An expert-routing-based watermarking scheme for Mixture-of-Experts LLMs.

An offline voice dictation and speech-to-text solution for the Linux desktop.

Benchmark framework for out-of-distribution detection in reinforcement learning.

A method for mining published LP/MILP formulations into a reproducible dataset and taxonomy.

A lightweight streaming multimodal memory system for everyday-life assistance.

A framework for symbolic music generation based on explicit edit operations.

A robust generative watermarking framework for AI music generation

A fail-closed boundary monitor to block stale durable-effect attempts by LLM agents

Analysis of how data imbalance helps capable models avoid spurious correlations and shortcuts.

Offline agent policy optimization framework using DPO-style preference objectives on expert trajectories.

Evaluation framework and dataset for agentic abstention in LLM agents.

A framework for Riemannian Batch Normalization over Lie groups.

Sleep-based Hierarchical Accelerated Replay for long-range non-stationary temporal pattern recognition.

Online 3D Gaussian Splatting pipeline for surgical video SLAM and reconstruction.

A tool/dataset for recovering type representations from hidden states of code models.

A framework for failure localization in LLM-based multi-agent systems using judging mechanisms and multi-perspective verification.

Benchmark for measuring proactivity and autonomous problem solving in LLM agents.

A benchmark for evaluating LLMs in practical data analysis scenarios using governmental open data.

Local tool to identify which LLM calls a cheaper model could handle to reduce costs.

Software to optimize and accelerate AI inference across diverse AI chipsets.