Report location: Computer Building313
Report time:February 19, 2025 (Wednesday) from 4:30-6:00 pm
Report title: Vision-oriented-Augmented Bet365 lotto review of multimodal pretrained model for semantic association learning
Profile:
Li Zechao, School of Computer Bet365 lotto review and Engineering, Nanjing University of Technology/Professor and Vice Dean of the School of Artificial Intelligence/Software, whose research interests are mainly Bet365 lotto review intelligent analysis, computer vision, etc., presided over the National Outstanding Youth Science Foundation, the National Science and Technology Major Project for the New Generation of Artificial Intelligence, and the National Natural Science Fund joint fund key projects, Jiangsu Province climbing project, Jiangsu Province Outstanding Youth Project, etc.; selected as the National "Ten Thousand Talents Program" young talents; published more than 70 CCF Class A journals and conference papers; won 2 first prizes for science and technology in Jiangsu Province
Report introduction:
In recent years, in exploring the possible development directions of general artificial intelligence, Bet365 lotto review large models have become an important direction that has attracted much attention and have attracted widespread attention from the academic and industrial circles. The research tasks of Bet365 lotto review large models cover multiple aspects such as Bet365 lotto review question and answer and reasoning, graphic and text generation, image understanding and reasoning.-Research on Bet365 lotto review enhancement of multimodal pre-training model for semantic association learning, and adaptation of multimodal pre-training big models and downstream vision-semantic association learning tasks are carried out around the two aspects of external Bet365 lotto review and internal Bet365 lotto review. The problem study focuses on the adaptation of downstream tasks such as small sample recognition, image understanding, visual question and answer, semantic segmentation, image retrieval, visual positioning, etc. based on multimodal pre-trained large models, and finally introduces the application situation in actual business.
