Explaining Question Answering Models through Text Generation

Abstract

Large pre-trained language models (LMs) have been shown to perform surprisingly well when fine-tuned on tasks that require commonsense and world knowledge. However, in end-to-end architectures, it is difficult to explain what is the knowledge in the LM that allows it to make a correct prediction. In this work, we propose a model for multi-choice question answering, where a LM-based generator generates a textual hypothesis that is later used by a classifier to answer the question. The hypothesis provides a window into the information used by the fine-tuned LM that can be inspected by humans. A key challenge in this setup is how to constrain the model to generate hypotheses that are meaningful to humans. We tackle this by (a) joint training with a simple similarity classifier that encourages meaningful hypotheses, and (b) by adding loss functions that encourage natural text without repetitions. We show on several tasks that our model reaches performance that is comparable to end-to-end architectures, while producing hypotheses that elucidate the knowledge used by the LM for answering the question.

References in corpus (22)

• Sequence to Sequence Learning with Neural Networks
• A Unified Approach to Interpreting Model Predictions
• Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
• Categorical Reparameterization with Gumbel-Softmax
• HuggingFace's Transformers: State-of-the-art Natural Language Processing
• Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation
• XLNet: Generalized Autoregressive Pretraining for Language Understanding
• The Curious Case of Neural Text Degeneration
• ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
• CTRL: A Conditional Transformer Language Model for Controllable Generation
• Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
• The Natural Language Decathlon: Multitask Learning as Question Answering
• CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge
• A Simple Method for Commonsense Reasoning
• Neural Text Generation with Unlikelihood Training
• Plug and Play Language Models: A Simple Approach to Controlled Text Generation
• Performance Impact Caused by Hidden Bias of Training Data for Recognizing Textual Entailment
• Language Models as Knowledge Bases?
• RACE: Large-scale ReAding Comprehension Dataset From Examinations
• Understanding Straight-Through Estimator in Training Activation Quantized Neural Nets
• Controlling Output Length in Neural Encoder-Decoders
• Sequence Tutor: Conservative Fine-Tuning of Sequence Generation Models with KL-control

Cited by in corpus (7)

• Interactive Question Answering Systems: Literature Review
• A Survey on Explainability in Machine Reading Comprehension
• Are Training Resources Insufficient? Predict First Then Explain!
• You Can Do Better! If You Elaborate the Reason When Making Prediction
• Denoising Large-Scale Image Captioning from Alt-text Data using Content Selection Models
• Grounding 'Grounding' in NLP
• Positioning yourself in the maze of Neural Text Generation: A Task-Agnostic Survey