Grounded Object Centric Learning
About
The extraction of modular object-centric representations for downstream tasks is an emerging area of research. Learning grounded representations of objects that are guaranteed to be stable and invariant promises robust performance across different tasks and environments. Slot Attention (SA) learns object-centric representations by assigning objects to \textit{slots}, but presupposes a \textit{single} distribution from which all slots are randomly initialised. This results in an inability to learn \textit{specialized} slots which bind to specific object types and remain invariant to identity-preserving changes in object appearance. To address this, we present \emph{\textsc{Co}nditional \textsc{S}lot \textsc{A}ttention} (\textsc{CoSA}) using a novel concept of \emph{Grounded Slot Dictionary} (GSD) inspired by vector quantization. Our proposed GSD comprises (i) canonical object-level property vectors and (ii) parametric Gaussian distributions, which define a prior over the slots. We demonstrate the benefits of our method in multiple downstream tasks such as scene generation, composition, and task adaptation, whilst remaining competitive with SA in popular object discovery benchmarks.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Image Classification | MM-A out-of-distribution (OOD) | Task Accuracy36.2 | 6 | |
| Classification | PokerRules standard (test) | Task Accuracy95.46 | 6 | |
| Image Classification | MM-A in-distribution (test) | Accuracy93 | 6 | |
| Classification | PokerRules Extrapolation: 5 cards (In-distribution class) | Task Accuracy44.26 | 5 | |
| Image Classification | MM-A Extrapolation 4 digits | Task Accuracy52.2 | 5 | |
| Image Classification | MM-A Extrapolation 5 digits | Task Accuracy25.06 | 5 | |
| Addition | CLEVR-Addition 7 objects (extrapolation) | Task Accuracy3.44 | 3 | |
| Addition | CLEVR-Addition (test) | Task Accuracy88.79 | 3 |