[2019] Vaishak Belle is one of 43 new Royal Society University Research Fellows (URFs) for 2019. He took up the post from the start of October. He will work on Efficient Inference and Learning in Probabilistic Logical Models. Image While artificial intelligence (AI) has revolutionized scientific, governmental and commercial enterprises alike, many fundamental problems remain for enabling systems that are scalable and robust while also being able to leverage human knowledge. In that regard, unifying logical reasoning and probabilistic learning is a longstanding goal for AI, which has led to the development of so-called probabilistic logical models. Vaishak’s project looks at issues pertaining to the expressiveness versus tractability trade-off for reasoning and learning with such models. In particular, the project's aims are to develop new foundational and algorithmic ideas for such models over big, uncertain, possibly continuous data. He is especially interested in understanding: whether we can push the envelope on efficient reasoning?; whether we can push the envelope on efficient learning?; and finally whether we can develop paradigms for abstracting models, in the sense of how predictions at the level of the data representation can be mapped to a user's understanding of the domain. In the long term, our inquiry is motivated by the growing interest in so-called explainable AI, human-in-the-loop decision making and the democratization of data, in allowing non-experts to communicate effectively with machines and understand its decisions. Vaishak BelleSchool of Informatics The University Research Fellows is a scheme for outstanding scientists who are in the early stages of their research career and have the potential to become leaders in their field. Related links Royal Society announces University Research Fellowships for 2019 This article was published on 2024-03-18
