Multi-Agent Reinforcement Learning (MARL) holds significant promise for enabling cooperative driving among Connected and Automated Vehicles (CAVs). However, its practical application is hindered by a critical limitation, i.e., insufficient fault tolerance against observational faults. Such faults, which appear as perturbations in the vehicles' perceived data, can substantially compromise the performance of MARL-based driving systems. Addressing this problem presents two primary challenges. One is to generate adversarial perturbations that effectively stress the policy during training, and the other is to equip vehicles with the capability to mitigate the impact of corrupted observations. To overcome the challenges, we propose a fault-tolerant MARL method for cooperative on-ramp vehicles incorporating two key agents. First, an adversarial fault injection agent is co-trained to generate perturbations that actively challenge and harden the vehicle policies. Second, we design a novel fault-tolerant vehicle agent equipped with a self-diagnosis capability, which leverages the inherent spatio-temporal correlations in vehicle state sequences to detect faults and reconstruct credible observations, thereby shielding the policy from misleading inputs. Experiments in a simulated highway merging scenario demonstrate that our method significantly outperforms baseline MARL approaches, achieving near-fault-free levels of safety and efficiency under various observation fault patterns.

Underwater visual localization remains challenging due to wavelength-dependent attenuation, poor texture, and non-Gaussian sensor noise. We introduce MARVO, a physics-aware, learning-integrated odometry framework that fuses underwater image formation modeling, differentiable matching, and reinforcement-learning optimization. At the front-end, we extend transformer-based feature matcher with a Physics Aware Radiance Adapter that compensates for color channel attenuation and contrast loss, yielding geometrically consistent feature correspondences under turbidity. These semi dense matches are combined with inertial and pressure measurements inside a factor-graph backend, where we formulate a keyframe-based visual-inertial-barometric estimator using GTSAM library. Each keyframe introduces (i) Pre-integrated IMU motion factors, (ii) MARVO-derived visual pose factors, and (iii) barometric depth priors, giving a full-state MAP estimate in real time. Lastly, we introduce a Reinforcement-Learningbased Pose-Graph Optimizer that refines global trajectories beyond local minima of classical least-squares solvers by learning optimal retraction actions on SE(2).

We find that in Hanabi, one of the most complex and popular benchmarks for zero-shot coordination and ad-hoc teamplay, a standard implementation of independent PPO with a slightly higher entropy coefficient 0.05 instead of the typically used 0.01, achieves a new state-of-the-art in cross-play between different seeds, beating by a significant margin all previous specialized algorithms, which were specifically designed for this setting. We provide an intuition for why sufficiently high entropy regularization ensures that different random seed produce joint policies which are mutually compatible. We also empirically find that a high $λ_{\text{GAE}}$ around 0.9, and using RNNs instead of just feed-forward layers in the actor-critic architecture, strongly increase inter-seed cross-play. While these results demonstrate the dramatic effect that hyperparameters can have not just on self-play scores but also on cross-play scores, we show that there are simple Dec-POMDPs though, in which standard policy gradient methods with increased entropy regularization are not able to achieve perfect inter-seed cross-play, thus demonstrating the continuing necessity for new algorithms for zero-shot coordination.

To reliably navigate ever-shifting real-world environments, agents must grapple with incomplete knowledge and adapt their behavior through experience. However, current evaluations largely focus on tasks that leave no ambiguity, and do not measure agents' ability to adaptively learn and reason through the experiences they accrued. We exemplify the need for this in-context experiential learning in a product recommendation context, where agents must navigate shifting customer preferences and product landscapes through natural language dialogue. We curate a benchmark for experiential learning and active exploration (BELA) that combines (1) rich real-world products from Amazon, (2) a diverse collection of user personas to represent heterogeneous yet latent preferences, and (3) a LLM user simulator powered by the persona to create rich interactive trajectories. We observe that current frontier models struggle to meaningfully improve across episodes, underscoring the need for agentic systems with strong in-context learning capabilities.

As embodied agents operate in increasingly complex environments, the ability to perceive, track, and reason about individual object instances over time becomes essential, especially in tasks requiring sequenced interactions with visually similar objects. In these non-Markovian settings, key decision cues are often hidden in object-specific histories rather than the current scene. Without persistent memory of prior interactions (what has been interacted with, where it has been, or how it has changed) visuomotor policies may fail, repeat past actions, or overlook completed ones. To surface this challenge, we introduce LIBERO-Mem, a non-Markovian task suite for stress-testing robotic manipulation under object-level partial observability. It combines short- and long-horizon object tracking with temporally sequenced subgoals, requiring reasoning beyond the current frame. However, vision-language-action (VLA) models often struggle in such settings, with token scaling quickly becoming intractable even for tasks spanning just a few hundred frames. We propose Embodied-SlotSSM, a slot-centric VLA framework built for temporal scalability. It maintains spatio-temporally consistent slot identities and leverages them through two mechanisms: (1) slot-state-space modeling for reconstructing short-term history, and (2) a relational encoder to align the input tokens with action decoding. Together, these components enable temporally grounded, context-aware action prediction. Experiments show Embodied-SlotSSM's baseline performance on LIBERO-Mem and general tasks, offering a scalable solution for non-Markovian reasoning in object-centric robotic policies.

