Sean J. Wang

Zijian Carl Ma, Sean J. Wang, Sijbren Kramer et al.

Historical medical archives and traditional medicines hold immense potential for drug discovery and remain a primary source for current drug development. However, pre-ontological prose and idiosyncratic taxonomies prevent the standardization and medical modernization of the data for use in current biomedical pipelines. Furthermore, no existing LLM agent system, whether tool-calling, retrieval-augmented, or agentic deep-research, can convert such text into verifiable drug-discovery leads at scale. We close this gap with DeepRoot, a multi-agent LLM system that jointly builds and utilizes a verified knowledge graph, showing that grounding and reasoning -- often conflated -- are separable axes the system can compose for therapeutic reasoning. Applied to the Shen Nong Ben Cao Jing, DeepRoot recovers $10$ of $21$ held-out compound-disease treatment pairs at R@$20$ ($47.6\%$ vs $4.8\%$ for a raw corpus LLM and $\sim\!2.4\%$ random) and dominates an LLM-as-judge audit for reasoning quality over baseline LLMs and LLMs with direct tool-call access to the same APIs DeepRoot itself queries. Tool-using LLMs hallucinate evidence on $87\%$ of claims, versus 7-10% for DeepRoot. Graph-only inference hallucinates $0\%$ but ranks lowest on reasoning coherence; DeepRoot KG+LLM is the only condition to win on both axes, pointing toward a route for systematic mining and repurposing of historical medical knowledge.

YuanLab. ai, Shawn Wu, Jiangang Luo et al.

Although Mixture-of-Experts (MoE) Large Language Models (LLMs) deliver superior accuracy with a reduced number of active parameters, their pre-training represents a significant computationally bottleneck due to underutilized experts and limited training efficiency. This work introduces a Layer-Adaptive Expert Pruning (LAEP) algorithm designed for the pre-training stage of MoE LLMs. In contrast to previous expert pruning approaches that operate primarily in the post-training phase, the proposed algorithm enhances training efficiency by selectively pruning underutilized experts and reorganizing experts across computing devices according to token distribution statistics. Comprehensive experiments demonstrate that LAEP effectively reduces model size and substantially improves pre-training efficiency. In particular, when pre-training the 1010B Base model from scratch, LAEP achieves a 48.3\% improvement in training efficiency alongside a 33.3% parameter reduction, while still delivering excellent performance across multiple domains.

Jeffrey R. Peters, Sean J. Wang, Amit Surana et al.

A decomposition-based coverage control scheme is proposed for multi-agent, persistent surveillance missions operating in a communication-constrained, dynamic environment. The proposed approach decouples high-level task assignment from low-level motion planning in a modular framework. Coverage assignments and surveillance parameters are managed by a central base station, and transmitted to mobile agents via unplanned and asynchronous exchanges. Coverage updates promote load balancing, while maintaining geometric and temporal characteristics that allow effective pairing with generic path planners. Namely, the proposed scheme guarantees that (i) coverage regions are connected and collectively cover the environment, (ii) subregions may only go uncovered for bounded periods of time, (iii) collisions (or sensing overlaps) are inherently avoided, and (iv) under static event likelihoods, the collective coverage regions converge to a Pareto-optimal configuration. This management scheme is then paired with a generic path planner satisfying loose assumptions. The scheme is illustrated through simulated surveillance missions.