The 1st International Workshop on
Acceleration and Optimization of Multi-modal Computing (AOMC)

Call for Papers

With the rapid development of various powerful sensors and mobile devices, multimodal applications which leverage speech, touch, vision, and gesture have attracted considerable attention. Multi-modal computing exhibited impressive accuracy improvements over traditional uni-modal DNNs. However, this accuracy improvement comes at the cost of significantly increased computation, which limits the practical value of multi-modal DNN applications, especially in energy- and latency-constrained scenarios. Therefore, it is desirable to accelerate and optimize multi-modal DNN applications to facilitate the deployment of real-life multi-modal DNN applications. 

The workshop will focus on the acceleration and optimization of multi-modal DNN applications. The primary objective of this workshop is to catalyze advancements in computer architecture specifically tailored for multi-modal DNN applications. We are inviting proposals from researchers and developers actively engaged in the field. Proposals should emphasize novel approaches and technologies in accelerating multi-modal DNN applications through innovative computer architecture solutions. We encourage submissions that offer insights into research progress, practical challenges, and future directions. Examples of this workshop include, but are not limited to:

• Impact of multi-modal DNN applications on computer architecture research 
• APIs and programming models for multi-modal DNN applications 
• IoT, mobile and embedded system design for multi-modal DNN applications 
• Emerging memory technologies and their impacts on multi-modal applications 
• Specialized multi-modal DNN accelerators and reconfigurable systems 
• Hardware-software co-design for multi-modal DNN applications 
• Design space exploration for multi-modal DNN applications 
• Implementation and optimization of multi-modal computing systems