Optimal Regret

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 13Week of Sep 28

Latest papers 27

All topics

False Floors: LLM Safety Routing Evaluations Break Under Distribution Shift
Oct 1, 2026Amit Singh Bhatti, Vishal VaddinaLarge Language Model Safety Distribution Shifts
Quantiphi Analytics
From Switching to Dynamic Regret: A Simple Reduction via Unbiased Random Sequences
Sep 17, 2026Yibo Wang, Wenhao Yang, Sifan Yang +3Optimal Regret Learning-Augmented Algorithms
State Key Laboratory for Novel Software Technology, Nanjing University · School of Artificial Intelligence, Nanjing University · School of Software Technology, Zhejiang University
Bandits with Probing: Optimal Regret and the Limits of Winner Feedback
Sep 14, 2026Yongjie GuanOptimal Regret Minimax
Zhejiang University of Technology
Bandits in Prod: Hyperparameter Optimization at Inference Time
Sep 1, 2026Louis Abraham, Tuan-Anh Nguyen, Nicolas DevatineGeneral Grid Search Framework Agentic Optimization
Tiime Paris, France
Tracking the Best Strategy in an Extensive-Form Game
Aug 10, 2026Stephen Pasteris, Rahul Savani, Theodore TurocyOptimal Regret Imperfect-Information Games
The Alan Turing Institute · London, United Kingdom · The University of Liverpool +3
Optimal and Efficient Contextual Combinatorial Semi-bandits with General Function Approximation
Jul 15, 2026Hao Qin, Chicheng ZhangContextual Bandit Framework Stochastic Multi-Armed Bandits
The University of Arizona
Nonlinear Bandit
Jul 8, 2026Tianshuo Zheng, Ting Wu, Zhi-Hua Zhou +1Optimal Regret Loss Function
School of Mathematics, Nanjing University, Nanjing, 210093, China · School of Artificial Intelligence, Nanjing University, National Key Laboratory for Novel Software Technology, Nanjing, 210023, China · School of Mathematics and Physics, Xi’an Jiaotong-Liverpool University, Suzhou, 215123, China
Constrained Online Convex Optimization without Slater's Condition
Jun 30, 2026Kihyun Yu, Junehee Lee, Dabeen LeeConvex Loss \Widetilde{\Mathcal{O}}(\Sqrt{T})$ Regret
1KAIST · 2Seoul National University
Stochastic Linear Contextual Bandits with Bounded Noise: A Set-Membership Approach
Jun 18, 2026Haonan Xu, Yingying LiContextual Bandit Framework Optimal Regret
The Grainger College of Engineering University of Illinois Urbana-Champaign Champaign, IL, 61801
Matching Markets meet Cumulative Prospect Theory: Towards Optimal and Adversarially Robust Learning
Jun 18, 2026Ananya Kunisetty, Avishek GhoshOptimal Regret Matching Markets
Indian Institute of Technology Bombay, Mumbai 400076, India
A Solver-Free Training Method for Predict-then-Optimize
Jun 17, 2026Beichen Wan, Mo LiuOptimization Modeling Surrogate Gradient
Department of Statistics and Operations Research, University of North Carolina at Chapel Hill, NC, USA.
Toward Simultaneously Optimal Regret in U-Calibration
Jun 16, 2026Rafael Frongillo, Haipeng Luo, Nishant A. Mehta +1Optimal Regret Regret
University of Colorado Boulder · University of Southern California and Google Research · University of Victoria +1
Optimal Gap-Dependent Regret for Private Stochastic Decision-Theoretic Online Learning
May 27, 2026Tommaso Cesari, Roberto ColomboniOptimal Regret Standard Differential-Privacy
School of Electrical Engineering and Computer Science University of Ottawa Ottawa, Canada · School of Mathematics University of Bristol Bristol, United Kingdom
Online Learning on Hidden-Convex Losses via Algorithmic Equivalence: Optimal Regret, Geometric Barrier, and Bandit Feedback
May 25, 2026Anas Barakat, Andreas Kontogiannis, Vasilis Pollatos +2Convex Loss \Widetilde{\Mathcal{O}}(\Sqrt{T})$ Regret
Singapore University of Technology and Design · National Technical University of Athens · Archimedes, Athena Research Center, Greece +3
Infra-Bayesian Reinforcement Learning Agents Outperform Classical RL For Worst-Case Robustness
May 22, 2026Manish Aryal, Faiyaz Azam, Agnivo Banerjee +11Offline Reinforcement Learning Optimal Regret
Purdue University · Carnegie Mellon University · WorldQuant University +11
Regret-Based $(ε,δ)$ -optimal Stopping Criteria for Bayesian Optimization
May 21, 2026Haowei Wang, Jingyi Wang, Qiyu WeiBayesian Optimization Optimal Stopping
National University of Singapore, Singapore · Lawrence Livermore National Laboratory, CA, USA · The University of Manchester, UK
Tighter Regret Bounds for Contextual Action-Set Reinforcement Learning
May 15, 2026Zijun Chen, Zihan ZhangOptimal Regret Offline Reinforcement Learning
Department of Computer Science and Engineering Hong Kong University of Science and Technology
Regret Equals Covariance: A Closed-Form Characterization for Stochastic Optimization
May 13, 2026Irene AldridgeOptimal Regret Regret
IGT-OMD: Implicit Gradient Transport for Decision-Focused Learning under Delayed Feedback
May 12, 2026Benjamin Amoh, Geoffrey G. Parker, Wesley MarreroBilevel Optimization Optimal Regret
Thayer School of Engineering, Dartmouth College
Learning Weakly Communicating Average-Reward CMDPs: Strong Duality and Improved Regret
May 12, 2026Kihyun Yu, Beomhan Baek, Dabeen LeeMarkov Decision Processes Optimal Regret
1KAIST · 2Seoul National University
Online Sharp-Calibrated Bayesian Optimization
May 11, 2026Marshal Arijona Sinaga, Julien Martinelli, Teemu Turpeinen +1Bayesian Optimization Gaussian Process
ELLIS Institute Finland · Aalto University · University of Manchester
PFN-TS: Thompson Sampling for Contextual Bandits via Prior-Data Fitted Networks
May 11, 2026Yan Shuo Tan, Kenyon Ng, Ruizhe Deng +3Thompson Sampling Tabular Prior-Data Fitted Network
National University of Singapore · Monash University · Renmin University of China +1
Optimal Regret for Single Index Bandits
May 10, 2026Devdan Dey, Sujoy Bhore, Avishek GhoshStochastic Multi-Armed Bandits Optimal Regret
Department of Computer Science and Engineering, Indian Institute of Technology Bombay.
Cost-Ordered Feasibility for Multi-Armed Bandits with Cost Subsidy
May 8, 2026Ishank Juneja, Carlee Joe-Wong, Osman YağanStochastic Multi-Armed Bandits Optimal Regret
Department of Electrical and Computer Engineering Carnegie Mellon University Pittsburgh, PA 15213
Bandit Learning in General Open Multi-agent Systems
May 7, 2026Mengfan XuBandits Optimal Regret
Mechanical and Industrial Engineering University of Massachusetts Amherst
From Confounding to Learning: Dynamic Service Fee Pricing on Third-Party Platforms
Dec 28, 2025Rui Ai, David Simchi-Levi, Feng ZhuDynamic Pricing Optimal Regret
MIT
Thompson Sampling for Infinite-Horizon Discounted Decision Processes
May 14, 2024Daniel Adelman, Cagla Keceli, Alba V. Olivares-NadalThompson Sampling Markov Decision Processes
Booth School of Business, The University of Chicago, Chicago, IL 60637, USA · United Airlines, Chicago, IL 60605, USA · UNSW Business School, UNSW, Kensington NSW 2033, Australia

