English
Related papers

Related papers: Settling the Reward Hypothesis

200 papers

We describe a statistical hypothesis test for the presence of a signal. The test allows the researcher to fix the signal location and/or width a priori, or perform a search to find the signal region that maximizes the signal. The background…

Data Analysis, Statistics and Probability · Physics 2007-05-23 Wolfgang Rolke , Angel Lopez

Rewards typically express desirabilities or preferences over a set of alternatives. Here we propose that rewards can be defined for any probability distribution based on three desiderata, namely that rewards should be real-valued, additive…

Artificial Intelligence · Computer Science 2009-12-31 Pedro A. Ortega , Daniel A. Braun

We study the design of optimal incentives in sequential processes. To do so, we consider a basic and fundamental model in which an agent initiates a value-creating sequential process through costly investment with random success. If…

Theoretical Economics · Economics 2023-11-22 Jens Gudmundsson , Jens Leth Hougaard , Juan D. Moreno-Ternero , Lars Peter Østerdal

We study the theoretical aspects of Reinforced Language Models (RLMs) from a bi-objective optimization perspective. Specifically, we consider the RLMs as a Pareto optimization problem that maximizes the two conflicting objectives, i.e.,…

Machine Learning · Computer Science 2023-11-27 Changhun Lee , Chiehyeon Lim

We characterize the optimal reward functions (scoring rules) that incentivize an agent to acquire information and report it truthfully to the principal. The optimal scoring rules let the agent make a simple binary bet in single-dimensional…

Computer Science and Game Theory · Computer Science 2025-10-03 Jason D. Hartline , Yingkai Li , Liren Shan , Yifan Wu

Reward machines allow the definition of rewards for temporally extended tasks and behaviors. Specifying "informative" reward machines can be challenging. One way to address this is to generate reward machines from a high-level abstract…

Machine Learning · Computer Science 2024-08-16 Giovanni Varricchione , Natasha Alechina , Mehdi Dastani , Brian Logan

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 function that never…

Machine Learning · Computer Science 2024-12-13 Joar Skalse , Lucy Farnik , Sumeet Ramesh Motwani , Erik Jenner , Adam Gleave , Alessandro Abate

Inferring reward functions from human behavior is at the center of value alignment - aligning AI objectives with what we, humans, actually want. But doing so relies on models of how humans behave given their objectives. After decades of…

Machine Learning · Computer Science 2023-10-31 Joey Hong , Kush Bhatia , Anca Dragan

We review the theory of renewal reward processes, which describes renewal processes that have some cost or reward associated with each cycle. We present a new simplified proof of the renewal reward theorem that mimics the proof of the…

Probability · Mathematics 2014-04-23 Maria Vlasiou

We present a novel method for learning a set of disentangled reward functions that sum to the original environment reward and are constrained to be independently obtainable. We define independent obtainability in terms of value functions…

Machine Learning · Computer Science 2019-03-06 Christopher Grimm , Satinder Singh

We consider a renewal-reward process with multivariate rewards. Such a process is constructed from an i.i.d.\ sequence of time periods, to each of which there is associated a multivariate reward vector. The rewards in each time period may…

Probability · Mathematics 2014-08-08 Brendan Patch , Yoni Nazarathy , Thomas Taimre

Reward modeling has emerged as a crucial component in aligning large language models with human values. Significant attention has focused on using reward models as a means for fine-tuning generative models. However, the reward models…

Computation and Language · Computer Science 2026-02-04 Brian Christian , Hannah Rose Kirk , Jessica A. F. Thompson , Christopher Summerfield , Tsvetomira Dumbalska

We introduce a notion of real-valued reward testing for probabilistic processes by extending the traditional nonnegative-reward testing with negative rewards. In this richer testing framework, the may and must preorders turn out to be…

Logic in Computer Science · Computer Science 2011-07-07 Yuxin Deng , Rob van Glabbeek , Matthew Hennessy , Carroll Morgan

Reward design is a critical part of the application of reinforcement learning, the performance of which strongly depends on how well the reward signal frames the goal of the designer and how well the signal assesses progress in reaching…

Machine Learning · Computer Science 2022-08-01 Yixiang Wang , Yujing Hu , Feng Wu , Yingfeng Chen

Solving a reinforcement learning problem typically involves correctly prespecifying the reward signal from which the algorithm learns. Here, we approach the problem of reward signal design by using an evolutionary approach to perform a…

Multiagent Systems · Computer Science 2021-05-19 Rafal Muszynski , Katja Hofmann , Jun Wang

The goal of this article is to give a positive answer to Rockafellar's maximality of the sum conjecture in the linear multi-valued operator case.

Functional Analysis · Mathematics 2007-05-23 M. D. Voisei

We introduce a "high probability" framework for repeated games with incomplete information. In our non-equilibrium setting, players aim to guarantee a certain payoff with high probability, rather than in expected value. We provide a high…

Computer Science and Game Theory · Computer Science 2015-09-30 Payam Delgosha , Amin Gohari , Mohammad Akbarpour

In this paper we introduce a qualitative decision and game theory based on belief (B) and desire (D) rules. We show that a group of agents acts as if it is maximizing achieved joint goals.

Multiagent Systems · Computer Science 2007-05-23 Mehdi Dastani , Leendert van der Torre

We study a problem where a group of agents has to decide how a joint reward should be shared among them. We focus on settings where the share that each agent receives depends on the subjective opinions of its peers concerning that agent's…

Computer Science and Game Theory · Computer Science 2013-05-23 Arthur Carvalho , Kate Larson

What is happiness for reinforcement learning agents? We seek a formal definition satisfying a list of desiderata. Our proposed definition of happiness is the temporal difference error, i.e. the difference between the value of the obtained…

Artificial Intelligence · Computer Science 2015-05-19 Mayank Daswani , Jan Leike