An open-source embedding database for building AI applications with RAG.

Packet-level Adversarial Attacks framework for evading network traffic detection systems.

Code and data for automated geospatial vector data quality assessment.

A 14B parameter model designed for predicting BDI/E cognitive states and user utterances.

Code for measuring graph-to-graph semantic similarity in Knowledge Graphs.

A public benchmark of real-world point-of-care clinical queries for evaluating AI tools.

Framework for certified robustness in pedestrian trajectory prediction for autonomous driving.

Code for pulmonary embolism risk stratification using vascular graphs and medical records.

A role-playing voice agent for improved turn-taking in multi-party spoken conversations.

Leaderboard tracking the performance of LLM agents on the LiveClawBench benchmark.

Dataset of agent trajectories used in the LiveClawBench benchmark.

Framework for integrating textual data into time series in the frequency domain.

A multimodal benchmark to evaluate deepfake detectors on out-of-distribution content.

Coding-agent adapter for auto-configuring scientific simulators like GEOS and OpenFOAM.

An agentic AI framework for deep scientific review and verification of manuscripts.

Evidence-Driven self-correction framework for long video understanding.

An agent multiplexer that lives in your terminal.

Stack used for the Supersede reinforcement learning environment.

Real-time gridlock prevention and traffic signal optimization framework.

A benchmark for evaluating hidden social norm compliance in embodied planning agents.