Workshop Description
Much of the data that is fueling the current rapid advances in data science and computer vision is very high dimensional or complex. This poses challenges both in terms of building algorithms that can capture meaningful structure and also building analytical techniques that help to understand what that structure means. Mathematicians working in topology, algebra, and geometry have more than a hundred year’s worth of finely-developed machinery whose purpose is to give structure to, help build intuition about, and generally better understand spaces beyond those that we can easily visualize. This workshop will show-case work which brings methods from topology, algebra, and geometry and uses them to answer challenging questions in computer vision while often posing new questions in mathematics in the process. We interpret mathematics broadly and welcome submissions ranging from manifold methods to optimal transport to topological data analysis to mathematically informed deep learning. We envision our session as an opportunity for the researchers building state-of-the-art methods to connect with researchers who have challenging computer vision problems for which standard off-the-shelf techniques do not work.
With this workshop we hope to create a vehicle for disseminating computer vision techniques that utilize rich mathematics and address core challenges in computer vision as described in the ICCV call for papers. Our intention is to build community and facilitate increased exposure of innovative approaches rooted in mathematical theory and understanding. We expect the approaches to address a specific challenge and demonstrate utility on interesting datasets while lowering the barrier for entry with respect to comparison to other approaches or across multiple datasets. Mathematically derived techniques address a specific problem and while they may provide invaluable insights on novel real-world datasets they may not yield strong performance gains on many data-rich benchmarking datasets. Through intellectual cross-pollination between data-driven and mathematically-inspired communities we believe this workshop will support the continued development of both groups and enable new solutions to problems in computer vision.
Topic Areas of Interest Include, but are not limited to:
Geometric Deep Learning
Optimal Transport
Topological Data Analysis
Graph-based Methods
Manifold Methods
Abstract algebra in computer vision
Important Dates:
Paper Submission Deadline: July 10, 2021
Final Decisions to Authors: August 3, 2021
Camera-Ready Deadline (required for inclusion in proceedings): August 17, 2021
Main Conference: October 11-17, 2021
Workshop Date: TBD
Contact information: For questions and comments, please write to tagcv2021@gmail.com.