Paul Weng

TMLR Journal 2025 Journal Article

A Survey of Reinforcement Learning from Human Feedback

• Timo Kaufmann
• Paul Weng
• Viktor Bengs
• Eyke Hüllermeier

Reinforcement learning from human feedback (RLHF) is a variant of reinforcement learning (RL) that learns from human feedback instead of relying on an engineered reward function. Building on prior work on the related setting of preference-based reinforcement learning (PbRL), it stands at the intersection of artificial intelligence and human-computer interaction. This positioning provides a promising approach to enhance the performance and adaptability of intelligent systems while also improving the alignment of their objectives with human values. The success in training large language models (LLMs) has impressively demonstrated this potential in recent years, where RLHF has played a decisive role in directing the model's capabilities towards human objectives. This article provides an overview of the fundamentals of RLHF, exploring how RL agents interact with human feedback. While recent focus has been on RLHF for LLMs, our survey covers the technique across multiple domains. We provide our most comprehensive coverage in control and robotics, where many fundamental techniques originate, alongside a dedicated LLM section. We examine the core principles that underpin RLHF, how algorithms and human feedback work together, and discuss the main research trends in the field. Our goal is to give researchers and practitioners a clear understanding of this rapidly growing field.

ICML Conference 2025 Conference Paper

Comparing Comparisons: Informative and Easy Human Feedback with Distinguishability Queries

• Xuening Feng
• Zhao-Hui Jiang
• Timo Kaufmann
• Eyke Hüllermeier
• Paul Weng
• Yifei Zhu

Learning human objectives from preference feedback has significantly advanced reinforcement learning (RL) in domains where objectives are hard to formalize. However, traditional methods based on pairwise trajectory comparisons face notable challenges, including the difficulty in comparing trajectories with subtle differences and the limitation of conveying only ordinal information, limiting direct inference of preference strength. In this paper, we introduce a novel distinguishability query, enabling humans to express preference strength by comparing two pairs of trajectories. Labelers first indicate which of two pairs is easier to distinguish, then provide preference feedback only on the easier pair. Our proposed query type directly captures preference strength and is expected to reduce the cognitive load on the labeler. We further connect this query to cardinal utility and difference relations and develop an efficient query selection scheme to achieve a better trade-off between query informativeness and easiness. Experimental results demonstrate the potential of our method for faster, data-efficient learning and improved user-friendliness in RLHF benchmarks, particularly in classical control settings where preference strength is critical for expected utility maximization.

AAAI Conference 2025 Conference Paper

DUO: Diverse, Uncertain, On-Policy Query Generation and Selection for Reinforcement Learning from Human Feedback

• Xuening Feng
• Zhaohui Jiang
• Timo Kaufmann
• Puchen Xu
• Eyke Hüllermeier
• Paul Weng
• Yifei Zhu

Defining a reward function is usually a challenging but critical task for the system designer in reinforcement learning, especially when specifying complex behaviors. Reinforcement learning from human feedback (RLHF) emerges as a promising approach to circumvent this. In RLHF, the agent typically learns a reward function by querying a human teacher using pairwise comparisons of trajectory segments. A key question in this domain is how to reduce the number of queries necessary to learn an informative reward function since asking a human teacher too many queries is impractical and costly. To tackle this question, we propose DUO, a novel method for diverse, uncertain, on-policy query generation and selection in RLHF. Our method produces queries that are (1) more relevant for policy training (via an on-policy criterion), (2) more informative (via a principled measure of epistemic uncertainty), and (3) diverse (via a clustering-based filter). Experimental results on a variety of locomotion and robotic manipulation tasks demonstrate that our method can outperform state-of-the-art RLHF methods given the same total budget of queries, while being robust to possibly irrational teachers.

