AI in Research Institution

Zhejiang Cancer Hospital and Alibaba DAMO Academy developed DAMO GRAPE, the first non-contrast CT AI for gastric cancer screening, achieving 85.1% sensitivity and 96.8% specificity, with retrospective studies showing earlier detection of 11 cases by 2-10 months. The model is being deployed in Zhejia

Creval Sistemi e Servizi, internal IT unit of Italian banking group Credito Valtellinese (Creval), deployed IBM Watson-powered virtual assistant Alfredo across 360+ bank branches, reducing back-office service desk call volume by 80%.

Berkeley Lab's A-Lab is an autonomous solid-state synthesis platform that combined ML-driven synthesis planning with robotic powder mixing and characterization to synthesize 36 of 57 target inorganic materials in 17 days, spanning 33 elements and 40 structural prototypes.

China Agricultural University's 'Shennong' (神农) agricultural LLM — version 3.0 (Oct 2025) with 36 specialized agents — serves 100,000+ Chinese farmers, covers 90% of ag disciplines and 80% of ag scenarios, with a 3 PB knowledge base (10M knowledge-graph entries, 20M labeled images, 50M production records). Built on MOE architecture, training cost cut in half from v2.0. Real outcome: a Shenyang corn farmer cut cost from ~480 yuan/mu to <400 yuan/mu in 2025.

iAorta, an AI system co-developed by Alibaba DAMO Academy and Zhejiang University's FAHZU, detects acute aortic syndrome (AAS) from routine non-contrast CT scans in emergency departments. Deployed at 10 hospitals in Zhejiang Province, it cut AAS misdiagnosis from 48.8% to 4.8% (90% reduction) and shortened diagnosis time from 4.3 hours to 1.7 hours. Published in Nature Medicine 2025 and entered NMPA Innovation Device Special Review in 2026.

University of Cape Town's EduBot served 43,600 students and 245,000 answers in 90 days post-AI upgrade, with 93% resolved without human help, 4.8/5 satisfaction, and 28% drop in human handoff requests. Supports 65 languages, 6,000 international students.

The British Antarctic Survey PolarDT project (2021-2030) is building digital twins of the Antarctic and Arctic environments for climate monitoring, sea ice analysis, and environmental research. Using AI and machine learning, the initiative creates comprehensive digital replicas of polar regions to support scientific research and environmental decision-making. The project addresses critical gaps in polar monitoring by enabling more frequent, accurate, and comprehensive observations of rapidly changing polar environments through advanced simulation and data integration.

The National Supercomputing Center Changsha locally deployed DeepSeek R1 model on domestic MXChip GPUs, launching Hunan's first city-level application CS-DeepSeek for urban management, social services, and public safety, alongside a medical research AI assistant for thoracic surgery clinical pathway optimization at Hunan Chest Hospital.