{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/dnn-feature-map-compression-using-learned","title":"DNN Feature Map Compression using Learned Representation over GF(2)","arxiv_id":"1808.05285","date":"2018-08-15","proceeding":"ICLR 2018 1","authors":["Denis A. Gudovskiy","Alec Hodgkinson","Luca Rigazio"],"abstract":"In this paper, we introduce a method to compress intermediate feature maps of\ndeep neural networks (DNNs) to decrease memory storage and bandwidth\nrequirements during inference. Unlike previous works, the proposed method is\nbased on converting fixed-point activations into vectors over the smallest\nGF(2) finite field followed by nonlinear dimensionality reduction (NDR) layers\nembedded into a DNN. Such an end-to-end learned representation finds more\ncompact feature maps by exploiting quantization redundancies within the\nfixed-point activations along the channel or spatial dimensions. We apply the\nproposed network architectures derived from modified SqueezeNet and MobileNetV2\nto the tasks of ImageNet classification and PASCAL VOC object detection.\nCompared to prior approaches, the conducted experiments show a factor of 2\ndecrease in memory requirements with minor degradation in accuracy while adding\nonly bitwise computations.","url_abs":"http://arxiv.org/abs/1808.05285v1","url_pdf":"http://arxiv.org/pdf/1808.05285v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"dnn-feature-map-compression-using-learned","repo_url":"https://github.com/gudovskiy/fmap_compression","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"dimensionality-reduction","task_name":"Dimensionality Reduction"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"quantization","task_name":"Quantization"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"fire-module","method_name":"Fire Module"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"squeezenet","method_name":"SqueezeNet"},{"method_slug":"xavier-initialization","method_name":"Xavier Initialization"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}