{"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/analysis-and-optimization-of-convolutional","title":"Analysis and Optimization of Convolutional Neural Network Architectures","arxiv_id":"1707.09725","date":"2017-07-31","proceeding":null,"authors":["Martin Thoma"],"abstract":"Convolutional Neural Networks (CNNs) dominate various computer vision tasks\nsince Alex Krizhevsky showed that they can be trained effectively and reduced\nthe top-5 error from 26.2 % to 15.3 % on the ImageNet large scale visual\nrecognition challenge. Many aspects of CNNs are examined in various\npublications, but literature about the analysis and construction of neural\nnetwork architectures is rare. This work is one step to close this gap. A\ncomprehensive overview over existing techniques for CNN analysis and topology\nconstruction is provided. A novel way to visualize classification errors with\nconfusion matrices was developed. Based on this method, hierarchical\nclassifiers are described and evaluated. Additionally, some results are\nconfirmed and quantified for CIFAR-100. For example, the positive impact of\nsmaller batch sizes, averaging ensembles, data augmentation and test-time\ntransformations on the accuracy. Other results, such as the positive impact of\nlearned color transformation on the test accuracy could not be confirmed. A\nmodel which has only one million learned parameters for an input size of\n32x32x3 and 100 classes and which beats the state of the art on the benchmark\ndataset Asirra, GTSRB, HASYv2 and STL-10 was developed.","url_abs":"http://arxiv.org/abs/1707.09725v1","url_pdf":"http://arxiv.org/pdf/1707.09725v1.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":"analysis-and-optimization-of-convolutional","repo_url":"https://github.com/MartinThoma/clana","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"data-augmentation","task_name":"Data Augmentation"},{"task_slug":"object-recognition","task_name":"Object Recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}