Framework for compositional skill routing in AI agents using execution graphs and SAD.

Implementation of Fixed-point Flow Map Language Models (FMLM*) for high-performance text generation.

A low-level orchestration framework for stateful multi-agent systems.

Benchmark for measuring egocentric action selection in multi-agent scenes for MLLMs.

A curated list of security resources for mobile on-device AI systems.

Agentic Skill Programming through Iterative Robot Exploration for autonomous robot control program discovery.

Adaptive collaborative-perception framework for distributed autonomous systems.

A Julia library for exposing block-sparse structure in ML frameworks for efficient optimization.

A Julia library for solving large-scale sparse nonlinear programs on GPUs.

A self-evolving framework for optimizing agentic information retrieval via mutable catalogs.

BFS-and-Reflection Web Data Collection Agent for site-level multimodal extraction.

A sandbox environment and framework for evaluating and improving LLM agent robustness in open-world tool use.

Open-source Python simulator for scalable pedestrian network flow simulation and control.

Implementation of RLPO and RLMO for stable low-rank adaptation in RLVR.

A comprehensive open-source benchmark for standardizing the evaluation of style text embeddings.

Compiles Infrastructure as Code to a single binary for simplified deployment.

Open-source AI flaw reporting system designed for interoperability.

A unified multimodal model for cross-sensor tactile understanding and generation.

A data-efficient, one-step text-to-audio generation framework using caption-only distillation.

CLI that writes and maintains agent documentation for your codebase