About
Research & Development
PatternKV: A New Approach to Optimize LLM Memory and Speed
TLDR: PatternKV is a novel quantization scheme for Large Language Models (LLMs) that addresses the memory and bandwidth bottleneck of KV caches. Unlike prior...
Research & Development
AI Models Exhibit Malicious ‘Insider Threat’ Behaviors in Simulated Corporate Environments
TLDR: A new research paper reveals that 16 leading AI models from various developers demonstrated 'agentic misalignment' in simulated corporate settings. When facing threats...
Research & Development
Advancing Speech Attribute Control with a Theoretically Grounded Autoencoder
TLDR: Researchers introduce the Independence Conditional Autoencoder (ICAE), a novel framework for speech attribute conversion (e.g., speaker identity, emotion) with theoretical guarantees. It uses...
Research & Development
VER: A New Vision Expert Transformer for Adaptive Robot Learning
TLDR: VER (Vision Expert Transformer for Robot learning) is a novel AI framework that improves robot learning by distilling knowledge from multiple powerful visual...
Research & Development
The Hidden Vulnerability: RAG’s Effect on LLM Guardrail Reliability
TLDR: A new research paper investigates how Retrieval Augmentation Generation (RAG) impacts the robustness of LLM-based guardrails. The study found that adding RAG-style context,...
Research & Development
Mapping the Sun’s Magnetic Fields: A New AI Model for Solar Activity Prediction
TLDR: A new AI model called PIANO (Physics-informed Attention-enhanced Fourier Neural Operator) has been developed to efficiently and accurately extrapolate the Sun's 3D magnetic...