-- Planning under partial observability is an essential capability of autonomous robots. The Partially Observable Markov Decision Process (POMDP) provides a powerful framework for planning under partial observability problems, capturing the stochastic effects of actions and the limited information available through noisy observations. POMDP solving could benefit tremendously from massive parallelization on today's hardware, but parallelizing POMDP solvers has been challenging. Most of these solvers rely on interleaving numerical optimization over actions with the estimation of their values, which creates dependencies and synchronization bottlenecks between parallel processes that can offset the benefits of paral-lelization. In this paper, we propose V ectorized Online POMDP Planner (VOPP), a novel parallel online solver that leverages a recent POMDP formulation which analytically solves part of the optimization component, leaving numerical computations to consist of only estimation of expectations. VOPP represents all data structures related to planning as a collection of tensors, and implements all planning steps as fully vectorized computations over this representation. The result is a massively parallel solver with no dependencies or synchronization bottlenecks between parallel processes. Experimental results indicate that VOPP is at least 20 more efficient in computing near-optimal solutions compared to an existing state-of-the-art parallel online solver .

To enable embodied agents to operate effectively over extended timeframes, it is crucial to develop models that form and access memories to stay contextualized in their environment. In the current paradigm of training transformer-based policies for embodied sequential decision-making tasks, visual inputs often overwhelm the context limits of transformers, while humans can maintain and utilize a lifetime of experience compressed as memories. Significant compression is possible in principle, as much of the input is irrelevant and can be abstracted. However, existing approaches predominantly focus on either recurrent models with fixed-size memory or transformers with full-context reliance. In this work, we propose Memo, a transformer-based architecture and training recipe for reinforcement learning (RL) on memory-intensive, long-horizon tasks. Memo incorporates the creation and retrieval of memory by interleaving periodic summarization tokens with the inputs of a model during training. We demonstrate Memo's effectiveness on a gridworld meta-RL benchmark and a multi-object navigation task in photo-realistic indoor settings. Memo outperforms naive long-context transformer baselines while being more compute and storage efficient. Additionally, Memo generalizes better to longer contexts at inference time and remains robust in streaming settings, where historical context must be truncated to fit inference constraints. Our code is available at: https://github.com/gunshi/memo.

We investigate how machine learning models acquire the ability to compose music and how musical information is internally represented within such models. We develop a composition algorithm based on a restricted Boltzmann machine (RBM), a simple generative model capable of producing musical pieces of arbitrary length. We convert musical scores into piano-roll image representations and train the RBM in an unsupervised manner. We confirm that the trained RBM can generate new musical pieces; however, by analyzing the model's responses and internal structure, we find that the learned information is not stored in a form directly interpretable by humans. This study contributes to a better understanding of how machine learning models capable of music composition may internally represent musical structure and highlights issues related to the interpretability of generative models in creative tasks.

Markov Chain Monte Carlo (MCMC) is a flexible approach to approximate sampling from intractable probability distributions, with a rich theoretical foundation and comprising a wealth of exemplar algorithms. While the qualitative correctness of MCMC algorithms is often easy to ensure, their practical efficiency is contingent on the `target' distribution being reasonably well-behaved. In this work, we concern ourself with the scenario in which this good behaviour is called into question, reviewing an emerging line of work on `robust' MCMC algorithms which can perform acceptably even in the face of certain pathologies. We focus on two particular pathologies which, while simple, can already have dramatic effects on standard `local' algorithms. The first is roughness, whereby the target distribution varies so rapidly that the numerical stability of the algorithm is tenuous. The second is flatness, whereby the landscape of the target distribution is instead so barren and uninformative that one becomes lost in uninteresting parts of the state space. In each case, we formulate the pathology in concrete terms, review a range of proposed algorithmic remedies to the pathology, and outline promising directions for future research.

Cooperative multi-agent reinforcement learning (MARL) commonly adopts centralized training with decentralized execution (CTDE), where centralized critics leverage global information to guide decentralized actors. However, centralized-decentralized mismatch (CDM) arises when the suboptimal behavior of one agent degrades others' learning. Prior approaches mitigate CDM through value decomposition, but linear decompositions allow per-agent gradients at the cost of limited expressiveness, while nonlinear decompositions improve representation but require centralized gradients, reintroducing CDM. To overcome this trade-off, we propose the multi-agent cross-entropy method (MCEM), combined with monotonic nonlinear critic decomposition (NCD). MCEM updates policies by increasing the probability of high-value joint actions, thereby excluding suboptimal behaviors. For sample efficiency, we extend off-policy learning with a modified k-step return and Retrace. Analysis and experiments demonstrate that MCEM outperforms state-of-the-art methods across both continuous and discrete action benchmarks.