Supervising the Transfer of Reasoning Patterns in VQA
arXiv:2106.05597
Abstract
Methods for Visual Question Anwering (VQA) are notorious for leveraging dataset biases rather than performing reasoning, hindering generalization. It has been recently shown that better reasoning patterns emerge in attention layers of a state-of-the-art VQA model when they are trained on perfect (oracle) visual inputs. This provides evidence that deep neural networks can learn to reason when training conditions are favorable enough. However, transferring this learned knowledge to deployable models is a challenge, as much of it is lost during the transfer. We propose a method for knowledge transfer based on a regularization term in our loss function, supervising the sequence of required reasoning operations. We provide a theoretical analysis based on PAC-learning, showing that such program prediction can lead to decreased sample complexity under mild hypotheses. We also demonstrate the effectiveness of this approach experimentally on the GQA dataset and show its complementarity to BERT-like self-supervised pre-training.
References in corpus (8)
- Adam: A Method for Stochastic Optimization
- Gaussian Error Linear Units (GELUs)
- Shortcut Learning in Deep Neural Networks
- The Neuro-Symbolic Concept Learner: Interpreting Scenes, Words, and Sentences From Natural Supervision
- Quantifying the Carbon Emissions of Machine Learning
- What Can Neural Networks Reason About?
- Hopfield Networks is All You Need
- Unshuffling Data for Improved Generalization