After 20 dead versions and about 2 month of work, my RL agent (NASMU) passed its walk-forward backtest across 2020–2026. But the most interesting part wasn’t the results — it was what the model actually learned. The setup: – PPO + xLSTM (4 blocks), BTC/USDT 4h bars – 35 features distilled from López de Prado, Hilpisch, Kaabar, Chan and others – Triple Barrier labeling (TP/SL/Timeout) – HMM for regime detection (bull/bear/sideways) – Running on a Xeon E5-1650 v2 […]

Ver mais

Recently, there have been numerous attempts to enhance the sample efficiency of off-policy reinforcement learning (RL) agents when interacting with the environment, including architecture improvements and new algorithms. Despite these advances, they overlook the potential of directly constraining the initial representations of the input data, which can intuitively alleviate the distribution shift issue and stabilize training. In this paper, we introduce the Tanh function into the initial layer to fulfill such a constraint. We theoretically unpack the convergence […]

Ver mais

Language models are revolutionizing the biochemistry domain, assisting scientists in drug design and chemical synthesis with high efficiency. Yet current approaches struggle between small language models prone to hallucination and limited knowledge retention, and large cloud-based language models plagued by privacy risks and high inference costs. To bridge this gap, we introduce ChemCRAFT, a novel framework leveraging agentic reinforcement learning to decouple chemical reasoning from knowledge storage. Instead of forcing the model to memorize vast chemical data, our […]

Ver mais

arXiv:2604.06299v1 Announce Type: new Abstract: This paper studies the co-design of actuators, sensors, and communication in the distributed setting, where a networked plant is partitioned into subsystems each equipped with a sub-controller interacting with other sub-controllers. The objective is to jointly minimize control cost (measured by LQ cost) and material cost (measured by the number of actuators, sensors, and communication links used). We approach this using an evolutionary algorithm to selectively prune a baseline dense LQR controller. We […]

Ver mais

arXiv:2601.20533v1 Announce Type: new Abstract: Survival analysis has become a standard approach for modelling time to default by time-varying covariates in credit risk. Unlike most existing methods that implicitly assume a stationary data-generating process, in practise, mortgage portfolios are exposed to various forms of data drift caused by changing borrower behaviour, macroeconomic conditions, policy regimes and so on. This study investigates the impact of data drift on survival-based credit risk models and proposes a dynamic joint modelling framework […]

Ver mais

The integration of Large Language Models (LLMs) into clinical decision support is critically obstructed by their opaque and often unreliable reasoning. In the high-stakes domain of healthcare, correct answers alone are insufficient; clinical practice demands full transparency to ensure patient safety and enable professional accountability. A pervasive and dangerous weakness of current LLMs is their tendency to produce "correct answers through flawed reasoning." This issue is far more than a minor academic flaw; such process errors signal a […]

Ver mais

Reinforcement Learning (RL) has emerged as a crucial method for training or fine-tuning large language models (LLMs), enabling adaptive, task-specific optimizations through interactive feedback. Multi-Agent Reinforcement Learning (MARL), in particular, offers a promising avenue by decomposing complex tasks into specialized subtasks learned by distinct interacting agents, potentially enhancing the ability and efficiency of LLM systems. However, theoretical insights regarding when and why MARL outperforms Single-Agent RL (SARL) remain limited, creating uncertainty in selecting the appropriate RL framework. In […]

Ver mais

arXiv:2604.14158v1 Announce Type: new Abstract: Current evaluations of long-term memory in LLMs are fundamentally static. By fixating on simple retrieval and short-context inference, they neglect the multifaceted nature of complex memory systems, such as dynamic state tracking and hierarchical reasoning in continuous interactions. To overcome these limitations, we propose MemGround, a rigorous long-term memory benchmark natively grounded in rich, gamified interactive scenarios. To systematically assess these capabilities, MemGround introduces a three-tier hierarchical framework that evaluates Surface State Memory, […]

Ver mais