Question Answering with Texts and Tables through Deep Reinforcement Learning
This addresses the challenge of selecting models in sequential information retrieval for question answering, though it appears incremental as it builds on existing tools and benchmarks.
The paper tackled the problem of generating multi-hop answers to open domain questions requiring information from both texts and tables by proposing a novel architecture that uses reinforcement learning to select between state-of-the-art tools sequentially, achieving an F1-score of 19.03, which is comparable to existing iterative systems.
This paper proposes a novel architecture to generate multi-hop answers to open domain questions that require information from texts and tables, using the Open Table-and-Text Question Answering dataset for validation and training. One of the most common ways to generate answers in this setting is to retrieve information sequentially, where a selected piece of data helps searching for the next piece. As different models can have distinct behaviors when called in this sequential information search, a challenge is how to select models at each step. Our architecture employs reinforcement learning to choose between different state-of-the-art tools sequentially until, in the end, a desired answer is generated. This system achieved an F1-score of 19.03, comparable to iterative systems in the literature.