REU Site: Advances of Machine Learning in Theory & Applications (AMALTHEA)

点击翻译按钮获取中文摘要

The Advances of MAchine Learning in THEory and Applications (AMALTHEA) Research Experiences for Undergraduates (REU) Site aims to provide top quality educational experiences to a diverse community of undergraduate students through research participation in the area of Machine Learning (ML). The relevance and importance of ML is not limited to specialized technological innovations, as it was in the past. Nowadays, it also increasingly influences everyday life through its contributions to applications such as voice/face recognition, credit fraud detection, intelligent recommendation systems and many others. Furthermore, ML is inherently multi-disciplinary as it draws from advances in multiple disciplines such as engineering, computing, statistics, mathematics, physics and biology, to name a few major ones. Since its start in 2009, AMALTHEA, one of the first ML-focused REU sites, involves 10 undergraduate students per year from a broad spectrum of disciplines, and the educational experience spans 10 weeks in the summer. The participants are exposed to cutting-edge ML research, as well as professional development activities, such as technical seminars and career-related workshops. Moreover, these participants perform closely-mentored research, whose results are going to impact the field of ML itself, as well as how ML is applied in other scientific disciplines. Over the 2016-2019 time span, the project will directly impact a diverse group of 30 motivated students, the majority of which may not have access to such research participation opportunities otherwise.The project's thrust area is the theory of ML and how it can be integrated and applied to important real-life problems, hence exposing participants to both theory and applications. Past projects include applications such as automated anuran recognition from frog calls, uncovering criminal networks from crime locations, human-object interaction recognition, modelling of group dynamics in virtual worlds, speaker-independent speech recognition and license plate recognition among others. On the other hand, past contributions to the theory of ML have been steered towards topics such as functional data analysis, wavelet-based density estimation, non-linear dimensionality reduction, kernel methods and anomaly detection to name a few. Short video highlights of such projects can be found on AMALTHEA's YouTube channel located at https://goo.gl/2JhYoF. Finally, additional information about these topics and their outcomes can be found on the project's web site located at http://www.amalthea-reu.org.