The article analyzes how AI agents are shifting from merely usable to trustworthy, outlines three testing paradigms, presents two detailed real‑world case studies in finance and healthcare, and proposes the TAME framework for sustainable, robust agent testing.
In 2026, over 68% of AI service outages stem from silent prompt failures, and this article details a four‑step, data‑driven methodology that raised prompt robustness to 99.2% in a provincial health‑insurance audit system, cutting error rates from 17.3% to 0.8% and latency by 19%.
The ERNIE Open Innovation Hackathon’s multimodal track showcased a diverse set of award‑winning projects that leveraged the ERNIE‑4.5‑VL model to dramatically shorten video‑production cycles, create audio‑only smartphone assistants for seniors, enable personalized autism‑intervention platforms, generate AI‑driven music for videos, and more, demonstrating the practical impact of multimodal AI across real‑world scenarios.
This article explores how Huawei Cloud's Pangu large‑model platform powers the TianShiLi Digital Herbal AI system, merging centuries‑old Chinese medicine knowledge with cutting‑edge AI to accelerate drug discovery, enable intelligent diagnosis, and transform the entire TCM research and development workflow.
Ant Group and the China Academy of Information and Communications Technology released a research report outlining a trustworthy application framework for large models in rigorous sectors such as finance and healthcare, detailing technical safeguards, industry case studies, and guidance for scalable, secure AI deployment.
The article outlines how artificial intelligence will drive transformative changes in 2024 across healthcare, finance, smart cities, everyday tech, and governance, highlighting emerging trends, innovations, and ethical challenges that will reshape our world.
The article introduces Dingxiangyuan's medical knowledge‑graph ecosystem, describes the construction of a four‑layer taxonomy, presents the ACL‑published SMedBERT model that injects structured medical semantics into a pre‑trained language model, and discusses its deployment in search, query expansion, and semantic matching while outlining future challenges.
In a TVP forum hosted by Li Kaifu and Shen Chunhua, experts trace AI’s 70‑year journey, discuss the origins of the book “AI Future in Progress,” analyze investment stages, AI‑cloud synergy, NLP breakthroughs, medical applications, societal impacts, data privacy, and the challenges facing traditional enterprises.
This article reviews Alibaba Cloud senior algorithm expert Chen Mosha's presentation on medical NLP, covering Alibaba's healthcare business, data types, electronic medical record quality inspection, span‑based and nested NER models, term normalization, clinical trial outcome prediction, knowledge‑enhanced language models, and the CBLUE benchmark dataset.
MedicalNet, Tencent’s open-source framework, aggregates diverse small 3D medical imaging datasets into a large pre-training corpus, applies dataset filtering and joint spatial-pixel normalization, and provides encoder-decoder models that accelerate convergence and boost accuracy for AI-driven diagnosis in data-scarce medical imaging scenarios.
Fei‑Fei Li, co‑director of Stanford’s new Human‑Centric AI Institute, discusses how AI’s intelligence and values must be taught by humans, the importance of democratizing AI, addressing gender and bias in tech, and her vision for AI’s role in healthcare, education, and societal good.
This article describes the challenges and solutions encountered while developing an AI-driven medical dialogue assistant, covering data acquisition, preprocessing, model selection such as DCNN and Bi‑LSTM‑CRF, question generation, and system architecture, with insights applicable to similar healthcare chatbot projects.
Tencent's AI Breast System combines mammography, pathology, MRI and ultrasound analysis using a multi‑scale, progressive TMuNet model that processes four views, learns from physician feedback, and delivers lesion localization, malignancy scoring and automated reports, achieving up to 92% sensitivity and reducing annotation time.
