Keynote Speakers

DeepTutor: The Nature and Effectiveness of a State-of-the-Art Conversational Intelligent Tutoring System

Dr Vasile RusProf. Dr. Vasile Rus
William Dunavant Professor
Director of the Data Science Center
Systems Testing Research Fellow of the Fedex Institute of Technology
Department of Computer Science (CS)
Institute for Intelligent Systems (IIS)
The University of Memphis

DeepTutor: The Nature and Effectiveness of a State-of-the-Art Conversational Intelligent Tutoring System

  Abstract: I will present in this talk the design and effectiveness of DeepTutor, a state-of-the-art dialogue-based intelligent tutoring system that promotes deep learning of science topics. Tutoring is reportedly one of the most effective forms of instruction that often yields superior learning gains compared to other forms of instruction such as reading a textbook or traditional classroom instruction. Encouraged by the effectiveness of one-on-one human tutoring, computer tutors that mimic the human tutoring process have been successfully built with the hope that a computer tutor could be afforded by every child with access to a computer. Nevertheless, building computer tutors is computationally challenging. I will describe some of challenges and solutions we proposed such as a branch-and-bound algorithm to address the student input assessment task, which we modeled as a combinatorial optimization problem. Furthermore, I will present the results of a large scale, after-school experiment with high-school students that showed that DeepTutor is as effective as an average human tutor. I will conclude the talk with implications of our work on more general dialogue-based systems such as the emerging intelligent personal assistants like Apple’s Siri, Microsoft’s Cortana, Google’s Now, or Amazon’s Alexa.

BIO: Dr. Vasile Rus received his Bachelors degree in Computer Science from Technical University of Cluj-Napoca in June 1997 with a Diploma Thesis entitled Distributed and Collaborative Configuration Management, masterpieced while at LSR Laboratory, Grenoble Institute of Technology, France. Dr. Rus earned his Masters of Science in Computer Science and Doctor of Philosophy in Computer Science degrees from Southern Methodist University at Dallas, Texas in May 1999 and May 2002, respectively.

Dr. Vasile Rus joined the Department of Computer Science at the University of Memphis in 2004. He is also a member of the Institute for Intelligent Systems at the University of Memphis. Before that, Dr. Rus was an Assistant Professor at Indiana University (2003-2004).

Dr. Rus conducts state-of-the-art research and teaching in the area of language and information processing. He has been exploring fundamental topics such as natural language based knowledge representations, semantic similarity, and question answering as well as applications such as intelligent tutoring systems and software defect knowledge management. Dr. Rus has received research awards to support his work from the National Science Foundation, Institute for Education Sciences, Office of Naval Research, and other federal agencies. He has been a Principal Investigator or co-Principal Investigator on awards totaling more than $6.2 million. For his work on automated methods to handle software defect reports in large-scale software development projects, Dr. Rus has been named a Systems Testing Research Fellow of the FedEx Institute of Technology.

Dr. Rus has published more than 80 scientific articles in premier peer-reviewed international conferences and journals, as well as book chapters. He has received several best paper awards at international conferences, and all of his PhD students have earned research awards for their work under him. Furthermore, his students have received summer internships at prestigious research labs such as Stanford Research Institute, AT&T Research Labs, Vulcan, and IBM.

Among other accomplishments, Dr. Rus is an Associate Editor of the International Journal of Artificial Intelligence Tools and has been coordinating the local chapter of the North American Computational Linguistics Olympiad, which involves working with high-school students and teachers. In 2011, one of the local high-school students made it to Team USA and won a silver and gold medal at the International Linguistics Olympiad.


Deep Learning in Automotive Embedded Systems

Dr. Oliver Sbanski
Director Engineering Systems Vision
Robert Bosch GmbH

Deep Learning in Automotive Embedded Systems


Upcoming advanced driver assistance functions and automated driving demand near-perfect performance and maximum availability of sensor information. Using the example of lane keeping in scenarios without visible lane markings motivates, why existing algorithms need to be further developed and are currently complemented or replaced by deep learning technologies. Several examples of AI and deep learning applications in current Bosch automated driving systems are given in the presentation. One of the major challenges for automotive applications is the much higher resource consumption of deep learning approaches versus other machine learning techniques. Bosch has tackled this challenge and we will show that the deep learning performance of a 250 Watt GPU can be implemented into an embedded automotive camera using less than 5 Watt power if the target HW is considered consistently throughout the development.

BIO: Dr. Oliver Sbanski studied Physics at the University of Würzburg, Germany, and finished 2001 with a PhD in Physical Chemistry. He joined Bosch 2001 in the automotive business unit of active safety systems developing ABS/ESP SW & algorithms. Since 2011, he leads the system and computer vision development for video based driver assistance systems, including front facing cameras and surround view systems. Being a key sensor towards the Bosch vision of "safe, agile and automated driving for everyone", the Bosch video camera already today contributes significantly to comfortable and injure-free driving. Recent advances in embedded hardware design allow for implementation of leading edge AI & deep learning algorithms in automotive sensors, which take the performance and possibilities to new levels.