2
B站成WAIC官方AI科技视频平台，月均超1.9亿用户消费AI内容
Bilibili has been designated as the official video platform for the World Artificial Intelligence Conference, reporting high user engagement with AI-related content.
2
国家统计局：经济新动能不断成长壮大，日益挑起中国经济的大梁
China's National Bureau of Statistics reported that high-tech manufacturing and AI-related industries, such as integrated circuits, experienced significant growth in the first half of the year.
2
Prompt-Driven Exploration
This paper proposes using large language models and vision-language-action models to guide global exploration in reinforcement learning through prompt-driven perturbations.
2
CLAP: Direct VLM-to-VLA Adaptation via Language-Action Grounding
The CLAP framework enables the direct conversion of pretrained vision-language models into vision-language-action models for robotic control with minimal architectural changes.
2
Has China AI Really Surpassed the US? Behind Two Viral Charts, Five Questions Worth Asking
An analysis of the AI competition between China and the United States examines talent development and authorship data from major conferences like NeurIPS.
2
How Far Has China Built the Heart of Low-Altitude Aircraft? The State of Domestic Aviation Electric Propulsion
The Chinese aviation industry is advancing its low-altitude economy through new electric propulsion engine deliveries and regulatory certifications for electric aircraft.
2
文远知行L2++智驾业务加速出海，一段式端到端方案驶向全球
WeRide is expanding its L2++ autonomous driving technology internationally, conducting road tests for its end-to-end driving solution in Germany, France, and Japan.
2
An LLM-powered Agentic Recommendation System for Connected TV Content Discovery
The paper presents an agentic recommendation system powered by large language models to improve content discovery on connected televisions by integrating unstructured contextual signals.
2
OpsMem: Dual-Memory Reasoning with Cross-Memory Resonance for Failure Diagnosis
OpsMem is a dual-memory framework designed to help large language models diagnose software failures by coordinating short-term diagnostic states with long-term operational experience.
2
Neutralizing Structural Inequality in the Nigerian FinTech Sector
This paper proposes a hierarchical human-AI triage model designed to mitigate algorithmic discrimination caused by infrastructure-related network noise in Nigerian financial fraud detection.
2
马斯克据悉要求特斯拉员工转向使用Grok
Elon Musk has reportedly instructed Tesla employees to transition to using xAI's Grok 4.5, citing its lower token costs compared to competitors.
2
芯片商Etched累计融资近60亿，获台积电、辛顿、李飞飞押注
Silicon Valley AI chip startup Etched has raised a total of 800 million dollars at a five billion dollar valuation to develop Sohu, an ASIC designed exclusively for running Transformer models.
2
GPT-5.6一小时解开50年数学猜想，700词Prompt驾驭64个子Agent
A new prompting method using 64 multi-agent systems reportedly enabled an advanced language model to solve a 50-year-old mathematical conjecture within an hour.
2
Clash of the tech titans: China’s quiet industry vs loud US innovation
This opinion piece compares the leadership styles and innovation strategies of the US and Chinese tech industries, contrasting Western visibility with Eastern discipline.
2
[AINews] OpenAI launches GPT 5.6 Sol/Terra/Luna, Codex becomes ChatGPT superapp
OpenAI has released the GPT-5.6 model series, including Sol, Terra, and Luna variants, while integrating Codex capabilities into the ChatGPT platform.
2
Introducing Real World VoiceEQ: Measuring the human quality of voice AI
Hugging Face has introduced Real World VoiceEQ, a new framework designed to evaluate and measure the human-like quality of voice-based artificial intelligence.
2
腾讯洽购Manus？知情人士：腾讯仍将保留少数股东地位
Reports suggest Tencent is negotiating a deal involving AI agent startup Manus, though sources indicate Tencent will maintain a minority stake rather than acquiring a controlling interest.
2
KV-PRM: Efficient Process Reward Modeling via KV-Cache Transfer for Multi-Agent Test-Time Scaling
The paper proposes KV-PRM, a method that uses KV-cache transfer to make process reward modeling more computationally efficient during multi-agent test-time scaling.
2
TrustX Agent Risk Classification Framework (ARC): Risk-Tiering Internally Created Agentic AI Systems
The TrustX Agent Risk Classification Framework is introduced as a structured instrument to classify and govern internally developed agentic AI systems within enterprises.
2
Accelerating GPU Inference of Large Language Models with Moderately Unstructured Sparse Weight Matrices
Researchers developed a new GPU kernel designed to accelerate large language model inference by efficiently handling moderately unstructured sparse weight matrices.
2
SCATE: Learning to Supervise Coding Agents for Cost-Effective Test Generation
SCATE is a framework designed to reduce human supervision costs by training models to guide autonomous coding agents during automated test generation.
2
When your brain works differently, AI isn’t a luxury—it’s accessibility
An author describes how AI-powered desktop assistants can serve as crucial accessibility tools to support neurodivergent professionals with executive functioning.
2
Mitigating Early Training Collapse in CTR Models
This study analyzes the phenomenon of early training collapse in click-through rate prediction models and evaluates practical mitigation strategies using large-scale industrial datasets.
2
Verifier-Guided Twelve-Tone Composition: A Generate-Verify-Repair Harness for Symbolic Music Generation
A new neuro-symbolic framework improves the consistency of LLM-generated twelve-tone music by implementing a generate-verify-repair loop.