# Vector-based Representation is the Key: A Study on Disentanglement and Compositional Generalization

arXiv: https://arxiv.org/abs/2305.18063 · HTML: https://mlsanity.com/paper/2305.18063/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 13 of its 26 references that are in the map.
- First posted on arXiv 29 May 2023.

## Builds on (most cited first, 13 of 13)

- [A Style-Based Generator Architecture for Generative Adversarial Networks](https://mlsanity.com/paper/1812.04948/) (2018)
- [Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations](https://mlsanity.com/paper/1811.12359/) (2018)
- [Disentangling by Factorising](https://mlsanity.com/paper/1802.05983/) (2018)
- [On the Transfer of Inductive Bias from Simulation to the Real World: a New Disentanglement Dataset](https://mlsanity.com/paper/1906.03292/) (2019)
- [Bias and Generalization in Deep Generative Models: An Empirical Study](https://mlsanity.com/paper/1811.03259/) (2018)
- [Unsupervised Learning of Compositional Energy Concepts](https://mlsanity.com/paper/2111.03042/) (2021)
- [Structured Disentangled Representations](https://mlsanity.com/paper/1804.02086/) (2018)
- [Visual Representation Learning Does Not Generalize Strongly Within the Same Domain](https://mlsanity.com/paper/2107.08221/) (2021)
- [Compositional Generalization in Unsupervised Compositional Representation Learning: A Study on Disentanglement and Emergent Language](https://mlsanity.com/paper/2210.00482/) (2022)
- [Towards Building A Group-based Unsupervised Representation Disentanglement Framework](https://mlsanity.com/paper/2102.10303/) (2021)
- [Compositional generalization through abstract representations in human and artificial neural networks](https://mlsanity.com/paper/2209.07431/) (2022)
- [Visual Concepts Tokenization](https://mlsanity.com/paper/2205.10093/) (2022)
- [Structure by Architecture: Structured Representations without Regularization](https://mlsanity.com/paper/2006.07796/) (2020)

## Built on by (most cited first, 6 of 6)

- [Diffusion Model with Cross Attention as an Inductive Bias for Disentanglement](https://mlsanity.com/paper/2402.09712/) (2024)
- [Deciphering the Role of Representation Disentanglement: Investigating Compositional Generalization in CLIP Models](https://mlsanity.com/paper/2407.05897/) (2024)
- [Graph-based Unsupervised Disentangled Representation Learning via Multimodal Large Language Models](https://mlsanity.com/paper/2407.18999/) (2024)
- [Closed-Loop Unsupervised Representation Disentanglement with β-VAE Distillation and Diffusion Probabilistic Feedback](https://mlsanity.com/paper/2402.02346/) (2024)
- [Compositional Generalization via Forced Rendering of Disentangled Latents](https://mlsanity.com/paper/2501.18797/) (2025)
- [Rule Extrapolation in Language Models: A Study of Compositional Generalization on OOD Prompts](https://mlsanity.com/paper/2409.13728/) (2024)

## Abstract (arXiv)

Recognizing elementary underlying concepts from observations (disentanglement) and generating novel combinations of these concepts (compositional generalization) are fundamental abilities for humans to support rapid knowledge learning and generalize to new tasks, with which the deep learning models struggle. Towards human-like intelligence, various works on disentangled representation learning have been proposed, and recently some studies on compositional generalization have been presented. However, few works study the relationship between disentanglement and compositional generalization, and the observed results are inconsistent. In this paper, we study several typical disentangled representation learning works in terms of both disentanglement and compositional generalization abilities, and we provide an important insight: vector-based representation (using a vector instead of a scalar to represent a concept) is the key to empower both good disentanglement and strong compositional generalization. This insight also resonates the neuroscience research that the brain encodes information in neuron population activity rather than individual neurons. Motivated by this observation, we further propose a method to reform the scalar-based disentanglement works (β-TCVAE and FactorVAE) to be vector-based to increase both capabilities. We investigate the impact of the dimensions of vector-based representation and one important question: whether better disentanglement indicates higher compositional generalization. In summary, our study demonstrates that it is possible to achieve both good concept recognition and novel concept composition, contributing an important step towards human-like intelligence.