Technology Category
Machine Learning: Representation LearningMultiagent Systems: Multiagent LearningSearch and Optimization: Learning to Search
Disentangling Representations through Multi-task Learning
🤖 AI Summary
This study investigates how multi-task learning drives agents to spontaneously develop disentangled representations—orthogonal, generalizable internal coordinate systems that separate latent factors of the world. Methodologically, it integrates recurrent neural networks (RNNs), which implement continuous attractor dynamics for disentanglement, with Transformer architectures, complemented by latent-variable decoding and an out-of-distribution (OOD) zero-shot generalization evaluation framework. Theoretically and empirically, it establishes for the first time that optimal multi-task evidence accumulation implicitly induces disentanglement; it further identifies critical conditions—governing noise level, task cardinality, and decision time—under which disentanglement emerges. Results show that RNNs achieve zero-shot OOD prediction of latent factors, while Transformers exhibit superior disentanglement, deeper world modeling, and enhanced conceptual interpretability, providing a novel mechanistic account and empirical foundation for feature-based generalization.
Technology Category
Application Category
📝 Abstract
Intelligent perception and interaction with the world hinges on internal representations that capture its underlying structure (''disentangled'' or ''abstract'' representations). Disentangled representations serve as world models, isolating latent factors of variation in the world along approximately orthogonal directions, thus facilitating feature-based generalization. We provide experimental and theoretical results guaranteeing the emergence of disentangled representations in agents that optimally solve multi-task evidence accumulation classification tasks, canonical in the neuroscience literature. The key conceptual finding is that, by producing accurate multi-task classification estimates, a system implicitly represents a set of coordinates specifying a disentangled representation of the underlying latent state of the data it receives. The theory provides conditions for the emergence of these representations in terms of noise, number of tasks, and evidence accumulation time. We experimentally validate these predictions in RNNs trained to multi-task, which learn disentangled representations in the form of continuous attractors, leading to zero-shot out-of-distribution (OOD) generalization in predicting latent factors. We demonstrate the robustness of our framework across autoregressive architectures, decision boundary geometries and in tasks requiring classification confidence estimation. We find that transformers are particularly suited for disentangling representations, which might explain their unique world understanding abilities. Overall, our framework establishes a formal link between competence at multiple tasks and the formation of disentangled, interpretable world models in both biological and artificial systems, and helps explain why ANNs often arrive at human-interpretable concepts, and how they both may acquire exceptional zero-shot generalization capabilities.
Problem
Research questions and friction points this paper is trying to address.
Explores how multi-task learning enables disentangled representations in AI systems.
Investigates conditions for disentangled representations in terms of noise and task complexity.
Demonstrates zero-shot generalization in AI models using disentangled representations.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Multi-task learning for disentangled representations
RNNs and transformers for zero-shot generalization
Theoretical conditions for disentangled representation emergence
🔎 Similar Papers
No similar papers found.
