Converge to Surprise: Evolutionary Self-supervised Image Clustering
This work addresses the problem of self-supervised image clustering without requiring predefined targets or class numbers, offering a new paradigm for unsupervised learning.
The paper proposes a self-supervised image clustering framework that eliminates the need for a per-step loss target by using a 'surprise score' based on the Principle of Maximum Entropy, optimized via an evolution-strategy outer loop and gradient-descent inner loop. It achieves state-of-the-art results in non-parametric self-supervised image clustering on standard benchmarks.
Most self-supervised image clustering models, actually almost all deep learning approaches, are based on gradient descent: In order to calculate the loss, every optimization step requires a clearly defined target, whether a contrastive split, a masked patch or entity, an EMA-teacher output, a pseudo-label, or a differentiable information-theoretic functional. We propose a self-supervised framework that drops this requirement for image clustering. Without any prior knowledge, we have to assume that each pixel is i.i.d. according to the Principle of Maximum Entropy. Taking this as our null hypothesis H0, we define a "surprise score" that measures how unlikely the model's output representation would be under H0. Maximizing the surprise score forces the deep learning model to reject H0 - equivalently, to discover non-random feature from data. Also, here is our fundamental assumption: a surprise score cannot, in general, be reduced to a per-step loss. Hence, we propose the "converge-to-surprise" scheme to optimize our model: an evolution-strategy (ES) outer loop, which directly maximizes the surprise score without needing its gradient, paired with a periodic gradient-descent inner loop, which uses the surprising clusters already discovered by ES as surrogate targets. On standard image benchmarks, our framework achieves new state-of-the-art results in non-parametric self-supervised image clustering - the strictest deep-clustering setting, in which the number of ground-truth classes is not given to the model.