Huang Fei
Chief Scientist and CAIO of AI at Honor Devices; former Vice President of Alibaba Cloud
Dr. Huang Fei has long been engaged in research and development in natural language processing and artificial intelligence. He currently serves as Chief Scientist and CAIO of AI at Honor Devices. Previously, he was Head of the Tongyi Natural Language Intelligence Lab at Alibaba and Vice President of Alibaba Cloud, as well as an adjunct professor at Zhejiang University. He led teams in the development of the Tongyi large language model system (including AliceMind, PLUG, Qwen, Embedding, DeepResearch, etc.), achieving human-surpassing results in machine reading comprehension, vision-language question answering, and Chinese language understanding, work that was recognized with the SAIL Award at the World Artificial Intelligence Conference. Building on the Tongyi Qianwen model, he guided his team to focus on industry applications, enabling software development (Tongyi Lingma / Qoder), role-based dialogue, office productivity, intelligent judicial solutions, and other industry use cases. Dr. Huang has published over 300 papers in top AI conferences and journals, holds more than 100 international and domestic patents, and has over 50,000 citations on Google Scholar with an H-index of 86. He has served as area chair and journal editor for ACL, TACL, and other academic venues, and as a member of large model expert committees. He has also shared insights and technology perspectives at numerous industry conferences and government-organized seminars. He earned his Ph.D. in Computer Science from Carnegie Mellon University and previously worked at IBM Research and Facebook, where he contributed to and led the development of machine translation systems that help 2 billion Facebook users overcome language barriers.
Topic
Beyond Cycles: The Evolution and Future of Artificial Intelligence
This talk will take “technology cycles” as its guiding thread, briefly reviewing the evolution of artificial intelligence, especially large models, and analyzing the dominant technical paradigms, capability boundaries, and real-world deployment challenges of today’s mainstream models. It will explore how technological progress is shifting from model capabilities to architectural capabilities, how resource investment is expanding from pure compute competition to a balance of efficiency and performance, and how the cognitive role of AI is evolving from productivity tools to organizational insight. Against practical constraints such as compute, data, security, and cost, the talk will further discuss how AI systems represented by large models can pursue sustainable and evolvable development paths, ultimately advancing toward augmented human intelligence (AHI).
