About
Research & Development
Unmasking ‘Reasoning Distraction’: A New Threat to AI Reliability
TLDR: Researchers identified "reasoning distraction," a new vulnerability where irrelevant but complex tasks embedded in prompts significantly reduce Large Reasoning Models' (LRMs) accuracy, sometimes...
Research & Development
Boosting Language Model Reasoning with Targeted Exploration Rewards
TLDR: MERCI is a new Reinforcement Learning algorithm designed to improve how Large Language Models (LLMs) explore solutions in complex reasoning tasks. By simplifying...
Research & Development
ELLSA: An AI Model for Integrated Multimodal Interaction
TLDR: ELLSA (End-to-end Listen, Look, Speak and Act) is a new AI model that unifies vision, text, speech, and action into a single, full-duplex...
Research & Development
DeepAnalyze-8B: An Agentic LLM for End-to-End Data Science Automation
TLDR: DeepAnalyze-8B is the first agentic large language model designed for fully autonomous data science, capable of handling the entire pipeline from raw data...
Analytical Insights & Perspectives
The Rise of Local AI: OpenAI’s GPT-OSS-20B and NVIDIA RTX AI PCs Drive a New Era of Personalized Generative AI
TLDR: Generative AI is shifting from cloud-centric to local, private applications, primarily driven by OpenAI's release of the open-source and open-weight GPT-OSS-20B model and...
Research & Development
AI Agents Enhance Corporate Credit Analysis Through Structured Debate
TLDR: A study developed two AI systems for corporate credit assessment using non-financial data: a single-agent system (NAS) and a multi-agent debate system (KPD-MADS)....