Statistical and structural identifiability in representation learning

• URL: http://arxiv.org/abs/2603.11970v1
• Date: Thu, 12 Mar 2026 14:19:58 GMT
• Title: Statistical and structural identifiability in representation learning
• Abstract summary: We formalize two distinct concepts: statistical identifiability and structural identifiability.<n>We prove a statistical $$-near-identifiability result for the representations of models with nonlinear decoders.<n>With additional assumptions on the data-generating process, statistical identifiability extends to structural identifiability.
• Score: 31.990445585569688
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Representation learning models exhibit a surprising stability in their internal representations. Whereas most prior work treats this stability as a single property, we formalize it as two distinct concepts: statistical identifiability (consistency of representations across runs) and structural identifiability (alignment of representations with some unobserved ground truth). Recognizing that perfect pointwise identifiability is generally unrealistic for modern representation learning models, we propose new model-agnostic definitions of statistical and structural near-identifiability of representations up to some error tolerance $ε$. Leveraging these definitions, we prove a statistical $ε$-near-identifiability result for the representations of models with nonlinear decoders, generalizing existing identifiability theory beyond last-layer representations in e.g. generative pre-trained transformers (GPTs) to near-identifiability of the intermediate representations of a broad class of models including (masked) autoencoders (MAEs) and supervised learners. Although these weaker assumptions confer weaker identifiability, we show that independent components analysis (ICA) can resolve much of the remaining linear ambiguity for this class of models, and validate and measure our near-identifiability claims empirically. With additional assumptions on the data-generating process, statistical identifiability extends to structural identifiability, yielding a simple and practical recipe for disentanglement: ICA post-processing of latent representations. On synthetic benchmarks, this approach achieves state-of-the-art disentanglement using a vanilla autoencoder. With a foundation model-scale MAE for cell microscopy, it disentangles biological variation from technical batch effects, substantially improving downstream generalization.

Related papers

• Disentangled representations via score-based variational autoencoders [21.955536401578616]
  We present the Score-based Autoencoder for Multiscale Inference (SAMI)<n>SAMI formulates a principled objective that learns representations through score-based guidance of the underlying diffusion process.<n>It can extract useful representations from pre-trained diffusion models with minimal additional training.
  arXiv  Detail & Related papers  (2025-12-18T23:42:10Z)
• MASCOTS: Model-Agnostic Symbolic COunterfactual explanations for Time Series [4.664512594743523]
  We introduce MASCOTS, a method that generates meaningful and diverse counterfactual observations in a model-agnostic manner.<n>By operating in a symbolic feature space, MASCOTS enhances interpretability while preserving fidelity to the original data and model.
  arXiv  Detail & Related papers  (2025-03-28T12:48:12Z)
• The Foundations of Tokenization: Statistical and Computational Concerns [51.370165245628975]
  Tokenization is a critical step in the NLP pipeline.<n>Despite its recognized importance as a standard representation method in NLP, the theoretical underpinnings of tokenization are not yet fully understood.<n>The present paper contributes to addressing this theoretical gap by proposing a unified formal framework for representing and analyzing tokenizer models.
  arXiv  Detail & Related papers  (2024-07-16T11:12:28Z)
• Identifiable Exchangeable Mechanisms for Causal Structure and Representation Learning [54.69189620971405]
  We provide a unified framework, termed Identifiable Exchangeable Mechanisms (IEM), for representation and structure learning.<n>IEM provides new insights that let us relax the necessary conditions for causal structure identification in exchangeable non-i.i.d. data.<n>We also demonstrate the existence of a duality condition in identifiable representation learning, leading to new identifiability results.
  arXiv  Detail & Related papers  (2024-06-20T13:30:25Z)
• Beyond DAGs: A Latent Partial Causal Model for Multimodal Learning [80.44084021062105]
  We propose a novel latent partial causal model for multimodal data, featuring two latent coupled variables, connected by an undirected edge, to represent the transfer of knowledge across modalities.<n>Under specific statistical assumptions, we establish an identifiability result, demonstrating that representations learned by multimodal contrastive learning correspond to the latent coupled variables up to a trivial transformation.<n>Experiments on a pre-trained CLIP model embodies disentangled representations, enabling few-shot learning and improving domain generalization across diverse real-world datasets.
  arXiv  Detail & Related papers  (2024-02-09T07:18:06Z)
• Flow Factorized Representation Learning [109.51947536586677]
  We introduce a generative model which specifies a distinct set of latent probability paths that define different input transformations. We show that our model achieves higher likelihoods on standard representation learning benchmarks while simultaneously being closer to approximately equivariant models.
  arXiv  Detail & Related papers  (2023-09-22T20:15:37Z)
• Leveraging Task Structures for Improved Identifiability in Neural Network Representations [31.863998589693065]
  We extend the theory of identifiability in supervised learning by considering the consequences of having access to a distribution of tasks. We show that linear identifiability is achievable in the general multi-task regression setting.
  arXiv  Detail & Related papers  (2023-06-26T17:16:50Z)
• Indeterminacy in Latent Variable Models: Characterization and Strong Identifiability [3.959606869996233]
  We construct a theoretical framework for analyzing the indeterminacies of latent variable models. We then investigate how we might specify strongly identifiable latent variable models.
  arXiv  Detail & Related papers  (2022-06-02T00:01:27Z)
• Improving the Reconstruction of Disentangled Representation Learners via Multi-Stage Modeling [54.94763543386523]
  Current autoencoder-based disentangled representation learning methods achieve disentanglement by penalizing the ( aggregate) posterior to encourage statistical independence of the latent factors. We present a novel multi-stage modeling approach where the disentangled factors are first learned using a penalty-based disentangled representation learning method. Then, the low-quality reconstruction is improved with another deep generative model that is trained to model the missing correlated latent variables.
  arXiv  Detail & Related papers  (2020-10-25T18:51:15Z)
• Structural Causal Models Are (Solvable by) Credal Networks [70.45873402967297]
  Causal inferences can be obtained by standard algorithms for the updating of credal nets. This contribution should be regarded as a systematic approach to represent structural causal models by credal networks. Experiments show that approximate algorithms for credal networks can immediately be used to do causal inference in real-size problems.
  arXiv  Detail & Related papers  (2020-08-02T11:19:36Z)
• On Linear Identifiability of Learned Representations [26.311880922890843]
  We study identifiability in the context of representation learning. We show that a large family of discriminative models are identifiable in function space, up to a linear indeterminacy. We derive sufficient conditions for linear identifiability and provide empirical support for the result on both simulated and real-world data.
  arXiv  Detail & Related papers  (2020-07-01T23:33:37Z)
• ICE-BeeM: Identifiable Conditional Energy-Based Deep Models Based on Nonlinear ICA [11.919315372249802]
  We consider the identifiability theory of probabilistic models. We show that our model can be used for the estimation of the components in the framework of Independently Modulated Component Analysis.
  arXiv  Detail & Related papers  (2020-02-26T14:43:30Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.