Bayesian Inverse Graphics for Few-Shot Concept Learning
This work addresses the challenge of data-efficient learning in computer vision, which is important for applications with limited training data, though it appears to be an incremental advance combining existing techniques.
The authors tackled the problem of few-shot concept learning in computer vision by developing a Bayesian inverse graphics model that learns from minimal data using probabilistic programs of objects. Their model outperformed existing few-shot neural classification algorithms and demonstrated generalization across varying conditions.
Humans excel at building generalizations of new concepts from just one single example. Contrary to this, current computer vision models typically require large amount of training samples to achieve a comparable accuracy. In this work we present a Bayesian model of perception that learns using only minimal data, a prototypical probabilistic program of an object. Specifically, we propose a generative inverse graphics model of primitive shapes, to infer posterior distributions over physically consistent parameters from one or several images. We show how this representation can be used for downstream tasks such as few-shot classification and pose estimation. Our model outperforms existing few-shot neural-only classification algorithms and demonstrates generalization across varying lighting conditions, backgrounds, and out-of-distribution shapes. By design, our model is uncertainty-aware and uses our new differentiable renderer for optimizing global scene parameters through gradient descent, sampling posterior distributions over object parameters with Markov Chain Monte Carlo (MCMC), and using a neural based likelihood function.