Instruction Makes a Difference
This work addresses document analysis for AI applications, but it is incremental as it builds on existing large language-vision models with a new dataset and fine-tuning approach.
The authors tackled the problem of document visual question answering (DocVQA) by introducing an instruction-following dataset and model, showing that instruction-tuning improves performance over zero-shot and traditional fine-tuning, with gains ranging from 11x to 32x and 0.1% to 4.2%, though still below human performance at 94.36%.
We introduce Instruction Document Visual Question Answering (iDocVQA) dataset and Large Language Document (LLaDoc) model, for training Language-Vision (LV) models for document analysis and predictions on document images, respectively. Usually, deep neural networks for the DocVQA task are trained on datasets lacking instructions. We show that using instruction-following datasets improves performance. We compare performance across document-related datasets using the recent state-of-the-art (SotA) Large Language and Vision Assistant (LLaVA)1.5 as the base model. We also evaluate the performance of the derived models for object hallucination using the Polling-based Object Probing Evaluation (POPE) dataset. The results show that instruction-tuning performance ranges from 11X to 32X of zero-shot performance and from 0.1% to 4.2% over non-instruction (traditional task) finetuning. Despite the gains, these still fall short of human performance (94.36%), implying there's much room for improvement.