AAAI Conference 2025 Conference Paper

Enhancing Online Reinforcement Learning with Meta-Learned Objective from Offline Data

• Shilong Deng
• Zetao Zheng
• Hongcai He
• Paul Weng
• Jie Shao

A major challenge in Reinforcement Learning (RL) is the difficulty of learning an optimal policy from sparse rewards. Prior works enhance online RL with conventional Imitation Learning (IL) via a handcrafted auxiliary objective, at the cost of restricting the RL policy to be sub-optimal when the offline data is generated by a non-expert policy. Instead, to better leverage valuable information in offline data, we develop Generalized Imitation Learning from Demonstration (GILD), which meta-learns an objective that distills knowledge from offline data and instills intrinsic motivation towards the optimal policy. Distinct from prior works that are exclusive to a specific RL algorithm, GILD is a flexible module intended for diverse vanilla off-policy RL algorithms. In addition, GILD introduces no domain-specific hyperparameter and minimal increase in computational cost. In four challenging MuJoCo tasks with sparse rewards, we show that three RL algorithms enhanced with GILD significantly outperform state-of-the-art methods.

ICLR Conference 2025 Conference Paper

Reinforcement Learning from Imperfect Corrective Actions and Proxy Rewards

• Zhao-Hui Jiang
• Xuening Feng
• Paul Weng
• Yifei Zhu
• Yan Song
• Tianze Zhou
• Yujing Hu
• Tangjie Lv

In practice, reinforcement learning (RL) agents are often trained with a possibly imperfect proxy reward function, which may lead to a human-agent alignment issue (i.e., the learned policy either converges to non-optimal performance with low cumulative rewards, or achieves high cumulative rewards but in an undesired manner). To tackle this issue, we consider a framework where a human labeler can provide additional feedback in the form of corrective actions, which expresses the labeler's action preferences although this feedback may possibly be imperfect as well. In this setting, to obtain a better-aligned policy guided by both learning signals, we propose a novel value-based deep RL algorithm called **I**terative learning from **Co**rrective actions and **Pro**xy rewards (ICoPro), which cycles through three phases: (1) Solicit sparse corrective actions from a human labeler on the agent's demonstrated trajectories; (2) Incorporate these corrective actions into the Q-function using a margin loss to enforce adherence to labeler's preferences; (3) Train the agent with standard RL losses regularized with a margin loss to learn from proxy rewards and propagate the Q-values learned from human feedback. Moreover, another novel design in our approach is to integrate pseudo-labels from the target Q-network to reduce human labor and further stabilize training. We experimentally validate our proposition on a variety of tasks (Atari games and autonomous driving on highway). On the one hand, using proxy rewards with different levels of imperfection, our method can better align with human and is more sample-efficient than baseline methods. On the other hand, facing corrective actions with different types of imperfection, our method can overcome the non-optimality of this feedback thanks to the guidance from proxy rewards.

NeurIPS Conference 2025 Conference Paper

Time Reversal Symmetry for Efficient Robotic Manipulations in Deep Reinforcement Learning

• Yunpeng Jiang
• Jianshu Hu
• Paul Weng
• Yutong Ban

Symmetry is pervasive in robotics and has been widely exploited to improve sample efficiency in deep reinforcement learning (DRL). However, existing approaches primarily focus on spatial symmetries—such as reflection, rotation, and translation—while largely neglecting temporal symmetries. To address this gap, we explore time reversal symmetry, a form of temporal symmetry commonly found in robotics tasks such as door opening and closing. We propose Time Reversal symmetry enhanced Deep Reinforcement Learning (TR-DRL), a framework that combines trajectory reversal augmentation and time reversal guided reward shaping to efficiently solve temporally symmetric tasks. Our method generates reversed transitions from fully reversible transitions, identified by a proposed dynamics-consistent filter, to augment the training data. For partially reversible transitions, we apply reward shaping to guide learning, according to successful trajectories from the reversed task. Extensive experiments on the Robosuite and MetaWorld benchmarks demonstrate that TR-DRL is effective in both single-task and multi-task settings, achieving higher sample efficiency and stronger final performance compared to baseline methods.

TMLR Journal 2025 Journal Article

Understanding and Reducing the Class-Dependent Effects of Data Augmentation with A Two-Player Game Approach

• Yunpeng Jiang
• Yutong Ban
• Paul Weng

Data augmentation is widely applied and has shown its benefits in different machine learning tasks. However, as recently observed, it may have an unfair effect in multi-class classification. While data augmentation generally improves the overall performance (and therefore is beneficial for many classes), it can actually be detrimental for other classes, which can be problematic in some application domains. In this paper, to counteract this phenomenon, we propose CLAM, a CLAss-dependent Multiplicative-weights method. To derive it, we first formulate the training of a classifier as a non-linear optimization problem that aims at simultaneously maximizing the individual class performances and balancing them. By rewriting this optimization problem as an adversarial two-player game, we propose a novel multiplicative weight algorithm, for which we prove the convergence. Interestingly, our formulation also reveals that the class-dependent effects of data augmentation is not due to data augmentation only, but is in fact a general phenomenon. Our empirical results over five datasets demonstrate that the performance of learned classifiers is indeed more fairly distributed over classes, with only limited impact on the average accuracy.

