Then, we perform sequential reasoning over the graph, iteratively traversing its nodes to answer a given question or draw a new inference.

Memory-based meta-learning is a powerful technique to build agents that adapt fast to any task within a target distribution. A previous theoretical study has argued that this remarkable performance is because the meta-training protocol incentivises agents to behave Bayes-optimally.

Large Language Models (LLMs) based agents have demonstrated remarkable potential in autonomous task-solving across complex, open-ended environments. A promising approach for improving the reasoning capabilities of LLM agents is to better utilize prior experiences in guiding current decisions. However, LLMs acquire experience either through implicit memory via training, which suffers from catastrophic forgetting and limited interpretability, or explicit memory via prompting, which lacks adaptability. In this paper, we introduce a novel agent-centric, trainable, multi-layered graph memory framework and evaluate how context memory enhances the ability of LLMs to utilize parametric information. The graph abstracts raw agent trajectories into structured decision paths in a state machine and further distills them into high-level, human-interpretable strategic meta-cognition. In order to make memory adaptable, we propose a reinforcement-based weight optimization procedure that estimates the empirical utility of each meta-cognition based on reward feedback from downstream tasks. These optimized strategies are then dynamically integrated into the LLM agent's training loop through meta-cognitive prompting. Empirically, the learnable graph memory delivers robust generalization, improves LLM agents' strategic reasoning performance, and provides consistent benefits during Reinforcement Learning (RL) training.

The role of hidden units in recurrent neural networks is typically seen as modeling memory, with research focusing on enhancing information retention through gating mechanisms. A less explored perspective views hidden units as active participants in the computation performed by the network, rather than passive memory stores. In this work, we revisit bilinear operations, which involve multiplicative interactions between hidden units and input embeddings. We demonstrate theoretically and empirically that they constitute a natural inductive bias for representing the evolution of hidden states in state tracking tasks. These are the simplest type of tasks that require hidden units to actively contribute to the behavior of the network. We also show that bilinear state updates form a natural hierarchy corresponding to state tracking tasks of increasing complexity, with popular linear recurrent networks such as Mamba residing at the lowest-complexity center of that hierarchy.

In the field of railway automation, one of the key challenges has been the development of effective computer vision systems due to the limited availability of high-quality, sequential data. Traditional datasets are restricted in scope, lacking the spatio temporal context necessary for real-time decision-making, while alternative solutions introduce issues related to realism and applicability. By focusing on route-specific, contextually relevant cues, we can generate rich, sequential datasets that align more closely with real-world operational logic. The concept of milestone determination allows for the development of targeted, rule-based models that simplify the learning process by eliminating the need for generalized recognition of dynamic components, focusing instead on the critical decision points along a route. We argue that this approach provides a practical framework for training vision agents in controlled, predictable environments, facilitating safer and more efficient machine learning systems for railway automation.

This work introduces a new framework for modeling financial markets through an interpretable probabilistic state machine. By clustering historical returns based on momentum and risk features across multiple time horizons, we identify distinct market states that capture underlying regimes, such as expansion phase, contraction, crisis, or recovery. From a transition matrix representing the dynamics between these states, we construct a probabilistic state machine that models the temporal evolution of the market. This state machine enables the generation of a custom distribution of returns based on a mixture of Gaussian components weighted by state frequencies. We show that the proposed benchmark significantly outperforms the traditional approach in capturing key statistical properties of asset returns, including skewness and kurtosis, and our experiments across random assets and time periods confirm its robustness.

Designing robotic systems to act autonomously in unforeseen environments is a challenging task. This work presents a novel approach to use formal verification, specifically Statistical Model Checking (SMC), to verify system properties of autonomous robots at design-time. We introduce an extension of the SCXML format, designed to model system components including both Robot Operating System 2 (ROS 2) and Behavior Tree (BT) features. Further, we contribute Autonomous Systems to Formal Models (AS2FM), a tool to translate the full system model into JANI. The use of JANI, a standard format for quantitative model checking, enables verification of system properties with off-the-shelf SMC tools. We demonstrate the practical usability of AS2FM both in terms of applicability to real-world autonomous robotic control systems, and in terms of verification runtime scaling. We provide a case study, where we successfully identify problems in a ROS 2-based robotic manipulation use case that is verifiable in less than one second using consumer hardware. Additionally, we compare to the state of the art and demonstrate that our method is more comprehensive in system feature support, and that the verification runtime scales linearly with the size of the model, instead of exponentially.

Then, we perform sequential reasoning over the graph, iteratively traversing its nodes to answer a given question or draw a new inference.

Structured decoding enables large language models (LLMs) to generate outputs in formats required by downstream systems, such as HTML or JSON. However, existing methods suffer from efficiency bottlenecks due to grammar compilation, state tracking, and mask creation. We observe that many real-world tasks embed strong prior knowledge about output structure. Leveraging this, we propose a decomposition of constraints into static and dynamic components -- precompiling static structures offline and instantiating dynamic arguments at runtime using grammar snippets. Instead of relying on pushdown automata, we employ a compositional set of operators to model regular formats, achieving lower transition latency. We introduce wgrammar, a lightweight decoding engine that integrates domain-aware simplification, constraint decomposition, and mask caching, achieving up to 250x speedup over existing systems. wgrammar's source code is publicly available at https://github.com/wrran/wgrammar.