Generating Accurate Pseudo-Labels in Semi-Supervised Learning and Avoiding Overconfident Predictions via Hermite Polynomial Activations

By Vishnu Suresh Lokhande, Songwong Tasneeyapant, Abhay Venkatesh, Sathya Ravi, Vikas Singh. 2019

30 citations recorded by the source corpus.

Rectified Linear Units (ReLUs) are among the most widely used activation function in a broad variety of tasks in vision. Recent theoretical results suggest that despite their excellent practical performance, in various cases, a substitution with basis expansions (e.g., polynomials) can yield significant benefits from both the optimization and generalization perspective. Unfortunately, the existing results remain limited to networks with a couple of layers, and the practical viability of these results is not yet known. Motivated by some of these results, we explore the use of Hermite polynomial expansions as a substitute for ReLUs in deep networks. While our experiments with supervised learning do not provide a clear verdict, we find that this strategy offers considerable benefits in semi-supervised learning (SSL) / transductive learning settings. We carefully develop this idea and show how the use of Hermite polynomials based activations can yield improvements in pseudo-label accuracies and sizable financial savings (due to concurrent runtime benefits). Further, we show via theoretical analysis, that the networks (with Hermite activations) offer robustness to noise and other attractive mathematical properties.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Detection and Continual Learning of Novel Face Presentation Attacks (2021, background): “Since deep networks suffer from overconfidence in their predictions [14], the system may not be able to identify novel attack types, generated at inference time after the initial training.”
Learning to Predict Gradients for Semi-Supervised Continual Learning (2022, methodology): “During the training process, at the i -th training step or episode, the generated logits z i w.r.t. the input x i is passed to the continual loss ℓ .”