Kubernetes Topology Intelligence Engine for improving LLM security patch generation.

Reactive real-time flow policies for robotic manipulation.

A curated list of face de-identification literature, datasets, and open source code.

Automatic framework for modality-level inconsistency detection in large-scale knowledge auditing.

A decoupled audio-visual AIGC detection system for identifying forged audio-visual content.

A linear-time sequence model that uses learned, input-dependent routing to maintain sparse memory slots for better long-context recall.

An autonomous agentic system for generating formal specifications and performing model checking.

A scalable diffusion framework for real-world image restoration based on residual rectified flow.

Multi-agent ensemble framework for universal biomedical image analysis with offline policy-guided routing.

Full-stack cloud with forkable VM snapshots and global SQLite.

Framework for medical MCQA using dynamic knowledge graphs and evidence judgment.

Open-source Vision-Language Model used for generating textual summaries of tables and figures.

A distributable harness for embedding security controls into AI coding agents.

Spatio-temporal Transformer for nuclear radiation forecasting with atmospheric diffusion module.

A reusable, agent-aware instantiation for heterogeneous collaborative knowledge work with LLM agents.

Neurosymbolic framework for patching LLM reasoning traces using Inductive Logic Programming.

Compiler-grounded hierarchical diagnosis system for optimizing Triton kernels on Ascend NPUs.

A disentangled multi-view Mamba framework for network traffic anomaly detection.

Efficient long-term agent memory framework that constructs context selectively at query time.

A 2T-token English corpus for LLM pretraining with optimized data curation.