Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks11 from public data
- Least-Loaded Expert Parallelism: Load Balancing An Imbalanced Mixture-of-Experts9
Least-Loaded Expert Parallelism (LLEP), a novel EP algorithm that dynamically reroutes excess tokens and associated expert parameters from overloaded devices to underutilized ones, which enables faster and higher-throughput post-training and inference, with ~1.9x faster for gpt-oss-120b.
- The Illusion of Multi-Agent Advantage8
It is shown that expert-architected MAS consistently outperforms automatically generated architectures in both raw performance and cost-efficiency on this dataset, demonstrating that existing evaluation frameworks mask critical architectural gaps and inefficiencies of complex MAS by failing to account for the marginal utility of increased computational cost.
- MAS-ProVe: Understanding the Process Verification of Multi-Agent Systems7
It is found that process-level verification does not consistently improve performance and frequently exhibits high variance, highlighting the difficulty of reliably evaluating partial multi-agent trajectories.
- 6
- 4
- SMILE: A Composite Lexical-Semantic Metric for Question-Answering Evaluation2
This work introduces SMILE: Semantic Metric Integrating Lexical Exactness, a novel approach that combines sentence-level semantic understanding with keyword-level semantic understanding and easy keyword matching, bridging the gap between lexical and semantic evaluation.
- GraphBit: A Graph-based Agentic Framework for Non-Linear Agent Orchestration1
GraphBit is introduced, an engine-orchestrated framework that defines workflows explicitly and deterministically as a directed acyclic graph (DAG), while a Rust-based engine governs routing, state transitions, and tool invocation, ensuring reproducibility and auditability.
- Reward Modeling for Multi-Agent Orchestration1
OrchRM is proposed, a self-supervised framework for evaluating orchestration quality without human annotations that operates directly at the orchestration level, enabling efficient and high-performing reward-guided orchestrator training and MAS test-time scaling.
- –
- –
- –
Publication data from OpenAlex; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-12. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.