研讨会信息
🎤 Speaker: Wangbo, ZHAO
A final-year Ph.D. candidate,
National University of Singapore
📰 Title: Dynamic Computation for Efficient AI: From Foundations to Applications
⏰ Time: 2:30 – 3:30 PM,
📆 Date: Nov. 17, 2025 (Monday)
📍 Venue: E4-201, GZ Campus
Online Zoom Meeting:
https://hkust-gz-edu-cn.zoom.us/j/92858266535?pwd=WrZCG5uUiUjAlZ2xAQMfVJaxahghqR.1
Meeting ID: 928 5826 6535
Passcode: ait
研讨会概要
Abstract
As AI models grow in scale and complexity, achieving efficiency and adaptability becomes increasingly critical. My research focuses on dynamic neural networks—models that can adjust their behavior and structure based on varying computational demands. This work aims to enable more efficient and scalable AI models by addressing key challenges in training, inference, and deployment:
1.Training: Traditional methods for transforming static models into dynamic ones often rely on full-parameter fine-tuning, which is prohibitively expensive for large-scale models. I propose leveraging parameter-efficient fine-tuning (PEFT) to construct dynamic models, significantly reducing training costs while maintaining performance.
2. Inference: While most research on dynamic networks emphasizes accelerating visual perception tasks, critical areas like visual generation and vision–language understanding remain underexplored. To address this, I extended dynamic modeling into these domains, introducing innovations such as DyDiT, DyDiT++, and RAPID3 for visual generation, as well as SGL for vision–language understanding. These advancements highlight the versatility of dynamic networks.
3.Deployment: Static models with fixed size and architecture are challenging to deploy across devices with varying computational and memory constraints. My proposed dynamic models adapt their size post-training, enabling seamless deployment across diverse platforms without retraining. This flexibility ensures efficient resource utilization from edge devices to cloud systems.
Looking forward, I aim to build a comprehensive and dynamic computational ecosystem by integrating model design with machine learning systems through joint optimization. This vision aims to foster the development of efficient, adaptable, and versatile AI models, driving meaningful progress toward real-world advancements in Artificial General Intelligence (AGI).
分享者简介
Wangbo, ZHAO
A final-year Ph.D. candidate,
National University of Singapore
Wangbo Zhao is a final-year Ph.D. candidate at the National University of Singapore. His research focuses on efficient deep learning, with a particular emphasis on dynamic models. He earned his master's and bachelor's degrees from Northwestern Polytechnical University in 2022 and 2019, respectively. Wangbo has published 22 papers in top-tier conferences and journals, including ICLR, NeurIPS, and CVPR, with nine of them as the first or corresponding author. He is a recipient of the President's Graduate Fellowship at NUS and the Google Ph.D. Fellowship 2025 in Machine Learning and ML Foundations. Currently, he is also a research scientist intern at Meta's SuperIntelligence Lab.
扫描加关注
获取更多AI学域消息
SCAN & FOLLOW US!