An open-source orchestrator that exposes 150+ security tools for LLM agents.

Distributed system for coordination of LLM workers to generate formally verified programs at scale.

Language-conditioning space optimization for black-box VLA models to improve robot control.

A framework integrating LLMs with symbolic regression to guide equation discovery via search control.

A bias mitigation approach for T2I diffusion models based on concept-graph alignment.

Explainable multi-head logarithmic structural mask for graph-structured data in Transformers.

Code implementation of Value Diffusion World Models for MPC latent planning.

A framework for benchmarking algorithmic adaptation in LLM code generation through constraint-shifting.

Model-agnostic framework to migrate advanced encoders to sparse-friendly normalized vocabularies.

Robot-collected touch-vision-language dataset for tactile generalization.

A hierarchical framework for robot manipulation using 3D trajectory guidance in VLA models.

Tabular Data Generation Toolkit for synthetic tabular data and fidelity assessment.

Unified framework for Vision-Language-Action agents with temporal memory and parallel action decoding.

Pose-aware multi-view architecture for skilled action detection and grading in ego-exo videos.

A semantic operating-system layer for accessibility-grounded AI agents.

A module for elastically orchestrating historical step information in agent contexts.

An LLM-oriented simulation platform and benchmark for multi-UAV collaborative task planning.

Unified whole-body humanoid controller driven by arbitrary body keypoints.

Event-level Neurodegenerative modeling in Continuous time with neural Ordinary Differential Equations.

Unified framework for dynamic self-debate within a single model using MoE paradigm.