Tag
This paper proposes an agentic LLM framework for automated structural analysis of 3D frame systems from natural language inputs, achieving 90% accuracy on ten representative 3D frames through a multi-agent pipeline.
This paper introduces SearchSwarm, a model trained on synthesized delegation intelligence to improve long-horizon deep research tasks via task decomposition and subagent coordination, achieving state-of-the-art results on BrowseComp benchmarks.
SAGE proposes a novelty gate for memory evolution in agentic LLMs, using a von Mises-Fisher-based density estimator to decide whether to add, merge, or ignore new facts, reducing LLM calls while maintaining memory quality.