Teaching Large Language Models To Perform Complex Reasoning

Large language models have become the main backbone of most state-of-the-art NLP systems. By pre-training on very large datasets with unsupervised objectives, these models are able to learn good representations for language composition, then transfer these over to a...

Extending Large Language Models Through Querying Symbolic Systems

Large language models have become the main backbone of most state-of-the-art NLP systems. By pre-training on very large datasets with unsupervised objectives, these models are able to learn good representations for language composition, then transfer these over to a...

Reasoning about Stochastic Games of Imperfect Information

In many games the outcome of the players’ actions is given stochastically rather than deterministically, e.g., in card games, board games with dice (Risk!), etc.However, the literature of logic-based languages for strategic reasoning has so far fallen short of...

Multi-Task Reinforcement Learning with Imagination-based Agents

Deep Reinforcement Learning (DRL) has proved to be a powerful technique that allows autonomous agents to learn optimal behaviours (aka policies) in unknown and complex environments through models of rewards and penalizations [1].By extending DRL with formal...

Common Sense Planning (for Robotics)

Task Planning (also known as Symbolic Planning or AI Planning) has proved to be a very useful technique to tackle the decision-making problem in robotics. Given a set of task goals, the planner can come up with a set of actions that will reach those goals once...

Neurosymbolic approaches to causal representation learning

Causal reasoning is essential to decision-making in real-world problems. However, observational data is rarely sufficient to infer causal relationships or estimate treatment effects due to confounding signals. Pearl (2009) proposes a sound and complete formal system...

Fast Reinforcement Learning using Memory-Augmented Neural Networks

Reinforcement learning resembles human learning with intelligence accumulated through experiment. To attain expert human-level performance on tasks such as Atari video games or chess, deep RL systems have required many orders of magnitude more training data than human...