In this work, we propose Asynchronous Perception Machine (APM), a computationally-efficient architecture for test-time-training (TTT). APM can process patches of an image one at a time in any order asymmetrically, and still encode semantic-awareness in the net. We demonstrate APM’s ability to recognize out-of-distribution images without dataset-specific pre-training, augmentation orany-pretext task. APM offers competitive performance over existing TTT approaches. To perform TTT, APM just distills test sample’s representation once. APM possesses a unique property: it can learn using just this single representation and starts predicting semantically-aware features. APM’s ability to recover semantic information from a global CLS token validates the insight that CLStokens encode geometric-information of a given scene and can be recovered using appropriate inductive-biases. This offers a novel-insight with consequences for representational-learning. APM demostrates potential applications beyond test-time-training: APM can scale up to a dataset of 2D images and yield semantic-clusterings in a single forward pass. APM also provides first empirical evidence towards validating Hinton at Al’s GLOM’s insight, i.e. if input percept is a field. Therefore, APM helps our community converge towards an implementation which can do both interpolation and perception on a shared-connectionist hardware. Ourcodebase has been made available at https://rajatmodi62.github.io/apmprojectpage/--------It now appears that some of the ideas in GLOM could be made to work.https://www.technologyreview.com/2021/04/16/1021871/geoffrey-hinton-glom-godfather-ai-neural-networks/.-""""""-. .' './ O O \| O | \ '------' / '. .' '-....-'A silent man in deep-contemplation.Silent man emerges only sometimes.And he loves all.