Seminars and Events

Artificial Intelligence Seminar

Building Robust and Explainable AI Agents for the Real World

Event Details

In this talk, I will present our efforts in building natural-language AI agents that use common sense to act robustly in open-world situations. State-of-the-art technology is inadequate for this purpose: background knowledge and rules provide explainability but cannot generalize to unseen situations, whereas neural models with natural generalizability are prone to making silly mistakes and cannot explain their decisions. The lack of common sense in today’s models results in a lack of public trust and hinders the adoption of AI technology. I will describe our natural language agents that combine the best of the symbolic and the neural worlds to perform well on open-world tasks without the need for task-specific training data. I will describe our knowledge organization and enrichment efforts, together with our robust and explainable neuro-symbolic commonsense methods that reason over this commonsense knowledge. Our neuro-symbolic reasoning mechanisms include procedural reasoning, reasoning by analogy, and reasoning by imagination. I will discuss different strategies to design systems with native explainability, such as engineering the knowledge dimensions used for pretraining, generation of scene graphs, and learning to produce knowledge paths. I will conclude with our ongoing efforts to apply these agents to real-world challenges, such as intelligent traffic monitoring and socially assistive technologies.

Speaker Bio

Filip Ilievski is a Research Assistant Professor at the USC Viterbi School of Engineering and a Research Lead at the Center on Knowledge Graphs within USC’s Information Sciences Institute (ISI). Filip holds a Ph.D. in Natural Language Processing from the Vrije Universiteit (VU) in Amsterdam. His research focuses on developing robust and explainable neuro-symbolic technology with positive real-world impact, based on neural methods and high-quality knowledge. Filip has made extensive contributions in identifying long-tail entities in text, performing robust and explainable commonsense reasoning, and managing large-scale knowledge resources. Over the past three years, he mentored around twenty Master’s and Ph.D. students, and has been collaborating with researchers at USC, CMU, Bosch Research, RPI, University of Amsterdam, and the University of Lyon. Filip has over 40 peer-reviewed publications in top-tier venues on commonsense reasoning, information extraction, and knowledge graphs. He has also been actively organizing workshops (AAAI’21), tutorials (AAAI’21, ISWC’20, ISWC’21, TheWebConf’22, KGC’22), symposiums (USC), and a special journal issue (Semantic Web Journal) on these topics.

YOU ONLY NEED TO REGISTER ONCE TO ATTEND THE ENTIRE SERIES – We will send you email announcements with details of the upcoming speakers.

Register in advance for this webinar:

After registering, you will receive an email confirmation containing information about joining the Zoom webinar.

IF speaker approves to be recorded for this AI Seminar talk, it will be posted on our USC/ISI YouTube page within 1-2 business days:

Host: Deborah Khider, POC: Alma Nava