ICML Conference 2024 Conference Paper

INViT: A Generalizable Routing Problem Solver with Invariant Nested View Transformer

• Han Fang
• Zhihao Song
• Paul Weng
• Yutong Ban

Recently, deep reinforcement learning has shown promising results for learning fast heuristics to solve routing problems. Meanwhile, most of the solvers suffer from generalizing to an unseen distribution or distributions with different scales. To address this issue, we propose a novel architecture, called Invariant Nested View Transformer (INViT), which is designed to enforce a nested design together with invariant views inside the encoders to promote the generalizability of the learned solver. It applies a modified policy gradient algorithm enhanced with data augmentations. We demonstrate that the proposed INViT achieves a dominant generalization performance on both TSP and CVRP problems with various distributions and different problem scales. Our source code and datasets are available in supplementary materials.

ICLR Conference 2024 Conference Paper

Revisiting Data Augmentation in Deep Reinforcement Learning

• Jianshu Hu
• Yunpeng Jiang
• Paul Weng

Various data augmentation techniques have been recently proposed in image-based deep reinforcement learning (DRL). Although they empirically demonstrate the effectiveness of data augmentation for improving sample efficiency or generalization, which technique should be preferred is not always clear. To tackle this question, we analyze existing methods to better understand them and to uncover how they are connected. Notably, by expressing the variance of the Q-targets and that of the empirical actor/critic losses of these methods, we can analyze the effects of their different components and compare them. We furthermore formulate an explanation about how these methods may be affected by choosing different data augmentation transformations in calculating the target Q-values. This analysis suggests recommendations on how to exploit data augmentation in a more principled way. In addition, we include a regularization term called tangent prop, previously proposed in computer vision, but whose adaptation to DRL is novel to the best of our knowledge. We evaluate our proposition and validate our analysis in several domains. Compared to different relevant baselines, we demonstrate that it achieves state-of-the-art performance in most environments and shows higher sample efficiency and better generalization ability in some complex environments.

TMLR Journal 2023 Journal Article

Differentiable Logic Machines

• Matthieu Zimmer
• Xuening Feng
• Claire Glanois
• Zhaohui Jiang
• Jianyi Zhang
• Paul Weng
• Dong Li
• Jianye Hao

The integration of reasoning, learning, and decision-making is key to build more general artificial intelligence systems. As a step in this direction, we propose a novel neural-logic architecture, called differentiable logic machine (DLM), that can solve both inductive logic programming (ILP) and reinforcement learning (RL) problems, where the solution can be interpreted as a first-order logic program. Our proposition includes several innovations. Firstly, our architecture defines a restricted but expressive continuous relaxation of the space of first-order logic programs by assigning weights to predicates instead of rules, in contrast to most previous neural-logic approaches. Secondly, with this differentiable architecture, we propose several (supervised and RL) training procedures, based on gradient descent, which can recover a fully-interpretable solution (i.e., logic formula). Thirdly, to accelerate RL training, we also design a novel critic architecture that enables actor-critic algorithms. Fourthly, to solve hard problems, we propose an incremental training procedure that can learn a logic program progressively. Compared to state-of-the-art (SOTA) differentiable ILP methods, DLM successfully solves all the considered ILP problems with a higher percentage of successful seeds (up to 3.5x). On RL problems, without requiring an interpretable solution, DLM outperforms other non-interpretable neural-logic RL approaches in terms of rewards (up to 3.9%). When enforcing interpretability, DLM can solve harder RL problems (e.g., Sorting, Path) than other interpretable RL methods. Moreover, we show that deep logic programs can be learned via incremental supervised training. In addition to this excellent performance, DLM can scale well in terms of memory and computational time, especially during the testing phase where it can deal with much more constants (>2x) than SOTA.

