For peg-in-hole tasks, humans rely on binocular visual perception to locate the peg above the hole surface and then proceed with insertion. This paper draws insights from this behavior to enable agents to learn efficient assembly strategies through visual reinforcement learning. Hence, we propose a Separate Primitive Policy (S2P) to simultaneously learn how to derive location and insertion actions. S2P is compatible with model-free reinforcement learning algorithms. Ten insertion tasks featuring different polygons are developed as benchmarks for evaluations. Simulation experiments show that S2P can boost the sample efficiency and success rate even with force constraints. Real-world experiments are also performed to verify the feasibility of S2P. Ablations are finally given to discuss the generalizability of S2P and some factors that affect its performance.

Reinforcement learning (RL) has made significant progress in various domains, but scaling it to long-horizon tasks with complex decision-making remains challenging. Skill learning attempts to address this by abstracting actions into higher-level behaviors. However, current approaches often fail to recognize semantically similar behaviors as the same skill and use fixed skill lengths, limiting flexibility and generalization. To address this, we propose Dynamic Contrastive Skill Learning (DCSL), a novel framework that redefines skill representation and learning. DCSL introduces three key ideas: state-transition based skill representation, skill similarity function learning, and dynamic skill length adjustment. By focusing on state transitions and leveraging contrastive learning, DCSL effectively captures the semantic context of behaviors and adapts skill lengths to match the appropriate temporal extent of behaviors. Our approach enables more flexible and adaptive skill extraction, particularly in complex or noisy datasets, and demonstrates competitive performance compared to existing methods in task completion and efficiency.

Machine learning (ML)-based planners have recently gained significant attention. They offer advantages over traditional optimization-based planning algorithms. These advantages include fewer manually selected parameters and faster development. Within ML-based planning, imitation learning (IL) is a common algorithm. It primarily learns driving policies directly from supervised trajectory data. While IL has demonstrated strong performance on many open-loop benchmarks, it remains challenging to determine if the learned policy truly understands fundamental driving principles, rather than simply extrapolating from the ego-vehicle's initial state. Several studies have identified this limitation and proposed algorithms to address it. However, these methods often use original datasets for evaluation. In these datasets, future trajectories are heavily dependent on initial conditions. Furthermore, IL often overfits to the most common scenarios. It struggles to generalize to rare or unseen situations. To address these challenges, this work proposes: 1) a novel closed-loop simulator supporting both imitation and reinforcement learning, 2) a causal benchmark derived from the Waymo Open Dataset to rigorously assess the impact of the copycat problem, and 3) a novel framework integrating imitation learning and reinforcement learning to overcome the limitations of purely imitative approaches. The code for this work will be released soon.

--This survey explores the development of meta-thinking capabilities in Large Language Models (LLMs) from a Multi-Agent Reinforcement Learning (MARL) perspective. The survey begins by analyzing current LLM limitations, such as hallucinations and the lack of internal self-assessment mechanisms. It then talks about newer methods, including RL from human feedback (RLHF), self-distillation, and chain-of-thought prompting, and each of their limitations. The crux of the survey is to talk about how multi-agent architectures, namely supervisor-agent hierarchies, agent debates, and theory of mind frameworks, can emulate human-like introspective behavior and enhance LLM robustness. By exploring reward mechanisms, self-play, and continuous learning methods in MARL, this survey gives a comprehensive roadmap to building introspective, adaptive, and trustworthy LLMs. Evaluation metrics, datasets, and future research avenues, including neuroscience-inspired architectures and hybrid symbolic reasoning, are also discussed. THE cognitive abilities, such as intelligence and creativity, have played a fundamental role in human discoveries and inventions. Understanding the relationship between these two cognitive abilities is important not only for the advancement of psychological theories but also for the improvement of educational practices [1]. However, researchers still hold different views on how intelligence and creativity interact, often leading to conflicting findings. A key question in this discourse is how intelligence enables structured problem-solving, while creativity fosters novel solutions that are essential for human cognition and artificial intelligence systems. Ahsan Bilal is with University of Oklahoma, Norman, OK, 73072, USA (e-mail: ahsan.bilal-1@ou.edu). Muhammad Ahmed Mohsin, Muhammad Umer are with Stanford University, Stanford, CA, 94305, USA (e-mail: muahmed, mumer@stanford.edu). Muhammad A wais Khan Bangash is with the School of Electrical and Computer Engineering, Oklahoma State University, Stillwater, OK, 74075 USA (e-mail: awais.bangash@okstate.edu). Muhammad Ali Jamshed is with University of Glasgow, G12 8QQ, Glasgow, UK (e-mail: muhammadali.jamshed@glasgow.ac.uk). Similarly, in problem-solving tasks, intelligence aids in analyzing constraints, while creativity allows for flexible and unconventional approaches. Moreover, the role of internal thought processes varies with task complexity. Simpler tasks require minimal reasoning, whereas more complex tasks demand deeper cognitive engagement. This principle extends to artificial intelligence, where more sophisticated models exhibit enhanced performance in tasks requiring higher-order thinking.

