Challenge

In the afternoon we will kick off with a small challenge: participants will be divided into groups and asked to come up with hard examples that trick existing text classifiers. Links to models for multiple tasks and languages are available below: each team should choose one model/task and use the demo available on the HuggingFace platform. The challenge will work as a warm-up session for the following group activity.

Abuse Detection

• English: BERT
• Italian: BERT
• German: BERT

Sentiment Analysis

• English: RoBERTa
• Spanish: RoBERTa
• French: BERT