ECAI Conference 2023 Conference Paper

Fair Deep Reinforcement Learning with Preferential Treatment

• Guanbao Yu
• Umer Siddique
• Paul Weng

Learning fair policies in reinforcement learning (RL) is important when the RL agent may impact many users. We investigate a variant of this problem where equity is still desired, but some users may be entitled to preferential treatment. In this paper, we formalize this more sophisticated fair optimization problem in deep RL using generalized fair social welfare functions (SWF), provide a theoretical discussion to justify our approach, explain how deep RL algorithms can be adapted to tackle it, and empirically validate our propositions on several domains. Our contributions are both theoretical and algorithmic, notably: (1) We obtain a general bound on the suboptimality gap in terms of SWF-optimality using average reward of a policy SWF-optimal for the discounted reward, which notably justifies using standard deep RL algorithms, even for the average reward; (2) Our algorithmic innovations include a state-augmented DQN-based method for learning either deterministic or stochastic policies, which also applies to the usual fair optimization setting without any preferential treatment.

AAMAS Conference 2023 Conference Paper

Learning Rewards to Optimize Global Performance Metrics in Deep Reinforcement Learning

• Junqi Qian
• Paul Weng
• Chenmien Tan

When applying reinforcement learning (RL) to a new problem, reward engineering is a necessary, but often difficult and errorprone task a system designer has to face. To avoid this step, we propose LR4GPM, a novel (deep) RL method that can optimize a global performance metric, which is supposed to be available as part of the problem description. LR4GPM alternates between two phases: (1) learning a (possibly vector) reward function used to fit the performance metric, and (2) training a policy to optimize an approximation of this performance metric based on the learned rewards. Such RL training is not straightforward since both the reward function and the policy are trained using non-stationary data. To overcome this issue, we propose several training tricks. We demonstrate the efficiency of LR4GPM on several domains. Notably, LR4GPM outperforms the winner of a recent autonomous driving competition organized at DAI’2020.

ICML Conference 2022 Conference Paper

Neuro-Symbolic Hierarchical Rule Induction

• Claire Glanois
• Zhao-Hui Jiang
• Xuening Feng
• Paul Weng
• Matthieu Zimmer
• Dong Li 0016
• Wulong Liu
• Jianye Hao

We propose Neuro-Symbolic Hierarchical Rule Induction, an efficient interpretable neuro-symbolic model, to solve Inductive Logic Programming (ILP) problems. In this model, which is built from a pre-defined set of meta-rules organized in a hierarchical structure, first-order rules are invented by learning embeddings to match facts and body predicates of a meta-rule. To instantiate, we specifically design an expressive set of generic meta-rules, and demonstrate they generate a consequent fragment of Horn clauses. As a differentiable model, HRI can be trained both via supervised learning and reinforcement learning. To converge to interpretable rules, we inject a controlled noise to avoid local optima and employ an interpretability-regularization term. We empirically validate our model on various tasks (ILP, visual genome, reinforcement learning) against relevant state-of-the-art methods, including traditional ILP methods and neuro-symbolic models.

ICML Conference 2021 Conference Paper

Learning Fair Policies in Decentralized Cooperative Multi-Agent Reinforcement Learning

• Matthieu Zimmer
• Claire Glanois
• Umer Siddique
• Paul Weng

We consider the problem of learning fair policies in (deep) cooperative multi-agent reinforcement learning (MARL). We formalize it in a principled way as the problem of optimizing a welfare function that explicitly encodes two important aspects of fairness: efficiency and equity. We provide a theoretical analysis of the convergence of policy gradient for this problem. As a solution method, we propose a novel neural network architecture, which is composed of two sub-networks specifically designed for taking into account these two aspects of fairness. In experiments, we demonstrate the importance of the two sub-networks for fair optimization. Our overall approach is general as it can accommodate any (sub)differentiable welfare function. Therefore, it is compatible with various notions of fairness that have been proposed in the literature (e. g. , lexicographic maximin, generalized Gini social welfare function, proportional fairness). Our method is generic and can be implemented in various MARL settings: centralized training and decentralized execution, or fully decentralized. Finally, we experimentally validate our approach in various domains and show that it can perform much better than previous methods, both in terms of efficiency and equity.

