The 3rd Workshop on Software and Hardware Co-design of Deep Learning Systems in Future Architectures (SHDA) invites researchs who are working on new directions for co-design of machine learning systems in future architectures from academia, government, and industry. 

With the recent advancements in artificial intelligence, deep learning systems and applications have become a driving force in multiple transdisciplinary domains. This evolution has been supported by the rapid improvements of advanced processor, accelerator, memory, storage, interconnect and system architectures, including architectures based on future and emerging hardware (e.g., quantum, superconducting, photonic, neuromorphic). However, existing research is focused on hardware accelerators, deep learning systems and applications separately, but the co-design among them is largely underexplored. To develop high-performance deep learning systems on advanced accelerators, our workshop will focus on the following three important topics: 

1. adaptive deep learning model exploration and training for target inference devices based on customized user demands, 
2. joint optimization of deep learning model design with future accelerator architecture/compiler design, and 
3. how to leverage state-of-the-art computational functionalities from advanced accelerators for application optimization. 

The scope of the workshop will not be limited to the above topics and the organizers will consider other topics that are closely related.

SHDA held in conjunction with The SupercomputingAsia (SCA)/The International Conference on High Performance Computing in Asia-Pacific Region (HPCAsia), which will take place in Osaka, Japan, from January 26-29, 2026. This half-day workshop will be held on the afternoon of January 26th and will include sessions for 20-min presentations, invited talks, and panel discussions. 

All accepted papers will be published in an ACM-affiliated proceedings publication.

The workshop for the website can be found here: https://shda-workshop.github.io.

Submission Guidelines

All authors should submit manuscripts for review in a single-column format. Manuscripts must be a minimum of 12 pages and at most 18 pages in PDF format including figures and references, formatted according to the ACM Proceedings Style. The ACM templates for Word and LaTeX are shown below. Also, ACM has partnered with Overleaf and the ACM LaTeX template on Overleaf platform is available. Please follow STEP 1 (Microsoft Word/LaTeX) at “2. The Workflow and Templates” in https://www.acm.org/publications/authors/submissions. Please note that there are two different instructions for Word and LaTeX template users.

Manuscripts should be submitted through the EasyChair system.

Submission deadline: 7 November 2025 (AoE)

Author notification: 28 November 2025 (AoE)

Camera-ready paper: 12 December 2025 (AoE)

Committees

Organizing and Program Committee

• Masaaki Kondo, RIKEN Center for Computational Science (R-CCS) and Keio University
• Jong Choi (choij@ornl.gov)
• Shruti Kulkarni (kulkarnisr@ornl.gov)
• Seung-Hwan Lim (lims1@ornl.gov)
• Tong Shu (Tong.Shu@unt.edu)
• Elaine Wong (wongey@ornl.gov)

Contact

All questions about submissions can be emailed to any of the organizers.