Fast-Slow

Follow this topic Ask about this topic

Momentum

3 papers in the last four weeks, against 2 the four weeks before. 0.0% of all new papers.

Jul 6Week of Sep 21

Latest papers 15

All topics

Unlearnable, or Unmeasured? On the Reliability of Difficulty Labels in RLVR
Sep 30, 2026Chandak Chakma, Syed Nazmus Sakib, Nafiul Haque +1Reinforcement Learning With Verifiable Reward Verifiable Rewards
Department of Robotics and Mechatronics Engineering, University of Dhaka
SpikeCredit: Temporal Credit Carrier for Reinforcement Learning with Sparse Rewards
Sep 28, 2026Yingchao Yu, Pengfei Sun, Wenxuan Pan +4Time-To-First-Spike Step-Level Credit
School of Information and Intelligent Science, Donghua University, Shanghai, China · School of Engineering, Westlake University, Hangzhou, China · Department of Electrical and Electronic Engineering, Imperial College London, London, United Kingdom +1
Slow Decay and Silenced Expression: Iterated Subliminal Trait Transfer in Language-Model Lineages
Sep 22, 2026Ryan Vo, Duc-Vu Nguyen, Matt Kretchmar +1Lineages Fast-Slow
Denison University, Granville, OH, USA · VNUHCM - University of Information Technology,, Ho Chi Minh City, Vietnam
TAO-Force: Unifying Force-Aware Perception and Fast-Slow Control for Contact-Rich Manipulation
Sep 16, 2026Bohan Gan, Xuanzhang Wen, Yongsheng Zhao +8Contact-Rich Manipulation Viewpoint-Dependent Active Perception
China Mobile (Hangzhou) Information Technology Co., Ltd., Hangzhou 310023, China
ReactVAU: A Slow-Fast Decoupled Framework for Streaming Video Anomaly Understanding
Sep 7, 2026Chia-Hui Chen, Shih-Ying Yeh, Fu-En Yang +2Training-Free Video Anomaly Detection Fast-Slow
National Tsing Hua University, Hsinchu, Taiwan · NVIDIA, Taipei, Taiwan
Dual-Layer Agentic Memory with Fast Write Routing and Slow Consolidation
Aug 23, 2026Wenzhi Li, Dong Nie, Rui Lan +7Agentic Memory Internalization
Zhejiang University · Indepedent Researcher · Xiaohongshu +2
WhiteMatter: All-to-All Cross-Layer Connections via KV Source Mixing
Aug 19, 2026Wenbo Zhang, Xiang RenTransformer Architectures Cross-Layer Interactions
University of Southern California
The Latent Bridge: A Continuous Slow-Fast Channel for Real-Time Game Agents
Jun 23, 2026Bojie Li, Noah ShiEfficient Latent Reasoning Inference Latency
Pine AI · University of Washington
Fast and Slow Variational Continual Learning
Jun 22, 2026Subarnaduti Paul, Yohan Jung, Mohammad Emtiyaz Khan +3Replay-Based Continual Learning Adaptive Optimizers
Faculty of Mathematics and Computer Science, University of Bremen, Bremen, Germany · Department of Computer Science & AI, Jeonbuk National University, Republic of Korea · Department of Computer Science, TU Darmstadt, Darmstadt, Germany +3
Attention is Just Another Name for Coupling? A Fast-Slow ODE Perspective on Hierarchical Pretraining
Jun 15, 2026Zhengyuan GaoTransformer Attention Pretraining
Independent scholar
Approaching I/O-optimality for Approximate Attention
May 22, 2026Pál András Papp, Aleksandros Sobczyk, Anastasios ZouziasKimi Delta Attention Matrix Multiplication
Computing Systems Lab, Huawei Technologies, Zurich, Switzerland
Closing the Loop: Unified 3D Scene Generation and Immersive Interaction via LLM-RL Coupling
May 7, 2026Anh H. Vo, Sungyo Lee, Phil-Joong Kim +23D Scene Generation Immersion
Department of Computer Engineering, Sejong University, Seoul, Republic of Korea
CCL-D: A High-Precision Diagnostic System for Slow and Hang Anomalies in Large-Scale Model Training
May 6, 2026Yida Gu, Fakang Wang, Jianhao Fu +17Fast-Slow Diagnosis
University of Chinese Academy of Sciences · Ant Group · Jilin University
SlowBA: An efficiency backdoor attack towards VLM-based GUI agents
Mar 9, 2026Junxian Li, Tu Lan, Haozhen Tan +2Graphical User Interface Agents Fast-Slow
School of Computer Science, Shanghai Jiao Tong University
Contrastive Deep Learning Reveals Age Biomarkers in Histopathological Skin Biopsies
Nov 25, 2024Kaustubh Chakradeo, Pernille Nielsen, Lise Mette Rahbek Gjerdrum +5Biomarker Deep Learning
University of Copenhagen, Section of Epidemiology, Department of Public Health, Copenhagen, Denmark · University of Copenhagen, Section of Epidemiology, Department of Public Health, Copenhagen, 1353, Denmark · Technical University of Denmark, Department of Applied Mathematics and Computer Science, Denmark +7

Sep 30, 2026, cs.AIJ/K move · Enter open · S save

Unlearnable, or Unmeasured? On the Reliability of Difficulty Labels in RLVR

Chandak Chakma, Syed Nazmus Sakib, Nafiul Haque, Shifat E. Arman

Department of Robotics and Mechatronics Engineering, University of Dhaka

Reinforcement Learning With Verifiable Reward Verifiable Rewards Unlearnable Examples Fast-Slow Prompt Engineering

Open paper AskPDF arXiv

Reinforcement learning with verifiable rewards (RLVR) has become an important approach for improving reasoning during post-training. Recent work suggests that some difficult prompts remain resistant to learning even when they occasionally produce correct solutions. We revisit this unlearnability phenomenon and find that the affected prompts do improve, at roughly one third of the learnable rate, while the difficulty-defined set used to study them is much less reproducible than expected. These difficulty labels are estimated from a limited number of sampled responses. Combining them across seeds can further change which prompts are selected instead of simply reducing measurement noise. We develop a sampling-based framework for quantifying this instability and determining how much evaluation is required for difficulty assignments to reproduce reliably. We also revisit the gradient-similarity evidence proposed to explain unlearnability and show that part of the observed separation arises because difficult prompts provide fewer correct rollouts from which their gradients can be estimated. Matching this sample count weakens the gradient difference but does not remove it. Overall, the slow-learning phenomenon survives our reanalysis, while both the prompts used to define it and the evidence used to explain it require more careful measurement.

Momentum

Related topics

Latest papers 15