Multi-objective reinforcement learning (MORL) excels at handling rapidly changing preferences in tasks that involve multiple criteria, even for unseen preferences. However, previous dominating MORL methods typically generate a fixed policy set or preference-conditioned policy through multiple training iterations exclusively for sampled preference vectors, and cannot ensure the efficient discovery of the Pareto front. Furthermore, integrating preferences into the input of policy or value functions presents scalability challenges, in particular as the dimension of the state and preference space grow, which can complicate the learning process and hinder the algorithm's performance on more complex tasks. To address these issues, we propose a two-stage Pareto front discovery algorithm called Constrained MORL (C-MORL), which serves as a seamless bridge between constrained policy optimization and MORL. Concretely, a set of policies is trained in parallel in the initialization stage, with each optimized towards its individual preference over the multiple objectives. Then, to fill the remaining vacancies in the Pareto front, the constrained optimization steps are employed to maximize one objective while constraining the other objectives to exceed a predefined threshold. Empirically, compared to recent advancements in MORL methods, our algorithm achieves more consistent and superior performances in terms of hypervolume, expected utility, and sparsity on both discrete and continuous control tasks, especially with numerous objectives (up to nine objectives in our experiments). In many real-world control and planning problems, multiple and sometimes even conflicting objectives are getting involved. For instance, in industrial control scenarios (Salvendy, 2001; Wang et al., 2023), maximizing utility and minimizing energy consumption are of particular interest as objectives to be optimized. Since different decision makers have heterogeneous preferences over these objectives, there may exist multiple Pareto-optimal policies (Roijers et al., 2014). Classical reinforcement learning (RL) methods typically involve training individual policies exclusively to align with each preference weight vector over multiple rewards (Nagabandi et al., 2018; Gupta et al., 2018). Yet it may lead to an enormous computational burden due to the overly dependence on the model retraining and fine-tuning stages. Moreover, such policies are hard to directly generalize or transfer to newer tasks (Cobbe et al., 2019; Taiga et al., 2022).

The advent of large language models (LLMs) has revolutionized the field of natural language processing, yet they might be attacked to produce harmful content. Despite efforts to ethically align LLMs, these are often fragile and can be circumvented by jailbreaking attacks through optimized or manual adversarial prompts. To address this, we introduce the Information Bottleneck Protector (IBProtector), a defense mechanism grounded in the information bottleneck principle, and we modify the objective to avoid trivial solutions. The IBProtector selectively compresses and perturbs prompts, facilitated by a lightweight and trainable extractor, preserving only essential information for the target LLMs to respond with the expected answer. Moreover, we further consider a situation where the gradient is not visible to be compatible with any LLM. Our empirical evaluations show that IBProtector outperforms current defense methods in mitigating jailbreak attempts, without overly affecting response quality or inference speed. Its effectiveness and adaptability across various attack methods and target LLMs underscore the potential of IBProtector as a novel, transferable defense that bolsters the security of LLMs without requiring modifications to the underlying models. WARNING: This paper contains AI-generated text that is offensive in nature.

One of the notorious issues for Reinforcement Learning (RL) is poor sample efficiency. Compared to single agent RL, the sample efficiency for Multi-Agent Reinforcement Learning (MARL) is more challenging because of its inherent partial observability, non-stationary training, and enormous strategy space. Although much effort has been devoted to developing new methods and enhancing sample efficiency, we look at the widely used episodic training mechanism. In each training step, tens of frames are collected, but only one gradient step is made. We argue that this episodic training could be a source of poor sample efficiency. To better exploit the data already collected, we propose to increase the frequency of the gradient updates per environment interaction (a.k.a. Replay Ratio or Update-To-Data ratio). To show its generality, we evaluate $3$ MARL methods on $6$ SMAC tasks. The empirical results validate that a higher replay ratio significantly improves the sample efficiency for MARL algorithms. The codes to reimplement the results presented in this paper are open-sourced at https://anonymous.4open.science/r/rr_for_MARL-0D83/.

When facing an unsatisfactory prediction from a machine learning model, it is crucial to investigate the underlying reasons and explore the potential for reversing the outcome. We ask: To flip the prediction on a test point $x_t$, how to identify the smallest training subset $\mathcal{S}_t$ we need to relabel? We propose an efficient procedure to identify and relabel such a subset via an extended influence function. We find that relabeling fewer than 2% of the training points can always flip a prediction. This mechanism can serve multiple purposes: (1) providing an approach to challenge a model prediction by altering training points; (2) evaluating model robustness with the cardinality of the subset (i.e., $|\mathcal{S}_t|$); we show that $|\mathcal{S}_t|$ is highly related to the noise ratio in the training set and $|\mathcal{S}_t|$ is correlated with but complementary to predicted probabilities; (3) revealing training points lead to group attribution bias. To the best of our knowledge, we are the first to investigate identifying and relabeling the minimal training subset required to flip a given prediction.

Offline reinforcement learning (RL) methodologies enforce constraints on the policy to adhere closely to the behavior policy, thereby stabilizing value learning and mitigating the selection of out-of-distribution (OOD) actions during test time. Conventional approaches apply identical constraints for both value learning and test time inference. However, our findings indicate that the constraints suitable for value estimation may in fact be excessively restrictive for action selection during test time. To address this issue, we propose a Mildly Constrained Evaluation Policy (MCEP) for test time inference with a more constrained target policy for value estimation. Since the target policy has been adopted in various prior approaches, MCEP can be seamlessly integrated with them as a plug-in. We instantiate MCEP based on TD3-BC [Fujimoto and Gu, 2021] and AWAC [Nair et al., 2020] algorithms. The empirical results on MuJoCo locomotion tasks show that the MCEP significantly outperforms the target policy and achieves competitive results to state-of-the-art offline RL methods. The codes are open-sourced at https://github.com/egg-west/MCEP.git.

Portfolio methods represent a simple but efficient type of action abstraction which has shown to improve the performance of search-based agents in a range of strategy games. We first review existing portfolio techniques and propose a new algorithm for optimization and action-selection based on the Rolling Horizon Evolutionary Algorithm. Moreover, a series of variants are developed to solve problems in different aspects. We further analyze the performance of discussed agents in a general strategy game-playing task. For this purpose, we run experiments on three different game-modes of the Stratega framework. For the optimization of the agents' parameters and portfolio sets we study the use of the N-tuple Bandit Evolutionary Algorithm. The resulting portfolio sets suggest a high diversity in play-styles while being able to consistently beat the sample agents. An analysis of the agents' performance shows that the proposed algorithm generalizes well to all game-modes and is able to outperform other portfolio methods.

Designing agents that are able to achieve different play-styles while maintaining a competitive level of play is a difficult task, especially for games for which the research community has not found super-human performance yet, like strategy games. These require the AI to deal with large action spaces, long-term planning and partial observability, among other well-known factors that make decision-making a hard problem. On top of this, achieving distinct play-styles using a general algorithm without reducing playing strength is not trivial. In this paper, we propose Portfolio Monte Carlo Tree Search with Progressive Unpruning for playing a turn-based strategy game (Tribes) and show how it can be parameterized so a quality-diversity algorithm (MAP-Elites) is used to achieve different play-styles while keeping a competitive level of play. Our results show that this algorithm is capable of achieving these goals even for an extensive collection of game levels beyond those used for training.