method: HyperDQA_V02020-04-30

Authors: Anisha Gunjal, Vipul Gupta, Moinak Bhattacharya, Digvijay Singh

Affiliation: HyperVerge, Inc.

Description: Our method uses a deep learning based approach to extract text and positional embeddings from the image to locate the answer regions for the question. The model architecture is fine-tuned from a transformer based architecture.