Oct 1, 2026, cs.CRJ/K move · Enter open · S save

False Floors: LLM Safety Routing Evaluations Break Under Distribution Shift

Amit Singh Bhatti, Vishal Vaddina

Quantiphi Analytics

Large Language Model Safety Distribution Shifts LLM Defense Mechanisms Floor Optimal Regret

Open paper AskPDF arXiv

Safety routers send each request to one of several models and are judged against the best single model. A major routing benchmark picks that comparator on the evaluation data. In the benchmark's own setting this is harmless, but under distribution shift it is not. On HELM Safety the selection cost is 0.003-0.030 of harm under random splits and 0.045-0.113 under held-out categories, comparable to the whole deficit attributed to routing, with its direction holding under either published judge alone. It rises seven- to ninefold on AgentDojo when suites are held out. Across seven safety corpora chosen by rules fixed in advance, three meet a registered interval test and four beat a later permutation null, and three of the four interval misses are corpora where some models have zero observed harm. Prior work proves the direction of this bias. We size it on harm and accuracy, show that it is larger under the held-out splits we measure, and bound it by optimism plus a shift-dependent regret. Scored honestly under shift, routing buys little on these benchmarks. In most pool cells the nested router serves the honest baseline's model, and on the nearly saturated AgentDojo corpus a perfect pre-dispatch router is worth at most two points of harm. We also find a model's expressed recognition of a late injection steerable. On held-out reruns an attacker who knows which model it faces lowers GPT-5.4's judged recognition by 19.6 points, confirmed by an independent label. In an offline counterfactual composition into a controller, the same attack raises or lowers estimated harm depending on the fallback model. Safety routing should be evaluated under shift, against a baseline chosen without the test labels, and recognition-based defences should be scored on harm against an attacker who chooses what the model sees.

Momentum

Related topics

Latest papers 27