligence Governance
University of Bath, Bath, UK.
Prior Directions: Why GUI Grounding Gets Locked in the Past
Jul 29, 2026Weile Gong, Zijian Lu, Mingcai Chen +3Weak Visual Grounding Lock-In
Nanjing University of Posts and Telecommunications, Nanjing, China
LOCKS: Page-Local Compact Key Summaries for Efficient Long-Context Decoding
Jul 27, 2026Junsung HwangEfficient Long-Context Inference Key-Value Cache Compression
Hybrid Rigid-Soft Robotic Gripper with Shape Adaptation, Uniform Force Distribution, and Self-Locking Capabilities
Jul 16, 2026Xi Chen, Yun Wang, Lichao Yang +2Gripper Soft Pneumatic Actuators
Intelligent Equipment Research Center, Beijing Academy of Agriculture and Forestry Sciences, Beijing, China. · College of Engineering, China Agricultural University, Beijing, China. · College of Engineering, Shanxi Agricultural University, Jinzhong, China.
Relational Positioning as a Measurable Risk Object: History-Carried Lock-in and Self-Confabulation in Multi-Turn Human-AI Dialogue
Jul 13, 2026Jihong ChenAttitude Confabulation
Beijing Etown Academy
Overcoming Fourier Locking in Quantum Data Re-uploading Classifiers via Spectral Homotopy
Jul 13, 2026Spencer TopelAnisotropic Loss Landscapes Fourier
Moth, Brooklyn, New York
!Imperio, smolVLA: The Implications of Data Poisoning on Open Source Robotics
Jul 5, 2026Stefan Bühler, Mark SchuteraPoisoning AI Trustworthiness
Independent Researcher · Duale Hochschule Baden-Württemberg, Ravensburg
A Three-Phase Foundation Model for Tax-Aware Personalized Portfolio Management
Jun 30, 2026Ramin PishehvarPortfolio Construction Offline Reinforcement Learning
When Confidence Takes the Wrong Path: Diagnosing Retrieval-State Lock-In in RAG
Jun 22, 2026Sahib JulkaLock-In Correct
LMU University Hospital, LMU Munich · Marchioninistr. 15, 81377 Munich
The Governance of Human-LLM Interaction: Safety Gating, Civility Steering, and Affective Default Lock-In
Jun 6, 2026Manuele Reani, Hongjian Zhang, Hongyu TianLarge Language Model Safety Governance
School of Management and Economics, The Chinese University of Hong Kong, Shenzhen, China
Breaking the Lock-in: Diversifying Text-to-Image Generation via Representation Modulation
Jun 5, 2026Dahee Kwon, Haeun Lee, Jaesik ChoiModern Text-To-Image Models Text-To-Image
Korea Advanced Institute of Science and Technology (KAIST) · INEEJI
Physics in 2-Steps: Locking Motion Priors Before Visual Refinement Erases Them
Jun 4, 2026Woojung Han, Seil Kang, Youngjun Jun +3Denoising Trajectory Diffusion Models
1Yonsei University, South Korea · 2NVIDIA, Taiwan.
The Path Matters: Learning a Token-Commitment Policy for Diffusion Language Models
May 23, 2026Bohang Sun, Max Zhu, Francesco Caso +5Diffusion Language Models Frozen Encoder
Department of Computer Science and Technology, University of Cambridge · Department of Engineering Science, University of Oxford
Identifying and Mitigating Systemic Measurement Bias in Production LLM Inference Benchmarks
May 22, 2026Ashok Chandrasekar, Jason KrambergerLLM Inference Optimization Bottlenecks
Google
Targeting Clause Type Distributions: a Picklock for Random Satisfiability Problems
May 19, 2026J. Schwardt, J. C. BudichSatisfiability Large Neighborhood Search
Max Planck Institute for the Physics of Complex Systems, Nöthnitzer Str. 38, 01187 Dresden, Germany · Institute of Theoretical Physics, Technische Universität Dresden and Würzburg-Dresden Cluster of Excellence ctd.qmat, 01062 Dresden, Germany
A Fast and Energy-Efficient Latch-Based Memristive Analog Content-Addressable Memory
May 12, 2026Paul-Philipp Manea, Aishwarya Natarajan, Jim Ignowski +2Reram Crossbar Analog
PGI-14, Forschungszentrum Jülich, Jülich, Germany · RWTH Aachen University, Aachen, Germany · Hewlett Packard Enterprise, Fort Collins, CO, USA
Locking Pretrained Weights via Deep Low-Rank Residual Distillation
May 11, 2026Keitaro Sakamoto, Pierre Ablin, Federico Danieli +1Model-Agnostic Defense Model Weights
Apple
Matching Meaning at Scale: Evaluating Semantic Search for 18th-Century Intellectual History through the Case of Locke
May 10, 2026Yu Wu, Ananth Mahadevan, Filip Ginter +2Semantic Relationships Text Corpora
University of Helsinki · 2TurkuNLP, University of Turku · 3ELLIS Institute Finland
Adaptive Pluralistic Alignment: A pipeline for dynamic artificial democracy
May 2, 2026Rachel FreedmanPluralistic Alignment Lock-In
Department of Electrical Engineering and Computer Science, University of California, Berkeley, USA.
Path-Lock Expert: Separating Reasoning Mode in Hybrid Thinking via Architecture-Level Separation
Apr 29, 2026Shouren Wang, Wang Yang, Chuang Ma +7Thoughts Lock-In
1Case Western Reserve University · 2NII LLMC, Japan · 3Michigan State University
Beyond the Attention Stability Boundary: Agentic Self-Synthesizing Reasoning Protocols
Apr 27, 2026Dahlia Shehata, Ming LiAgentic Reasoning Skills
University of Waterloo · Canada
Breaking Lock-In: Preserving Steerability under Low-Data VLA Post-Training
Apr 25, 2026Suning Huang, Jiaqi Shao, Ke Wang +5Generalist Vision--Language--Action Frozen Vision-Language Models
Stanford University
Forward Dynamics of Variable Topology Mechanisms - The Case of Constraint Activation
Apr 21, 2026Andreas MuellerKinematic Chain Robotic Manipulator
Institute of Robotics, Johannes Kepler University Altenbergerstr. 69, 4040 Linz, Austria
Regulating Artificial Intimacy: From Locks and Blocks to Relational Accountability
Apr 20, 2026Henry Fraser, Jessica M. Szczuka, Raffaele F. CirielloArtificial Intelligence Companions Regulation
Queensland University of Technology · School of Law · ARC Centre of Excellence for Automated Decision-Making and Society (ADM+S) +3

