QuesBELM: A BERT based Ensemble Language Model for Natural Questions

A core goal in artificial intelligence is to build systems that can read the web, and then answer complex questions related to random searches about any topic. These question-answering (QA) systems could have a big impact on the way that we access information. In this paper, we addressed the task of question-answering (QA) systems on Google's Natural Questions (NQ) dataset containing real user questions issued to Google search and the answers found from Wikipedia by annotators. In our work, we systematically compare the performance of powerful variant models of Transformer architectures-'BERTbase, BERT-large-WWM and ALBERT-XXL' over Natural Questions dataset. We also propose a state-of-the-art BERT based ensemble language model-QuesBELM. QuesBELM leverages the power of existing BERT variants combined together to build a more accurate stacking ensemble model for question answering (QA) system. The model integrates top-K predictions from single language models to determine the best answer out of all. Our model surpassed the baseline language models with the Harmonic mean score of 0.731 and 0.582 for the long answer(LA) and short answer(SA) tasks respectively, reporting an average of 10% improvement over the baseline models.