There have been significant advances in computer vision in the past decade. Current computer vision systems usually learn a generic model. In order to handle the diversity of the visual world, the current approach is to scale up the model. Although scaling has been proven effective in the era of large language model, I argue that there are also other alternative approaches we should explore. In this talk, I will introduce some of our recent work on building robust computer vision systems via adaptation and continual learning. Instead of learning and deploying one generic model, our goal is to learn a model that can effectively and continuously adapt itself to different environments. I will present applications of this framework in several computer vision applications.

杏吧原创 the Speaker: Yang Wang is currently an associate professor in the Department of Computer Science and Software Engineering, Concordia University. Previously, he was a faculty member at the University of Manitoba. During 2020-2022, he worked as the Chief Scientist in Computer Vision at the Consumer Business Group, Huawei Canada. He did his PhD from Simon Fraser University, MSc from University of Alberta, and BEng from Harbin Institute of Technology. Before joining UManitoba, he worked as a NSERC postdoc at the University of Illinois at Urbana-Champaign. His research focuses on computer vision and machine learning.

Moderator: Prof. Yuhong Guo

This event will be in-person. Light refreshments will be served.

Putting Test-Time Adaptation to the Test: Updates for Robust Visual Recognition

When the data shifts between training and testing the accuracy of predictions can degrade. The problem is that the data changes, but our聽systems remain the same. Does it have to stay this way? In this talk we will examine how visual recognition can adapt and generalize to new and different data during testing. We will cover natural shifts, such as image corruptions, to highlight opportunities to update models (by entropy minimization and parameter mixing) and inputs (by diffusion) when the data differs. Then we will take a critical look at adversarial attacks through our case study of test-time adaptive defenses. On this front more adaptation is still needed in practice for empirical robustness, but there is progress in theory for certified robustness with our adaptive extension of randomized smoothing. As a last step, we will discuss what is next for test-time updates.

聽is an assistant professor at UBC in Vancouver, member of the Vector Institute, and senior research scientist at Google DeepMind. His research is on visual recognition, self-supervised learning without annotations, and robustness by adaptation. He earned his PhD at UC Berkeley advised by Prof. Trevor Darrell. He was the lead developer of the Caffe open-source deep learning framework from version 0.1 to 1.0. His research and service have received awards including the best paper honorable mention at CVPR’15 for fully convolutional networks and the Mark Everingham award at ICCV’17, the open-source award at MM’14, and the test-of-time award at MM’24 for Caffe. He is new to Canada and is excited to learn more about all of its provinces and seasons!

Moderator: Prof. Jim Green, Department of Systems and Computer Engineering

RSVP today for the CUIDS Distinguished Speaker Series.

Please click for more event details and to purchase a ticket.

State of the Climate 2024 and the Role of Nature-Based Climate Solutions

Join Prof. Damon Matthews as he provides an update on the state of global climate change, including anticipated future trajectories and the consequent implications for Canada and the world. He will speak about the potential role of nature-based climate solutions: efforts to leverage natural systems to store and sequester carbon while also contributing to other environmental and social goals. In this context, he will present some recent research showing that investing in natural systems has climate value, but that this value is of a different character than fossil fuel emissions reductions, raising questions about the appropriate use of carbon offsets as a contribution to climate mitigation goals.

is a Professor in the Department of Geography, Planning and Environment at Concordia University. Damon holds a PhD in climate science from the University of Victoria, and is a member of the College of New Scholars, Artists and Scientists of the Royal Society of Canada.聽He has published more than 130 research papers on topics ranging from quantifying the remaining carbon budget to assessing equitable approaches to allocate emission allowances to individual countries.聽He is internationally recognized for his work in policy-relevant climate science, as well as for initiatives such as the聽 that use digital visualization and web-based technologies to motivate and accelerate climate action.聽Damon is the Scientific Director of聽, and directs the NSERC CREATE in聽Leadership in Environmental and Digital Innovation for Sustainability (LEADS)聽program, which aims to train graduate student researchers at the intersection of sustainability science and digital innovation.

Moderator: Tracey Lauriault – Associate Professor in the School of Journalism and Communication

RSVP today for the CUIDS Distinguished Speaker Series.

Data Day includes presentations by 杏吧原创 researchers and industry experts, a keynote address, panel discussion and networking opportunities, as well as a poster competition to showcase graduate student research in data science across all disciplines.

Don鈥檛 miss our tenth annual Data Day, a must-attend event for those studying or working in data science, analytics and related fields!

Data Day 10.0 will focus on聽The Next Decade for Data Science. In keeping with this theme, the event program includes sessions on explainable AI, data science and heath care, and data science and the environment.