Image de couverture de livre

Back to Blackwell: Closing the Loop on Intransitivity in Multi-Objective Preference Fine-Tuning

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Zhang, Jiahao, Zhang, Lujing, Grimes, Keltin, Yu, Zhuohao, Swamy, Gokul, Wu, Zhiwei Steven
Format:	Preprint
Publié:	2026
Sujets:	Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

All Roads Lead to Likelihood: The Value of Reinforcement Learning in Fine-Tuning
par: Swamy, Gokul, et autres
Publié: (2025)

Multi-Agent Imitation Learning: Value is Easy, Regret is Hard
par: Tang, Jingwu, et autres
Publié: (2024)

Gained in Translation: Privileged Pairwise Judges Enhance Multilingual Reasoning
par: Sutawika, Lintang, et autres
Publié: (2026)

Inverse Reinforcement Learning without Reinforcement Learning
par: Swamy, Gokul, et autres
Publié: (2023)

Gone but Not Forgotten: Improved Benchmarks for Machine Unlearning
par: Grimes, Keltin, et autres
Publié: (2024)

From Curiosity to Caution: Mitigating Reward Hacking for Best-of-N with Pessimism
par: Yu, Zhuohao, et autres
Publié: (2026)

From Foresight to Forethought: VLM-In-the-Loop Policy Steering via Latent Alignment
par: Wu, Yilin, et autres
Publié: (2025)

A Minimaximalist Approach to Reinforcement Learning from Human Feedback
par: Swamy, Gokul, et autres
Publié: (2024)

Concept-ROT: Poisoning Concepts in Large Language Models with Model Editing
par: Grimes, Keltin, et autres
Publié: (2024)

The Virtues of Pessimism in Inverse Reinforcement Learning
par: Wu, David, et autres
Publié: (2024)

The Importance of Online Data: Understanding Preference Fine-tuning via Coverage
par: Song, Yuda, et autres
Publié: (2024)

From Firewalls to Frontiers: AI Red-Teaming is a Domain-Specific Evolution of Cyber Red-Teaming
par: Sinha, Anusha, et autres
Publié: (2025)

Closed-Loop Supervised Fine-Tuning of Tokenized Traffic Models
par: Zhang, Zhejun, et autres
Publié: (2024)

Hybrid Inverse Reinforcement Learning
par: Ren, Juntao, et autres
Publié: (2024)

ShakyPrepend: A Multi-Group Learner with Improved Sample Complexity
par: Zhang, Lujing, et autres
Publié: (2026)

Fair Risk Control: A Generalized Framework for Calibrating Multi-group Fairness Risks
par: Zhang, Lujing, et autres
Publié: (2024)

Dimension-Free Decision Calibration for Nonlinear Loss Functions
par: Tang, Jingwu, et autres
Publié: (2025)

From Instance Selection to Fixed-Pool Data Recipe Search for Supervised Fine-Tuning
par: Wu, Haodong, et autres
Publié: (2026)

Diffusing States and Matching Scores: A New Framework for Imitation Learning
par: Wu, Runzhe, et autres
Publié: (2024)

Multi-agent decision making: A Blackwell's informativeness approach
par: Zhang, Zheng, et autres
Publié: (2026)

RoaD: Rollouts as Demonstrations for Closed-Loop Supervised Fine-Tuning of Autonomous Driving Policies
par: Garcia-Cobo, Guillermo, et autres
Publié: (2025)

Preference-based Multi-Objective Reinforcement Learning
par: Mu, Ni, et autres
Publié: (2025)

Objective Matters: Fine-Tuning Objectives Shape Safety, Robustness, and Persona Drift
par: Vennemeyer, Daniel, et autres
Publié: (2026)

Efficient Imitation under Misspecification
par: Espinosa-Dice, Nicolas, et autres
Publié: (2025)

A Generalized Model for Multidimensional Intransitivity
par: Duan, Jiuding, et autres
Publié: (2024)

Reducing Blackwell and Average Optimality to Discounted MDPs via the Blackwell Discount Factor
par: Grand-Clément, Julien, et autres
Publié: (2023)

Multi-Objective Reward and Preference Optimization: Theory and Algorithms
par: Agnihotri, Akhil
Publié: (2025)

Intransitive Player Dominance and Market Inefficiency in Tennis Forecasting: A Graph Neural Network Approach
par: Clegg, Lawrence, et autres
Publié: (2025)

COS-DPO: Conditioned One-Shot Multi-Objective Fine-Tuning Framework
par: Ren, Yinuo, et autres
Publié: (2024)

Interpretable Preferences via Multi-Objective Reward Modeling and Mixture-of-Experts
par: Wang, Haoxiang, et autres
Publié: (2024)

Beyond One-Preference-Fits-All Alignment: Multi-Objective Direct Preference Optimization
par: Zhou, Zhanhui, et autres
Publié: (2023)

Closed-Loop Koopman Operator Approximation
par: Dahdah, Steven, et autres
Publié: (2023)

Multi-Task LLM with LoRA Fine-Tuning for Automated Cancer Staging and Biomarker Extraction
par: Shao, Jiahao, et autres
Publié: (2026)

Preference Fine-Tuning of LLMs Should Leverage Suboptimal, On-Policy Data
par: Tajwar, Fahim, et autres
Publié: (2024)

AutoFT: Learning an Objective for Robust Fine-Tuning
par: Choi, Caroline, et autres
Publié: (2024)

LoopTool: Closing the Data-Training Loop for Robust LLM Tool Calls
par: Zhang, Kangning, et autres
Publié: (2025)

EMORL: Ensemble Multi-Objective Reinforcement Learning for Efficient and Flexible LLM Fine-Tuning
par: Kong, Lingxiao, et autres
Publié: (2025)

Selection of LLM Fine-Tuning Data based on Orthogonal Rules
par: Li, Xiaomin, et autres
Publié: (2024)

Your Learned Constraint is Secretly a Backward Reachable Tube
par: Qadri, Mohamad, et autres
Publié: (2025)

When and How Human Curation Backfires: Preference Alignment under Multi-Model Self-Consuming Loop
par: Zhang, Yang, et autres
Publié: (2026)