ICML Conference 2020 Conference Paper

Learning Fair Policies in Multi-Objective (Deep) Reinforcement Learning with Average and Discounted Rewards

• Umer Siddique
• Paul Weng
• Matthieu Zimmer

As the operations of autonomous systems generally affect simultaneously several users, it is crucial that their designs account for fairness considerations. In contrast to standard (deep) reinforcement learning (RL), we investigate the problem of learning a policy that treats its users equitably. In this paper, we formulate this novel RL problem, in which an objective function, which encodes a notion of fairness that we formally define, is optimized. For this problem, we provide a theoretical discussion where we examine the case of discounted rewards and that of average rewards. During this analysis, we notably derive a new result in the standard RL setting, which is of independent interest: it states a novel bound on the approximation error with respect to the optimal average reward of that of a policy optimal for the discounted reward. Since learning with discounted rewards is generally easier, this discussion further justifies finding a fair policy for the average reward by learning a fair policy for the discounted reward. Thus, we describe how several classic deep RL algorithms can be adapted to our fair optimization problem, and we validate our approach with extensive experiments in three different domains.

IJCAI Conference 2019 Conference Paper

Exploiting the Sign of the Advantage Function to Learn Deterministic Policies in Continuous Domains

• Matthieu Zimmer
• Paul Weng

In the context of learning deterministic policies in continuous domains, we revisit an approach, which was first proposed in Continuous Actor Critic Learning Automaton (CACLA) and later extended in Neural Fitted Actor Critic (NFAC). This approach is based on a policy update different from that of deterministic policy gradient (DPG). Previous work has observed its excellent performance empirically, but a theoretical justification is lacking. To fill this gap, we provide a theoretical explanation to motivate this unorthodox policy update by relating it to another update and making explicit the objective function of the latter. We furthermore discuss in depth the properties of these updates to get a deeper understanding of the overall approach. In addition, we extend it and propose a new trust region algorithm, Penalized NFAC (PeNFAC). Finally, we experimentally demonstrate in several classic control problems that it surpasses the state-of-the-art algorithms to learn deterministic policies.

ICML Conference 2017 Conference Paper

Multi-objective Bandits: Optimizing the Generalized Gini Index

• Róbert Busa-Fekete
• Balázs Szörényi
• Paul Weng
• Shie Mannor

We study the multi-armed bandit (MAB) problem where the agent receives a vectorial feedback that encodes many possibly competing objectives to be optimized. The goal of the agent is to find a policy, which can optimize these objectives simultaneously in a fair way. This multi-objective online optimization problem is formalized by using the Generalized Gini Index (GGI) aggregation function. We propose an online gradient descent algorithm which exploits the convexity of the GGI aggregation function, and controls the exploration in a careful way achieving a distribution-free regret $\tilde{O}(T^{-1/2} )$ with high probability. We test our algorithm on synthetic data as well as on an electric battery control problem where the goal is to trade off the use of the different cells of a battery in order to balance their respective degradation rates.

AAAI Conference 2017 Conference Paper

Optimizing Quantiles in Preference-Based Markov Decision Processes

• Hugo Gilbert
• Paul Weng
• Yan Xu

In the Markov decision process model, policies are usually evaluated by expected cumulative rewards. As this decision criterion is not always suitable, we propose in this paper an algorithm for computing a policy optimal for the quantile criterion. Both finite and infinite horizons are considered. Finally we experimentally evaluate our approach on random MDPs and on a data center control problem.

UAI Conference 2016 Conference Paper

Model-Free Reinforcement Learning with Skew-Symmetric Bilinear Utilities

• Hugo Gilbert
• Bruno Zanuttini
• Paul Weng
• Paolo Viappiani
• Esther Nicart

In reinforcement learning, policies are typically evaluated according to the expectation of cumulated rewards. Researchers in decision theory have argued that more sophisticated decision criteria can better model the preferences of a decision maker. In particular, Skew-Symmetric Bilinear (SSB) utility functions generalize von Neumann and Morgenstern’s expected utility (EU) theory to encompass rational decision behaviors that EU cannot accommodate. In this paper, we adopt an SSB utility function to compare policies in the reinforcement learning setting. We provide a model-free SSB reinforcement learning algorithm, SSB Q-learning, and prove its convergence towards a policy that is -optimal according to SSB. The proposed algorithm is an adaptation of fictitious play [Brown, 1951] combined with techniques from stochastic approximation [Borkar, 1997]. We also present some experimental results which evaluate our approach in a variety of settings.

