The effort to integrate logic with deep learning has intensified in recent years and has the potential to give rise to a new computational paradigm in which symbolic knowledge is used to assist deep learning systems or extend their capabilities, while offering, at the same time, a path towards the grounding of symbols and the induction of knowledge from low-level sensory data.
The synergy of these two different worlds is the topic of the workshop “When deep learning meets logic”, which will take place online from February 15 to 17th, 2021. This workshop aims at the following: (i) present the applications that are enabled by this computational paradigm; (ii) explore the state-of-the-art and understand its level of maturity and adoption by the industrial sector; in particular, techniques and theory developed from both the deep learning and the logic communities will be presented; and (iii) identify some of the big questions that are open in this area and single out problems that require further investigation.
According to the AI researcher Efi Tsamoura, who is leading the organization of this event at Samsung AI Centre - Cambridge, “When deep learning meets logic” is the first workshop that covers all aspects related to the integration of deep neural networks with logic and symbolic computation: from (machine) learning theory to a calculus defined to capture the brain cell computations. Not only “When deep learning meets logic” covers a wide range of topics, but also is the first workshop which presents both foundational and applied research on the area, such as applications in visual QA, QA over knowledge graphs and program induction. “Despite that the integration of these two different paradigms has been the topic of different workshops affiliated with venues like AAAI, IJCAI, NeurIPS and CVPR, our workshop is the first one which does not focus on a single topic but rather offers a holistic view over the exciting research area of mixing deep neural networks with symbolic systems” said Efi Tsamoura.
The speakers of the workshop will be world-class academics and industrial researchers coming from the logic, complexity and deep learning communities, including Professor Leslie Valiant from Harvard University who has been awarded the Turing award for his transformative contributions to the theory of computation; and Professor Christos Papadimitriou from Columbia University who has won multiple awards for this contributions to complexity and game theory including the IEEE’s John von Neumann Medal, the IEEE Computer Society Charles Babbage Award and the Gödel Prize.
The full list of speakers is: Professor Leslie Valiant from Harvard University, Professor Martin Grohe from RWTH Aachen University, Assistant Professor Jiajun Wu from Stanford University, Associate Professor Daisy Zhe Wang from the University of Florida, Professor Madhusudan Parthasarathy from the University of Illinois, Balder ten Cate from Google, Assistant Professor Jacob Andreas from Massachusetts Institute of Technology (MIT), Associate Professor Le Song from Georgia Institute of Technology, Ryan Riegel from IBM Research and Professor Christos Papadimitriou from Columbia University.
The workshop is expected to bring many benefits to Samsung Research including getting insights on new applications, attracting new talent and positioning Samsung Research as the world leader in AI. However, not only Samsung will benefit from this event but also the broader academic community: registering to the workshop is free and open to the public, while the recorded talks will be provided online and without any registration. The workshop aims to bring together the best people and to learn from them.
“When deep learning meets logic” is sponsored by Samsung Research and is co-organized by Efi Tsamoura (Samsung AI Centre - Cambridge), Vaishak Belle (University of Edinburgh), Phokion Kolaitis (University of California, Santa Cruz and IBM Research) and Loizos Michael (Open University of Cyprus).
The workshop’s schedule is available in the workshop website at
To register, please follow the link: