Recently, the Successor Features and Generalized Policy Improvement (SF&GPI) framework has been proposed as a method for learning, composing, and transferring predictive knowledge and behavior.
Many important tasks are defined in terms of object.
SFL leverages the ability of successor features (SF) to capture transition dynamics, using it to drive exploration by estimating state-novelty and to enable high-level planning by abstracting the state-space as a non-parametric landmark-based graph.
In this work, we show that one can learn object-interaction tasks from scratch without supervision by learning an attentive object-model as an auxiliary task during task learning with an object-centric relational RL agent.
While machine learning (ML) models are being increasingly trusted to make decisions in different and varying areas, the safety of systems using such models has become an increasing concern.