IJCAI Conference 2015 Conference Paper

Optimization of Probabilistic Argumentation with Markov Decision Models

• Emmanuel Hadoux
• Aur
• eacute; lie Beynier
• Nicolas Maudet
• Paul Weng
• Anthony Hunter

One prominent way to deal with conflicting viewpoints among agents is to conduct an argumentative debate: by exchanging arguments, agents can seek to persuade each other. In this paper we investigate the problem, for an agent, of optimizing a sequence of moves to be put forward in a debate, against an opponent assumed to behave stochastically, and equipped with an unknown initial belief state. Despite the prohibitive number of states induced by a naive mapping to Markov models, we show that exploiting several features of such interaction settings allows for optimal resolution in practice, in particular: (1) as debates take place in a public space (or common ground), they can readily be modelled as Mixed Observability Markov Decision Processes, (2) as argumentation problems are highly structured, one can design optimization techniques to prune the initial instance. We report on the experimental evaluation of these techniques.

ICML Conference 2015 Conference Paper

Qualitative Multi-Armed Bandits: A Quantile-Based Approach

• Balázs Szörényi
• Róbert Busa-Fekete
• Paul Weng
• Eyke Hüllermeier

We formalize and study the multi-armed bandit (MAB) problem in a generalized stochastic setting, in which rewards are not assumed to be numerical. Instead, rewards are measured on a qualitative scale that allows for comparison but invalidates arithmetic operations such as averaging. Correspondingly, instead of characterizing an arm in terms of the mean of the underlying distribution, we opt for using a quantile of that distribution as a representative value. We address the problem of quantile-based online learning both for the case of a finite (pure exploration) and infinite time horizon (cumulative regret minimization). For both cases, we propose suitable algorithms and analyze their properties. These properties are also illustrated by means of first experimental studies.

IJCAI Conference 2015 Conference Paper

Solving MDPs with Skew Symmetric Bilinear Utility Functions

• Hugo Gilbert
• Olivier Spanjaard
• Paolo Viappiani
• Paul Weng

In this paper we adopt Skew Symmetric Bilinear (SSB) utility functions to compare policies in Markov Decision Processes (MDPs). By considering pairs of alternatives, SSB utility theory generalizes von Neumann and Morgenstern’s expected utility (EU) theory to encompass rational decision behaviors that EU cannot accommodate. We provide a game-theoretic analysis of the problem of identifying an SSB-optimal policy in finite horizon MDPs and propose an algorithm based on a double oracle approach for computing an optimal (possibly randomized) policy. Finally, we present and discuss experimental results where SSB-optimal policies are computed for a popular TV contest according to several instantiations of SSB utility functions.

UAI Conference 2013 Conference Paper

Approximation of Lorenz-Optimal Solutions in Multiobjective Markov Decision Processes

• Patrice Perny
• Paul Weng
• Judy Goldsmith
• Josiah P. Hanna

This paper is devoted to fair optimization in Multiobjective Markov Decision Processes (MOMDPs). A MOMDP is an extension of the MDP model for planning under uncertainty while trying to optimize several reward functions simultaneously. This applies to multiagent problems when rewards define individual utility functions, or in multicriteria problems when rewards refer to different features. In this setting, we study the determination of policies leading to Lorenz-nondominated tradeoffs. Lorenz dominance is a refinement of Pareto dominance that was introduced in Social Choice for the measurement of inequalities. In this paper, we introduce methods to efficiently approximate the sets of Lorenz-non-dominated solutions of infinite-horizon, discounted MOMDPs. The approximations are polynomial-sized subsets of those solutions.

IJCAI Conference 2013 Conference Paper

Interactive Value Iteration for Markov Decision Processes with Unknown Rewards

• Paul Weng
• Bruno Zanuttini

To tackle the potentially hard task of defining the reward function in a Markov Decision Process, we propose a new approach, based on Value Iteration, which interweaves the elicitation and optimization phases. We assume that rewards whose numeric values are unknown can only be ordered, and that a tutor is present to help comparing sequences of rewards. We first show how the set of possible reward functions for a given preference relation can be represented as a polytope. Then our algorithm, called Interactive Value Iteration, searches for an optimal policy while refining its knowledge about the possible reward functions, by querying a tutor when necessary. We prove that the number of queries needed before finding an optimal policy is upperbounded by a polynomial in the size of the problem, and we present experimental results which demonstrate that our approach is efficient in practice.

