When: Apr 28, 2021 – 11:00 – 11:45 AM
Where: Google meet link
Many tasks in AI can be divided into roughly two categories: those that require low-level perception, and those that require high-level reasoning. At the
same time, there is a growing consensus that being capable of tackling both types of tasks is essential to achieve true (artificial) intelligence.
At this aim I will present DeepProbLog, a neural probabilistic logic programming language that incorporates deep learning by means of neural predicates.
I will show how existing inference and learning techniques of the underlying probabilistic logic programming language ProbLog can be adapted for the new language. To demonstrate that DeepProbLog supports both symbolic and subsymbolic representations and inference I will present one of the experiments performed by the authors. This work proposes a framework where general-purpose neural networks and expressive probabilistic-logical modeling and reasoning are integrated in a way that exploits the full expressiveness and strengths of both worlds and can be trained end-to-end based on examples.
In conclusion I will show a possible, real application for this model.
From the paper: Robin Manhaeve, Sebasijan Dumancic, Angelika Kimmig, Thomas Demeester, Luc De Raedt. “Neural Probabilistic Logic Programming in DeepProbLog”. (2021) – Artificial Intelligence, volume 298 – Elsevier