In recent years, the fields of graphics and vision have experienced unprecedented growth and innovation. The integration of advanced computational methods, particularly those involving AI and machine learning, has revolutionized how we process, understand, and interact with visual data. Specifically, compared to traditional graphics and vision applications that rely on physical models, the emerging applications incorporate neural networks and gradient propagation-based optimization processes to deliver photorealistic rendering quality and human-level understanding ability.
However, these newly introduced operations also pose challenges in computer system and architecture design to unleash the potential of emerging graphics and vision applications on local devices in daily life. It is imperative for the computer architecture community to engage in the development of emerging graphics and vision innovations at this early stage. By providing system and architecture insights, we can aid application developers in creating more efficient and effective graphics and vision applications through the co-design of algorithms, architecture, and systems.
The overarching goal of this workshop is to bring together researchers, developers, and users of emerging graphics and vision applications to present their latest research with focus on innovations in system and architecture. We believe that this workshop will encourage new perspectives on how system and architecture can enable and accelerate more emerging graphics and vision applications.
2024 International Symposium on Computer Architecture
Neural Radiance Fields
Diffusion Models
Neural Rendering
Point Cloud Processing
Vision Transformers
Vision Foundation Models
Hardware Acceleration of Transformers
Hardware Acceleration of Neural Graphics
Length and Format: Submissions should be up to 4 pages in length, excluding references. While this is not a strict limit, authors are encouraged to keep their submissions within this length for conciseness. All submissions must be in PDF format.
Template: Submissions should follow the ISCA'24 LaTeX Template. This template is designed to ensure consistency and readability for all submissions and must be adhered to strictly.
Submission Platform: All papers must be submitted through the specific OpenReview platform.
Review Process: The review process for the workshop will be double-blind. Therefore, authors must not include their names or affiliations on the submitted documents, except in the designated space on the submission form. The identity of authors and reviewers will be kept confidential from each other.
Yingyan (Celine) Lin, Georgia Institute of Technology
Yang (Katie) Zhao, University of Minnesota, Twin Cities
Selvakumar Panneer, Intel Labs
Abdulrahman Mahmoud, Harvard University
Muhammad Husnain Mubarik, SMTS Graphics and ML at AMD
Chaojian Li, Georgia Institute of Technology
For any questions, please contact us at evgaisca24@groups.gatech.edu