Researchers at Northwestern Polytechnical University have unveiled a new AI system that filters out irrelevant video content and uses progressive reasoning to boost multi-turn video-chat accuracy by up to 7%, setting new benchmarks and enabling smarter visual assistants across fields.

A new AI-driven image recommendation method that learns directly from shoppers’ clicks delivers smoother, more personalized fashion suggestions—boosting offline accuracy by 0.46% and driving a 0.88% sales lift on Taobao’s women’s clothing.

Researchers from Fudan University, the Chinese Academy of Sciences, and Peking University have developed a prediction-guided testing method that cuts quantum device verification experiments by half, slashing time and cost for next-generation quantum computer calibration.

Researchers have unveiled a real-time dashboard that detects, analyzes, and defends federated learning AI systems—boosting trust and security in sensitive fields like healthcare, finance, and smart grids.

Researchers at Soochow University have unveiled a transparent neural network that verifies whether events described in documents actually occurred and highlights the exact sentences that support its conclusions, boosting fact-checking accuracy and interpretability for journalists, legal teams, and policymakers.