LEAF-QA: Locate, Encode & Attend for Figure Question Answering
This addresses the problem of multimodal QA for researchers and practitioners in AI, focusing on chart understanding, but is incremental as it builds on prior datasets like FigureQA and DVQA.
The authors tackled multimodal question answering on complex real-world charts by introducing LEAF-QA, a dataset of 250,000 annotated figures with 2 million QA pairs, and LEAF-Net, an architecture that advances state-of-the-art on existing benchmarks.
We introduce LEAF-QA, a comprehensive dataset of $250,000$ densely annotated figures/charts, constructed from real-world open data sources, along with ~2 million question-answer (QA) pairs querying the structure and semantics of these charts. LEAF-QA highlights the problem of multimodal QA, which is notably different from conventional visual QA (VQA), and has recently gained interest in the community. Furthermore, LEAF-QA is significantly more complex than previous attempts at chart QA, viz. FigureQA and DVQA, which present only limited variations in chart data. LEAF-QA being constructed from real-world sources, requires a novel architecture to enable question answering. To this end, LEAF-Net, a deep architecture involving chart element localization, question and answer encoding in terms of chart elements, and an attention network is proposed. Different experiments are conducted to demonstrate the challenges of QA on LEAF-QA. The proposed architecture, LEAF-Net also considerably advances the current state-of-the-art on FigureQA and DVQA.