AI Large Model Technology Exploration and Application Forum (CNCC2024)

The forum "AI Large Model Technology Exploration and Application" will be held on October 24‑26, 2024 at Hengdian, Dongyang, Zhejiang, as part of CNCC2024.

Rapid development of large language models (LLMs) has expanded AI applications across industries, yet three major capability gaps remain: limited domain knowledge, difficulty handling complex decisions, and insufficient collaborative dialogue.

Ant Group, an early adopter of AI large models, will co‑organize the event with Zhejiang University, Xi'an Jiaotong University, Renmin University and other experts to address these challenges from both industry and academia.

Key discussion questions include: how to endow LLMs with expert‑level knowledge, how to emulate human expert reasoning for professional decision‑making, and how to bridge the gap between general language processing and domain‑specific knowledge interpretation.

The program features a series of invited talks:

Knowledge‑Enhanced LLMs and Knowledge Governance – Zhang Zhiqiang (Ant Group), focusing on integrating symbolic knowledge into LLM pre‑training, prompting, chain‑of‑thought, RAG and model alignment.

Research Paradigm Shift Driven by Large Models – Chen Huajun (Zhejiang University), covering the impact of large models on scientific research.

Data Infrastructure and Broad‑Spectrum Association Analysis – Cheng Xianzhong (Chinese Academy of Sciences), presenting data space concepts, infrastructure architecture and a new paradigm for value release.

Vertical‑Domain Reasoning Enhancement for Large Models – Shi Bin (Xi'an Jiaotong University), proposing a few‑shot fine‑tuning framework, pseudo‑labeling with factor‑graph modeling, and a bootstrapped chain‑of‑thought method.

Advances in Diffusion Models – Li Chongxuan (Renmin University), discussing continuous diffusion theory, efficient sampling, video and 3‑D generation, and recent progress in discrete diffusion.

Ant’s Multimodal Large Model Progress and Outlook – Chen Jingdong (Ant Group), reviewing multimodal model achievements and future directions.

Alipay Medical Large Model Technology and Applications – Wang Jian (Ant Group), describing a medical LLM built on the Baoling foundation model with extensive medical corpora and knowledge graphs, achieving GPT‑4‑level performance.

Each speaker’s biography highlights their research achievements, publications in top conferences (NeurIPS, ICML, ICLR, etc.), awards, and contributions to AI platforms such as Baoling, AGL, and knowledge graph systems.

CNCC2024 will host 18 invited talks, 3 plenary forums, 138 specialized forums, 34 thematic activities and over 100 exhibitions, featuring more than 800 speakers including Turing Award winners and academicians, with an expected attendance of over ten thousand participants.