Sep 28, 2026, cs.AIJ/K move · Enter open · S save

RSI-Master: Structuring Experiments to Guide Autonomous Model Improvement

Yaxin Du, Xiyuan Yang, Zhifan Zhou, Yujie Ge+9

Shanghai Jiao Tong University · Carnegie Mellon University · University of Waterloo

Recursive Self-Improvement Artificial Intelligence Research Research Automation Improvement Dream Lock-In

Open paper AskPDF arXiv

Recursive self-improvement (RSI) seeks to enable AI systems to participate in improving their own capabilities. A concrete pathway is autonomous model development, where agents iteratively explore post-training strategies to improve a base model. This setting faces two challenges: agents may exploit open-ended experimental actions through hacking, and repeated experimentation may lead to strategy lock-in, where an early direction is refined rather than reconsidered. We introduce RSI-Master, which addresses the two challenges at two levels: regularize step-wise actions, avoiding hacking behaviors, and promote well-structured exploration of research directions, avoiding strategy lock-in. RSI-Master consists of an Experiment OS, which enables regularized experimental actions and maintains persistent, traceable experimental records, and Reviewer-Guided Research Orchestration, which organizes Workers and Reviewers in a dynamically growing research DAG. Workers explore diverse research directions and Reviewers compare evidence across related experiments for subsequent explorations. On PostTrainBench with Qwen3-4B-Base, it averages 54.49 versus 46.53 for the strongest agent baseline, with a 0.0% hacking rate. Scaling to 35B model, RSI-Master surpasses the human-developed Instruct model on LiveCodeBench-v6 (41.21 vs. 37.36) and SciCode, and reaches a nonzero score on HorizonMath, a benchmark of unsolved research problems on which most frontier models score near zero.

Momentum

Related topics

Latest papers 30