# Learning without Prejudice: Avoiding Bias in Webly-Supervised Action Recognition

## TL;DR

This paper presents a novel webly-supervised approach for action recognition that trains independent CNNs on different data modalities to reduce bias and improve performance on benchmark datasets.

## Contribution

It introduces a method of training two independent CNNs on web data for action recognition, avoiding bias from data cleaning procedures, and demonstrates superior results on public benchmarks.

## Key findings

- Outperforms existing webly-supervised methods on UCF-101 and Thumos'14.
- Training independent CNNs on different data sources reduces bias.
- Enriching training data with heterogeneous web sources improves accuracy.

## Abstract

Webly-supervised learning has recently emerged as an alternative paradigm to traditional supervised learning based on large-scale datasets with manual annotations. The key idea is that models such as CNNs can be learned from the noisy visual data available on the web. In this work we aim to exploit web data for video understanding tasks such as action recognition and detection. One of the main problems in webly-supervised learning is cleaning the noisy labeled data from the web. The state-of-the-art paradigm relies on training a first classifier on noisy data that is then used to clean the remaining dataset. Our key insight is that this procedure biases the second classifier towards samples that the first one understands. Here we train two independent CNNs, a RGB network on web images and video frames and a second network using temporal information from optical flow. We show that training the networks independently is vastly superior to selecting the frames for the flow classifier by using our RGB network. Moreover, we show benefits in enriching the training set with different data sources from heterogeneous public web databases. We demonstrate that our framework outperforms all other webly-supervised methods on two public benchmarks, UCF-101 and Thumos'14.