Skip to yearly menu bar Skip to main content


Probing Emergent Semantics in Predictive Agents via Question Answering

Abhishek Das · Federico Carnevale · Hamza Merzic · Laura Rimell · Rosalia Schneider · Josh Abramson · Alden Hung · Arun Ahuja · Stephen Clark · Greg Wayne · Feilx Hill


Keywords: [ Deep Reinforcement Learning ] [ Neuroscience and Cognitive Science ] [ Representation Learning ] [ Applications - Neuroscience, Cognitive Science, Biology and Health ]


Recent work has shown how predictive modeling can endow agents with rich knowledge of their surroundings, improving their ability to act in complex environments. We propose question-answering as a general paradigm to decode and understand the representations that such agents develop, applying our method to two recent approaches to predictive modelling - action-conditional CPC (Guo et al., 2018) and SimCore (Gregor et al., 2019). After training agents with these predictive objectives in a visually-rich, 3D environment with an assortment of objects, colors, shapes, and spatial configurations, we probe their internal state representations with a host of synthetic (English) questions, without backpropagating gradients from the question-answering decoder into the agent. The performance of different agents when probed in this way reveals that they learn to encode factual, and seemingly compositional, information about objects, properties and spatial relations from their physical environment. Our approach is intuitive, i.e. humans can easily interpret the responses of the model as opposed to inspecting continuous vectors, and model-agnostic, i.e. applicable to any modeling approach. By revealing the implicit knowledge of objects, quantities, properties and relations acquired by agents as they learn, question-conditional agent probing can stimulate the design and development of stronger predictive learning objectives.

Chat is not available.