Contrastive Representations for Temporal Reasoning

Benjamin Eysenbach (Princeton) · Michał Bortkiewicz (Warsaw University of Technology) · Alicja Ziarko (Princeton University, University of Warsaw) · Michał Zawalski (NVIDIA) · Piotr Miłoś (Ideas NCBR, Polish Academy of Sciences)
action sequencescontrastive representations for temporal reasoningexternal search algorithmgeneralizationlearned representationsnegative samplingpuzzle-solvingrepresentation learningrubik’s cubesearch stepssokobanspurious featuresstate-based representationstemporal contrastive learningtemporal reasoning

In classical AI, perception relies on learning state-based representations, while planning