paper-with-me

홈 › Papers

Quantifying Differences in Reward Functions

2020-06-24 · ICLR 2021 1 · Adam Gleave, Michael Dennis, Shane Legg, Stuart Russell, Jan Leike

For many tasks, the reward function is inaccessible to introspection or too complex to be specified procedurally, and must instead be learned from user data. Prior work has evaluated learned reward functions by evaluating policies optimized for the learned reward. However, this method cannot distinguish between the learned reward function failing to reflect user preferences and the policy optimization process failing to optimize the learned reward. Moreover, this method can only tell us about behavior in the evaluation environment, but the reward may incentivize very different behavior in even a slightly different deployment environment. To address these problems, we introduce the Equivalent-Policy Invariant Comparison (EPIC) distance to quantify the difference between two reward functions directly, without a policy optimization step. We prove EPIC is invariant on an equivalence class of reward functions that always induce the same optimal policy. Furthermore, we find EPIC can be efficiently approximated and is more robust than baselines to the choice of coverage distribution. Finally, we show that EPIC distance bounds the regret of optimal policies even under different transition dynamics, and we confirm empirically that it predicts policy training success. Our source code is available at https://github.com/HumanCompatibleAI/evaluating-rewards.

📄 PDF Abstract BibTeX arXiv:2006.13900

Code (1)

HumanCompatibleAI/evaluating-rewards 공식 구현

Similar Papers 제목 키워드 기반

STARC: A General Framework For Quantifying Differences Between Reward Functions

2023-09-26 · Joar Skalse, Lucy Farnik, Sumeet Ramesh Motwani, Erik Jenner 외

In order to solve a task using reinforcement learning, it is necessary to first formalise the goal of that task as a reward function. However, for many real-world tasks, it is very difficult to manually specify a reward …

Understanding Reward Ambiguity Through Optimal Transport Theory in Inverse Reinforcement Learning

2023-10-18 · Ali Baheri

In inverse reinforcement learning (IRL), the central objective is to infer underlying reward functions from observed expert behaviors in a way that not only explains the given data but also generalizes to unseen scenario…

reinforcement-learning

Active Preference-Based Gaussian Process Regression for Reward Learning

2020-05-06 · Erdem Biyik, Nicolas Huynh, Mykel J. Kochenderfer, Dorsa Sadigh

Designing reward functions is a challenging problem in AI and robotics. Humans usually have a difficult time directly specifying all the desirable behaviors that a robot needs to optimize. One common approach is to learn…

regression

Understanding Impact of Human Feedback via Influence Functions

2025-01-10 · Taywon Min, Haeone Lee, Yongchan Kwon, Kimin Lee

In Reinforcement Learning from Human Feedback (RLHF), it is crucial to learn suitable reward models from human feedback to align large language models (LLMs) with human intentions. However, human feedback can often be no…

Assessment of Reward Functions for Reinforcement Learning Traffic Signal Control under Real-World Limitations

2020-08-26 · Alvaro Cabrejas-Egea, Shaun Howell, Maksis Knutins, Colm Connaughton

Adaptive traffic signal control is one key avenue for mitigating the growing consequences of traffic congestion. Incumbent solutions such as SCOOT and SCATS require regular and time-consuming calibration, can't optimise …

Deep Reinforcement LearningTraffic Signal Control