Personalized therapy, in which a therapeutic practice is adapted to an individual patient, can lead to improved health outcomes. Typically, this is accomplished by relying on a therapist's training and intuition along with feedback from a patient. However, this requires the therapist to become an expert on any technological components, such as in the case of Virtual Reality Exposure Therapy (VRET). While there exist approaches to automatically adapt therapeutic content to a patient, they generally rely on hand-authored, pre-defined rules, which may not generalize to all individuals. In this paper, we propose an approach to automatically adapt therapeutic content to patients based on physiological measures. We implement our approach in the context of virtual reality arachnophobia exposure therapy, and rely on experience-driven procedural content generation via reinforcement learning (ED-PCGRL) to generate virtual spiders to match an individual patient. Through a human subject study, we demonstrate that our system significantly outperforms a more common rules-based method, highlighting its potential for enhancing personalized therapeutic interventions.

Reinforcement Learning from Human Feedback (RLHF) is central in aligning large language models (LLMs) with human values and expectations. However, the process remains susceptible to governance challenges, including evaluator bias, inconsistency, and the unreliability of feedback. This study examines how the cognitive capacity of evaluators, specifically their level of rationality, affects the stability of reinforcement signals. A controlled experiment comparing high-rationality and low-rationality participants reveals that evaluators with higher rationality scores produce significantly more consistent and expert-aligned feedback. In contrast, lower-rationality participants demonstrate considerable variability in their reinforcement decisions ($p < 0.01$). To address these challenges and improve RLHF governance, we recommend implementing evaluator pre-screening, systematic auditing of feedback consistency, and reliability-weighted reinforcement aggregation. These measures enhance the fairness, transparency, and robustness of AI alignment pipelines.

This paper explores optimal data selection strategies for Reinforcement Learning with V erified Rewards (RLVR) training in the medical domain. While RLVR has shown exceptional potential for enhancing reasoning capabilities in large language models, most prior implementations have focused on mathematics and logical puzzles, with limited exploration of domain-specific applications like medicine. W e investigate four distinct data sampling strategies from MedQA-USMLE: random sampling (baseline), and filtering using Phi-4, Gemma-3-27b-it, and Gemma-3-12b-it models. Using Gemma-3-12b-it as our base model and implementing Group Relative Policy Optimization (GRPO), we evaluate performance across multiple benchmarks including MMLU, GSM8K, MMLU-Pro, and CMMLU. Our findings demonstrate that models trained on filtered data generally outperform those trained on randomly selected samples. Notably, training on self-filtered samples (using Gemma-3-12b-it for filtering) achieved superior performance in medical domains but showed reduced robustness across different benchmarks, while filtering with larger models from the same series yielded better overall robustness. These results provide valuable insights into effective data organization strategies for RLVR in specialized domains and highlight the importance of thoughtful data selection in achieving optimal performance. Y ou can access our repository to get the codes.

