Guiding Visual Question Generation

TL;DR

This paper introduces guided models for visual question generation that condition questions on specific objects and categories, improving relevance and coherence over existing methods.

Contribution

It proposes explicit and implicit guidance mechanisms for VQG, enabling better control and diversity in generated questions.

Findings

Over 9 BLEU-4 improvement over state-of-the-art

Guidance improves grammatical coherence and relevance

Models effectively condition on specified objects and categories

Abstract

In traditional Visual Question Generation (VQG), most images have multiple concepts (e.g. objects and categories) for which a question could be generated, but models are trained to mimic an arbitrary choice of concept as given in their training data. This makes training difficult and also poses issues for evaluation -- multiple valid questions exist for most images but only one or a few are captured by the human references. We present Guiding Visual Question Generation - a variant of VQG which conditions the question generator on categorical information based on expectations on the type of question and the objects it should explore. We propose two variants: (i) an explicitly guided model that enables an actor (human or automated) to select which objects and categories to generate a question for; and (ii) an implicitly guided model that learns which objects and categories to condition on, based on discrete latent variables. The proposed models are evaluated on an answer-category augmented VQA dataset and our quantitative results show a substantial improvement over the current state of the art (over 9 BLEU-4 increase). Human evaluation validates that guidance helps the generation of questions that are grammatically coherent and relevant to the given image and objects.