Learning Emotion from 100 Observations: Unexpected Robustness of Deep Learning under Strong Data Limitations

TL;DR

Despite common beliefs, deep learning models can effectively learn emotion recognition with only 100 observations, especially when using high-quality pre-trained embeddings, challenging the notion that large data is always necessary.

Contribution

The paper demonstrates that neural models can outperform traditional methods with minimal data in emotion analysis, highlighting the importance of pre-trained embeddings.

Findings

Neural models outperform $n$-gram ridge regression with just 100 data points.

Pre-trained embeddings are crucial for effective learning with limited data.

Deep learning shows unexpected robustness under strong data limitations.

Abstract

One of the major downsides of Deep Learning is its supposed need for vast amounts of training data. As such, these techniques appear ill-suited for NLP areas where annotated data is limited, such as less-resourced languages or emotion analysis, with its many nuanced and hard-to-acquire annotation formats. We conduct a questionnaire study indicating that indeed the vast majority of researchers in emotion analysis deems neural models inferior to traditional machine learning when training data is limited. In stark contrast to those survey results, we provide empirical evidence for English, Polish, and Portuguese that commonly used neural architectures can be trained on surprisingly few observations, outperforming $n$-gram based ridge regression on only 100 data points. Our analysis suggests that high-quality, pre-trained word embeddings are a main factor for achieving those results.