Probabilistic Neural-symbolic Models for Interpretable Visual Question Answering

TL;DR

This paper introduces probabilistic neural-symbolic models for visual question answering that produce interpretable programs, require fewer examples, and enable counterfactual reasoning to test model beliefs.

Contribution

It presents a novel probabilistic framework for neural-symbolic VQA models with interpretable programs and counterfactual reasoning capabilities, improving data efficiency and interpretability.

Findings

Better program and answer accuracy in low-data regimes

Enhanced interpretability of generated programs

Ability to probe model beliefs through counterfactual scenarios

Abstract

We propose a new class of probabilistic neural-symbolic models, that have symbolic functional programs as a latent, stochastic variable. Instantiated in the context of visual question answering, our probabilistic formulation offers two key conceptual advantages over prior neural-symbolic models for VQA. Firstly, the programs generated by our model are more understandable while requiring lesser number of teaching examples. Secondly, we show that one can pose counterfactual scenarios to the model, to probe its beliefs on the programs that could lead to a specified answer given an image. Our results on the CLEVR and SHAPES datasets verify our hypotheses, showing that the model gets better program (and answer) prediction accuracy even in the low data regime, and allows one to probe the coherence and consistency of reasoning performed.