The SelectGen Challenge: Finding the Best Training Samples for Few-Shot Neural Text Generation

Abstract

We propose a shared task on training instance selection for few-shot neural text generation. Large-scale pretrained language models have led to dramatic improvements in few-shot text generation. Nonetheless, almost all previous work simply applies random sampling to select the few-shot training instances. Little to no attention has been paid to the selection strategies and how they would affect model performance.The study of the selection strategy can help us to (1) make the most use of our annotation budget in downstream tasks and (2) better benchmark few-shot text generative models. We welcome submissions that present their selection strategies and the effects on the generation quality.

Cross-lingual Sentiment Analysis via AAE and BiGRU

“…Repeating this process for all languages is infeasible, as most low-resource languages do not have enough conversational data to support this type of training (Zhao et al, 2020;. Even for high-resource languages, collecting sufficient amount of high-quality data to cover various domains is still costly (Xu et al, 2020;Chang et al, 2021). Therefore, we believe crosslingual transfer is crucial for efficiently developing chatbots in multiple languages, through which the same resource can be reused across languages.…”