Code, data, models for the Sherlock corpus

This repo has the code for the paper "Question Answering over Electronic Devices: A New Benchmark Dataset and a Multi-Task Learning based QA Framework" accepted at EMNLP 2021 Findings.