In reinforcement learning, robust policies for high-stakes decision-making problems with limited data are usually computed by optimizing the \emph{percentile criterion}. The percentile criterion is approximately solved by constructing an \emph{ambiguity set} that contains the true model with high probability and optimizing the policy for the worst model in the set. Since the percentile criterion is non-convex, constructing ambiguity sets is often challenging. Existing work uses \emph{Bayesian credible regions} as ambiguity sets, but they are often unnecessarily large and result in learning overly conservative policies. To overcome these shortcomings, we propose a novel Value-at-Risk based dynamic programming algorithm to optimize the percentile criterion without explicitly constructing any ambiguity sets. Our theoretical and empirical results show that our algorithm implicitly constructs much smaller ambiguity sets and learns less conservative robust policies.

Sequential decision-making is desired to align with human intents and exhibit versatility across various tasks. Previous methods formulate it as a conditional generation process, utilizing return-conditioned diffusion models to directly model trajectory distributions. Nevertheless, the return-conditioned paradigm relies on pre-defined reward functions, facing challenges when applied in multi-task settings characterized by varying reward functions (versatility) and showing limited controllability concerning human preferences (alignment). In this work, we adopt multi-task preferences as a unified condition for both single- and multi-task decision-making, and propose preference representations aligned with preference labels. The learned representations are used to guide the conditional generation process of diffusion models, and we introduce an auxiliary objective to maximize the mutual information between representations and corresponding generated trajectories, improving alignment between trajectories and preferences. Extensive experiments in D4RL and Meta-World demonstrate that our method presents favorable performance in single- and multi-task scenarios, and exhibits superior alignment with preferences.

This paper considers the problem of learning a control policy for robot motion planning with zero-shot generalization, i.e., no data collection and policy adaptation is needed when the learned policy is deployed in new environments. We develop a federated reinforcement learning framework that enables collaborative learning of multiple learners and a central server, i.e., the Cloud, without sharing their raw data. In each iteration, each learner uploads its local control policy and the corresponding estimated normalized arrival time to the Cloud, which then computes the global optimum among the learners and broadcasts the optimal policy to the learners. Each learner then selects between its local control policy and that from the Cloud for next iteration. The proposed framework leverages on the derived zero-shot generalization guarantees on arrival time and safety. Theoretical guarantees on almost-sure convergence, almost consensus, Pareto improvement and optimality gap are also provided. Monte Carlo simulation is conducted to evaluate the proposed framework.

As a telecom provider, our company has a critical mission to maintain telecom services even during power outages. To accomplish the mission, it is essential to maintain the power of the telecom base stations. Here we consider a solution where electric vehicles (EVs) directly supply power to base stations by traveling to their locations. The goal is to find EV routes that minimize both the total travel distance of all EVs and the number of downed base stations. In this paper, we formulate this routing problem as a new variant of the Electric Vehicle Routing Problem (EVRP) and propose a solver that combines a rule-based vehicle selector and a reinforcement learning (RL)-based node selector. The rule of the vehicle selector ensures the exact environmental states when the selected EV starts to move. In addition, the node selection by the RL model enables fast route generation, which is critical in emergencies. We evaluate our solver on both synthetic datasets and real datasets. The results show that our solver outperforms baselines in terms of the objective value and computation time. Moreover, we analyze the generalization and scalability of our solver, demonstrating the capability toward unseen settings and large-scale problems. Check also our project page: https://ntt-dkiku.github.io/rl-evrpeps.

Off-policy Evaluation (OPE) methods are a crucial tool for evaluating policies in high-stakes domains such as healthcare, where exploration is often infeasible, unethical, or expensive. However, the extent to which such methods can be trusted under adversarial threats to data quality is largely unexplored. In this work, we make the first attempt at investigating the sensitivity of OPE methods to marginal adversarial perturbations to the data. We design a generic data poisoning attack framework leveraging influence functions from robust statistics to carefully construct perturbations that maximize error in the policy value estimates. We carry out extensive experimentation with multiple healthcare and control datasets. Our results demonstrate that many existing OPE methods are highly prone to generating value estimates with large errors when subject to data poisoning attacks, even for small adversarial perturbations. These findings question the reliability of policy values derived using OPE methods and motivate the need for developing OPE methods that are statistically robust to train-time data poisoning attacks.

