IRES: U.S.-Hungary Research Experience for Students on Non-Boolean Computer Architectures

点击翻译按钮获取中文摘要

The objective of this U.S.-Hungarian project is to provide early career international research experience (IRES) for U.S. students who will study emerging technologies based on cellular neural networks(CNN)with the goal of enhancing and developing information processing systems. The IRES program of activities will take place in Budapest, overseen by the principal investigator, Michael Niemier, in partnership with colleagues from the University of Notre Dame and faculty at Pázmány Péter Catholic University (PPKE). A CNN is an analog array processor architecture designed to significantly improve the power and performance of various computation-intensive applications, such as image processing, pattern recognition, and motion detection. Together, the U.S.-Hungarian team will pursue effective ways to move away from chip stored program models. By seeking additional functionalities and capabilities, the research mentors and their students expect to leverage new information processing technologies and algorithms to further enhance the functionality of and application spaces for the CNN paradigm.Technical Summary:Via collaboration and mentoring by faculty from the University of Notre Dame and Pázmány Péter Catholic University (PPKE) in Budapest, Hungary, the impact of emerging technologies on architectures, algorithms, and programming will be addressed by IRES student participants. Cooperative research activity will focus on determining whether emerging device technologies that are organized in a CNN-like fashion can be "programmed" and are amenable to template-based operations. U.S. student participants and their PPKE counterparts will study existing hardware designs for improvement by emerging technologies, possibly via lower energy, and if likely, what the resulting impact on programmability will be. For example, the viability of hardware realizations of new CNN-computing models (e.g., frameless computing and delay-based templates) that presently exist only mathematically will be explored. For new application spaces, team efforts will target tactile sensing and focus on developing CNN circuitry with integrated sensing technology. Also of interest are graphene based electronics and developing algorithms to process sensed data. If successful, the U.S.-Hungarian IRES team's combined research activities may lead us to consider new, integrated sensing and tactile data processing techniques for potential heavier load applications in areas such as robots or prosthetic feet.