Email: rohan-singh@aist.go.jp Abstract -- For the deployment of legged robots in real-world environments, it is essential to develop robust locomotion control methods for challenging terrains that may exhibit unexpected deformability and irregularity. In this paper, we explore the application of sim-to-real deep reinforcement learning (RL) for the design of bipedal locomotion controllers for humanoid robots on compliant and uneven terrains. Our key contribution is to show that a simple training curriculum for exposing the RL agent to randomized terrains in simulation can achieve robust walking on a real humanoid robot using only proprioceptive feedback. We train an end-to-end bipedal locomotion policy using the proposed approach, and show extensive real-robot demonstration on the HRP-5P humanoid over several difficult terrains inside and outside the lab environment. Further, we argue that the robustness of a bipedal walking policy can be improved if the robot is allowed to exhibit aperiodic motion with variable stepping frequency. We propose a new control policy to enable modification of the observed clock signal, leading to adaptive gait frequencies depending on the terrain and command velocity. Through simulation experiments, we show the effectiveness of this policy specifically for walking over challenging terrains by controlling swing and stance durations. This is primarily due to the strict temporal and spatial assumptions placed by such approaches on the foot trajectories and environmental contacts [1], [2]. When faced with an irregular or compliant (i.e.

The ability to train intelligent autonomous agents (such as mobile robots) on multiple tasks is crucial for adapting to dynamic real-world environments. However, state-of-the-art reinforcement learning (RL) methods only excel in single-task settings, and still struggle to generalize across multiple tasks due to task interference. Moreover, real-world environments also demand the agents to have data stream processing capabilities. Toward this, a state-of-the-art work employs Spiking Neural Networks (SNNs) to improve multi-task learning by exploiting temporal information in data stream, while enabling lowpower/energy event-based operations. However, it relies on fixed context/task-switching intervals during its training, hence limiting the scalability and effectiveness of multi-task learning. To address these limitations, we propose SwitchMT, a novel adaptive task-switching methodology for RL-based multi-task learning in autonomous agents. Specifically, SwitchMT employs the following key ideas: (1) a Deep Spiking Q-Network with active dendrites and dueling structure, that utilizes task-specific context signals to create specialized sub-networks; and (2) an adaptive task-switching policy that leverages both rewards and internal dynamics of the network parameters. Experimental results demonstrate that SwitchMT achieves superior performance in multi-task learning compared to state-of-the-art methods. It achieves competitive scores in multiple Atari games (i.e., Pong: -8.8, Breakout: 5.6, and Enduro: 355.2) compared to the state-of-the-art, showing its better generalized learning capability. These results highlight the effectiveness of our SwitchMT methodology in addressing task interference while enabling multi-task learning automation through adaptive task switching, thereby paving the way for more efficient generalist agents with scalable multi-task learning capabilities.

We introduce Iterative Dual Reinforcement Learning (IDRL), a new method that takes an optimal discriminator-weighted imitation view of solving RL. Our method is motivated by a simple experiment in which we find training a discriminator using the offline dataset plus an additional expert dataset and then performing discriminator-weighted behavior cloning gives strong results on various types of datasets. That optimal discriminator weight is quite similar to the learned visitation distribution ratio in Dual-RL, however, we find that current Dual-RL methods do not correctly estimate that ratio. In IDRL, we propose a correction method to iteratively approach the optimal visitation distribution ratio in the offline dataset given no addtional expert dataset. During each iteration, IDRL removes zero-weight suboptimal transitions using the learned ratio from the previous iteration and runs Dual-RL on the remaining subdataset. This can be seen as replacing the behavior visitation distribution with the optimized visitation distribution from the previous iteration, which theoretically gives a curriculum of improved visitation distribution ratios that are closer to the optimal discriminator weight. We verify the effectiveness of IDRL on various kinds of offline datasets, including D4RL datasets and more realistic corrupted demonstrations. IDRL beats strong Primal-RL and Dual-RL baselines in terms of both performance and stability, on all datasets.