On Linear Identifiability of Learned Representations
This work addresses a foundational issue in machine learning by providing theoretical guarantees for representation identifiability, which is incremental as it builds on recent nonlinear ICA advances.
The paper tackles the problem of identifiability in representation learning, showing that many discriminative models are identifiable in function space up to a linear indeterminacy, with empirical support on simulated and real-world data.
Identifiability is a desirable property of a statistical model: it implies that the true model parameters may be estimated to any desired precision, given sufficient computational resources and data. We study identifiability in the context of representation learning: discovering nonlinear data representations that are optimal with respect to some downstream task. When parameterized as deep neural networks, such representation functions typically lack identifiability in parameter space, because they are overparameterized by design. In this paper, building on recent advances in nonlinear ICA, we aim to rehabilitate identifiability by showing that a large family of discriminative models are in fact identifiable in function space, up to a linear indeterminacy. Many models for representation learning in a wide variety of domains have been identifiable in this sense, including text, images and audio, state-of-the-art at time of publication. We derive sufficient conditions for linear identifiability and provide empirical support for the result on both simulated and real-world data.