AbstractData Augmentation (DA) — generating extra training samples beyond the original set has been widely-used in today’s unbiased VQA models to mitigate language biases. Current mainstream DA strategies are synthetic-based methods, which synthesize new by either editing some visual regions/words, or re-generating them from scratch. However, these synthetic always unnatural and error-prone. To...

FrBMedQA is the first French biomedical question answering dataset, containing 41k+ passage-question instances. It was automatically constructed in a cloze-style manner, from Wikipedia articles. To test validity and difficulty of we experimented with four statistical baseline models, bidirectional encoder representation transformers (BERT)-based model, two BERT-based language model. We also did...

Incorporating external knowledge to Visual Question Answering (VQA) has become a vital practical need. Existing methods mostly adopt pipeline approaches with different components for matching and extraction, feature learning, etc. However, such suffer when some component does not perform well, which leads error cascading poor overall performance. Furthermore, the majority of existing ignore ans...

"Question answering is the task of predicting answers for questions based on a context paragraph. It has become especially important, as large amounts textual data available online requires not only gathering information but also findings specific to questions. In this work, we present experiments evaluated XQuAD-ro question dataset that been recently published translation SQuAD into Romanian. ...