Poster
Provable Compositional Generalization for Object-Centric Learning
Thaddäus Wiedemer · Jack Brady · Alexander Panfilov · Attila Juhos · Matthias Bethge · Wieland Brendel
Halle B #155
Fri 10 May 6:45 a.m. PDT — 7:30 a.m. PDT
Learning representations that generalize to novel compositions of known concepts is crucial for bridging the gap between human and machine perception. One prominent effort is learning object-centric representations, which are widely conjectured to enable compositional generalization. Yet, it remains unclear when this conjecture will be true, as a principled theoretical or empirical understanding of compositional generalization is lacking. In this work, we investigate when compositional generalization is guaranteed for object-centric representations through the lens of identifiability theory. We show that autoencoders that satisfy structural assumptions on the decoder and enforce encoder-decoder consistency will learn object-centric representations that provably generalize compositionally. We validate our theoretical result and highlight the practical relevance of our assumptions through experiments on synthetic image data.