We propose a weakly-supervised framework for the semantic segmentation of circular-scan synthetic-aperture-sonar (CSAS) imagery. The first part of our framework is trained in a supervised manner, on image-level labels, to uncover a set of semi-sparse, spatially-discriminative regions in each image. The classification uncertainty of each region is then evaluated. Those areas with the lowest uncertainties are then chosen to be weakly labeled segmentation seeds, at the pixel level, for the second part of the framework. Each of the seed extents are progressively resized according to an unsupervised, information-theoretic loss with structured-prediction regularizers. This reshaping process uses multi-scale, adaptively-weighted features to delineate class-specific transitions in local image content. Content-addressable memories are inserted at various parts of our framework so that it can leverage features from previously seen images to improve segmentation performance for related images. We evaluate our weakly-supervised framework using real-world CSAS imagery that contains over ten seafloor classes and ten target classes. We show that our framework performs comparably to nine fully-supervised deep networks. Our framework also outperforms eleven of the best weakly-supervised deep networks. We achieve state-of-the-art performance when pre-training on natural imagery. The average absolute performance gap to the next-best weakly-supervised network is well over ten percent for both natural imagery and sonar imagery. This gap is found to be statistically significant.

A few interesting papers andposters at the Neuromatch 4 Conference (held in December 2021), with topics at the intersection of Neuroscience and AI. In the first talk by authors Ilena Jones and Konrad Kording, dendritic processes (forming a tree-like structure) are used to solve machine learning tasks as opposed to simple connection weights. It's interesting to see how modelling dendrites and synapses with constraints can solve machine learning tasks. In this paper, authors Eric Elmoznino and Michael Bonner discuss how scaling the manifold dimensionality increases the performance for vision related tasks, as opposed to various theories that the brain tries to represent information in a really compressed format. NeuralNLP is an improvement over NeuralNLP which uses rule-based NLP methods for searching various queries related to Fruit Fly Brain.