Creation of a Student Research Environment in Artificial Intelligence with Industrial Involvement

摘要

In this paper we describe ongoing efforts in creating and using an environment for undergraduate research projects in Artificial Intelligence (AI). Our goal is to provide hands-on and real-world experience to students. Industry involvement in this environment provides partnership and supervision, based on deliverables. Students learn real-world skills related to AI in industry and this makes them employable in the job market which is experiencing growing demand for AI applications. The initial direction of the undergraduate research supported by this environment is related to ground based surveillance from stationary cameras where various aspects of human detection, human movement detection, and human movement classification are key components. The central theme of the research is how to use machine learning and computer vision to accomplish these goals and what hardware and software libraries should be selected. More specifically, we describe the research environment that consists of a laboratory with different stationary and remotely controlled cameras, including workstations with high end Graphical Processing Units (GPUs). Open source software libraries for machine learning and computer vision are used for advanced algorithm development in this environment. As a result our research environment has the following characteristics - it is geared towards attractive areas of machine learning and computer vision; and has preconfigured hardware and software components for experiments with data.