Current Status and Issues of in-memory Accelerators for Deep Neural Networks

Summary form only given, as follows. A complete record of the panel discussion was not made available for publication as part of the conference proceedings. Deep Neural Networks (DNNs) have been widely utilized in a variety of applications such as image recognition and natural language processing. However, a large number of computation is required to execute DNNs. One of the promising solutions to ease this heavy burden is to design and use in-memory accelerators. In this talk, the current status and issues of in-memory accelerators compared with typical digital-based accelerators are discussed with some examples. Additionally, we introduce some steps for making in-memory accelerators practical, and also how to perform effective benchmarking.