who: Veronika Thost
when: 10/17 (Thursday) 11:45a – 1:15p
where: Computer Science Building Rm 150
food: Athena’s Pizza
generous sponsor:  ORACLE LABS

“Knowledge Representations & Reasoning
Meets Machine Learning”

Abstract:
In many domains, there is structured knowledge which can be leveraged for reasoning in an informed way in order to obtain high quality answers. Symbolic approaches for knowledge representation and reasoning are less prominent today – mainly due to their lack of scalability – but their strength lies in the verifiable and interpretable reasoning that can be accomplished. In this talk, we present work about how symbolic knowledge representations and reasoning can be combined with machine learning. We will present new datasets to evaluate logical rule learning, show how reinforcement learning can support logical reasoning, and consider knowledge graphs as external information for learning (e.g., for recognizing textual entailment).

Bio:
Veronika is a Postdoctoral Researcher in the MIT-IBM Watson AI Lab at IBM Research, Cambridge, MA. Her work focuses on the combination of symbolic knowledge representations and reasoning with learning: on learning logical theories and to guide reasoning over such theories, and on how the information in knowledge graphs can be used as structured, external knowledge to support learning.

Previously, Veronika worked as a Postdoctoral Researcher at TU Dresden, Germany, where she also received her Ph.D. in Computer Science in 2017. At that time, she worked on query answering over knowledge graphs, existential rules, and in description logics. Her work has been published at IJCAI, KR, ISWC, ESWC, KR, JWS, and TOCL.