International Workshop on Continual Semi-Supervised Learning: Introduction, Benchmarks and Baselines

TL;DR

This paper formalizes the continual semi-supervised learning paradigm, introduces new benchmarks for activity recognition and crowd counting, and evaluates baseline models, highlighting the challenges of learning from unlabelled data streams.

Contribution

It defines CSSL, creates benchmarks and challenges for vision tasks, and proposes simple baseline methods to stimulate further research.

Findings

Learning from unlabelled data streams is highly challenging.

Baseline models show limited performance, indicating room for improvement.

The benchmarks facilitate standardized evaluation of CSSL methods.

Abstract

The aim of this paper is to formalize a new continual semi-supervised learning (CSSL) paradigm, proposed to the attention of the machine learning community via the IJCAI 2021 International Workshop on Continual Semi-Supervised Learning (CSSL-IJCAI), with the aim of raising field awareness about this problem and mobilizing its effort in this direction. After a formal definition of continual semi-supervised learning and the appropriate training and testing protocols, the paper introduces two new benchmarks specifically designed to assess CSSL on two important computer vision tasks: activity recognition and crowd counting. We describe the Continual Activity Recognition (CAR) and Continual Crowd Counting (CCC) challenges built upon those benchmarks, the baseline models proposed for the challenges, and describe a simple CSSL baseline which consists in applying batch self-training in temporal sessions, for a limited number of rounds. The results show that learning from unlabelled data streams is extremely challenging, and stimulate the search for methods that can encode the dynamics of the data stream.