Frontiers of AI Accelerators: Technologies, Circuits and Applications

Key topics

There is a strong demand for new technology and hardware architecture for more data-centric applications. Approaches are highly multidisciplinary in nature and the AI hardware problem requires complex specification matrices, including technology options, power reduction, timing acceleration, architectural optimisation, reliability enhancement, and application developments. To meet the increasing demands of large-scale data processing for data-centric applications, we will require expert knowledge in the following areas: • Advanced semiconductor memory technology • Integrated circuit design methodology • Data-centric system hardware architectures • Electronic Design Automation (EDA) tool development • Information coding and representation • Application specifications and innovation As a result, international experts specialised in the above areas interested in exploring the emerging area of AI accelerators will be invited to give in-depth lectures on the subjects they specialise in, with an emphasis on AI applications. Senior researchers from local organisations and academia will also be invited to conduct in-depth communication and exchange of ideas.