Offline reinforcement learning (RL) is a compelling framework for learning optimal policies from past experiences without additional interaction with the environment. Nevertheless, offline RL inevitably faces the problem of distributional shifts, where the states and actions encountered during policy execution may not be in the training dataset distribution. A common solution involves incorporating conservatism into the policy or the value function to safeguard against uncertainties and unknowns. In this work, we focus on achieving the same objectives of conservatism but from a different perspective. We propose COmpositional COnservatism with Anchor-seeking (COCOA) for offline RL, an approach that pursues conservatism in a compositional manner on top of the transductive reparameterization (Netanyahu et al., 2023), which decomposes the input variable (the state in our case) into an anchor and its difference from the original input. Our COCOA seeks both in-distribution anchors and differences by utilizing the learned reverse dynamics model, encouraging conservatism in the compositional input space for the policy or value function. Such compositional conservatism is independent of and agnostic to the prevalent behavioral conservatism in offline RL. We apply COCOA to four state-of-the-art offline RL algorithms and evaluate them on the D4RL benchmark, where COCOA generally improves the performance of each algorithm. The code is available at https://github.com/runamu/compositional-conservatism.

Video summarization is a crucial research area that aims to efficiently browse and retrieve relevant information from the vast amount of video content available today. With the exponential growth of multimedia data, the ability to extract meaningful representations from videos has become essential. Video summarization techniques automatically generate concise summaries by selecting keyframes, shots, or segments that capture the video's essence. This process improves the efficiency and accuracy of various applications, including video surveillance, education, entertainment, and social media. Despite the importance of video summarization, there is a lack of diverse and representative datasets, hindering comprehensive evaluation and benchmarking of algorithms. Existing evaluation metrics also fail to fully capture the complexities of video summarization, limiting accurate algorithm assessment and hindering the field's progress. To overcome data scarcity challenges and improve evaluation, we propose an unsupervised approach that leverages video data structure and information for generating informative summaries. By moving away from fixed annotations, our framework can produce representative summaries effectively. Moreover, we introduce an innovative evaluation pipeline tailored specifically for video summarization. Human participants are involved in the evaluation, comparing our generated summaries to ground truth summaries and assessing their informativeness. This human-centric approach provides valuable insights into the effectiveness of our proposed techniques. Experimental results demonstrate that our training-free framework outperforms existing unsupervised approaches and achieves competitive results compared to state-of-the-art supervised methods.

The development of robotic systems for palletization in logistics scenarios is of paramount importance, addressing critical efficiency and precision demands in supply chain management. This paper investigates the application of Reinforcement Learning (RL) in enhancing task planning for such robotic systems. Confronted with the substantial challenge of a vast action space, which is a significant impediment to efficiently apply out-of-the-shelf RL methods, our study introduces a novel method of utilizing supervised learning to iteratively prune and manage the action space effectively. By reducing the complexity of the action space, our approach not only accelerates the learning phase but also ensures the effectiveness and reliability of the task planning in robotic palletization. The experimental results underscore the efficacy of this method, highlighting its potential in improving the performance of RL applications in complex and high-dimensional environments like logistics palletization.

Unmanned Aerial Vehicles (UAVs) are integral in various sectors like agriculture, surveillance, and logistics, driven by advancements in 5G. However, existing research lacks a comprehensive approach addressing both data freshness and security concerns. In this paper, we address the intricate challenges of data freshness, and security, especially in the context of eavesdropping and jamming in modern UAV networks. Our framework incorporates exponential AoI metrics and emphasizes secrecy rate to tackle eavesdropping and jamming threats. We introduce a transformer-enhanced Deep Reinforcement Learning (DRL) approach to optimize task offloading processes. Comparative analysis with existing algorithms showcases the superiority of our scheme, indicating its promising advancements in UAV network management.

Direct Preference Optimization (DPO), which derives reward signals directly from pairwise preference data, has shown its effectiveness on aligning Large Language Models (LLMs) with human preferences. Despite its widespread use across various tasks, DPO has been criticized for its sensitivity to the SFT's effectiveness and its hindrance to the learning capacity towards human-preferred responses, leading to less satisfactory performance. To overcome those limitations, the theoretical understanding of DPO are indispensable but still lacking. To this end, we take a step towards theoretically analyzing and understanding the limitations of DPO. Specifically, we provide an analytical framework using the field theory to analyze the optimization process of DPO. By analyzing the gradient vector field of the DPO loss function, we find that the DPO loss function decreases the probability of producing human dispreferred data at a faster rate than it increases the probability of producing preferred data. This provides theoretical insights for understanding the limitations of DPO discovered in the related research experiments, thereby setting the foundation for its improvement.