Trick Me If You Can: Human-in-the-loop Generation of Adversarial Examples for Question Answering
This addresses the challenge of robust question answering for AI systems by exposing open issues like multi-hop reasoning, though it is incremental as it builds on existing adversarial evaluation methods.
The paper tackled the problem of generating complex and diverse adversarial examples for question answering by proposing a human-in-the-loop approach, where trivia enthusiasts crafted adversarial questions that systematically stumped neural and information retrieval models in live matches.
Adversarial evaluation stress tests a model's understanding of natural language. While past approaches expose superficial patterns, the resulting adversarial examples are limited in complexity and diversity. We propose human-in-the-loop adversarial generation, where human authors are guided to break models. We aid the authors with interpretations of model predictions through an interactive user interface. We apply this generation framework to a question answering task called Quizbowl, where trivia enthusiasts craft adversarial questions. The resulting questions are validated via live human--computer matches: although the questions appear ordinary to humans, they systematically stump neural and information retrieval models. The adversarial questions cover diverse phenomena from multi-hop reasoning to entity type distractors, exposing open challenges in robust question answering.