EWRL Workshop 2013 Workshop Paper

Preference-based Evolutionary Direct Policy Search

• Robert Busa-Fekete
• Balazs Szorenyi
• Paul Weng
• Weiwei Cheng
• Eyke Hullermeier

ICML Conference 2013 Conference Paper

Top-k Selection based on Adaptive Sampling of Noisy Preferences

• Róbert Busa-Fekete
• Balázs Szörényi
• Weiwei Cheng
• Paul Weng
• Eyke Hüllermeier

We consider the problem of reliably selecting an optimal subset of fixed size from a given set of choice alternatives, based on noisy information about the quality of these alternatives. Problems of similar kind have been tackled by means of adaptive sampling schemes called racing algorithms. However, in contrast to existing approaches, we do not assume that each alternative is characterized by a real-valued random variable, and that samples are taken from the corresponding distributions. Instead, we only assume that alternatives can be compared in terms of pairwise preferences. We propose and formally analyze a general preference-based racing algorithm that we instantiate with three specific ranking procedures and corresponding sampling schemes. Experiments with real and synthetic data are presented to show the efficiency of our approach.

ECAI Conference 2012 Conference Paper

Ordinal Decision Models for Markov Decision Processes

• Paul Weng

Setting the values of rewards in Markov decision processes (MDP) may be a difficult task. In this paper, we consider two ordinal decision models for MDPs where only an order is known over rewards. The first one, which has been proposed recently in MDPs [23], defines preferences with respect to a reference point. The second model, which can been viewed as the dual approach of the first one, is based on quantiles. Based on the first decision model, we give a new interpretation of rewards in standard MDPs, which sheds some interesting light on the preference system used in standard MDPs. The second model based on quantile optimization is a new approach in MDPs with ordinal rewards. Although quantile-based optimality is state-dependent, we prove that an optimal stationary deterministic policy exists for a given initial state. Finally, we propose solution methods based on linear programming for optimizing quantiles.

ICAPS Conference 2011 Conference Paper

Markov Decision Processes with Ordinal Rewards: Reference Point-Based Preferences

• Paul Weng

In a standard Markov decision process (MDP), rewards are assumed to be precisely known and of quantitative nature. This can be a too strong hypothesis in some situations. When rewards can really be modeled numerically, specifying the reward function is often difficult as it is a cognitively-demanding and/or time-consuming task. Besides, rewards can sometimes be of qualitative nature as when they represent qualitative risk levels for instance. In those cases, it is problematic to use directly standard MDPs and we propose instead to resort to MDPs with ordinal rewards. Only a total order over rewards is assumed to be known. In this setting, we explain how an alternative way to define expressive and interpretable preferences using reference points can be exploited.

ECAI Conference 2010 Conference Paper

On Finding Compromise Solutions in Multiobjective Markov Decision Processes

• Patrice Perny
• Paul Weng

ECAI Conference 2006 Conference Paper

An Axiomatic Approach in Qualitative Decision Theory with Binary Possibilistic Utility

• Paul Weng

Binary possibilistic utility unifies two previously proposed qualitative decision models: optimistic and pessimistic utilities. All these decision models have been axiomatized in a von Neumann-Morgenstern setting. These axiomatizations have shown the formal similarity of these qualitative utilities and expected utility. Unfortunately in this framework, the representation of uncertainty has to be clearly assumed to be given. In a more general setting, without this restriction, optimistic and pessimistic utilities have been axiomatized à la Savage. This paper proposes a study of the axiomatics of binary possibilistic utility in a Savagean framework.

UAI Conference 2006 Conference Paper

Axiomatic Foundations for a Class of Generalized Expected Utility: Algebraic Expected Utility

• Paul Weng

UAI Conference 2005 Conference Paper

Qualitative Decision Making Under Possibilistic Uncertainty: Toward more Discriminating Criteria

• Paul Weng