{"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/deep-reinforcement-one-shot-learning-for","title":"Deep Reinforcement One-Shot Learning for Artificially Intelligent Classification Systems","arxiv_id":"1808.01527","date":"2018-08-04","proceeding":null,"authors":["Anton Puzanov","Kobi Cohen"],"abstract":"In recent years there has been a sharp rise in networking applications, in\nwhich significant events need to be classified but only a few training\ninstances are available. These are known as cases of one-shot learning.\nExamples include analyzing network traffic under zero-day attacks, and computer\nvision tasks by sensor networks deployed in the field. To handle this\nchallenging task, organizations often use human analysts to classify events\nunder high uncertainty. Existing algorithms use a threshold-based mechanism to\ndecide whether to classify an object automatically or send it to an analyst for\ndeeper inspection. However, this approach leads to a significant waste of\nresources since it does not take the practical temporal constraints of system\nresources into account. Our contribution is threefold. First, we develop a\nnovel Deep Reinforcement One-shot Learning (DeROL) framework to address this\nchallenge. The basic idea of the DeROL algorithm is to train a deep-Q network\nto obtain a policy which is oblivious to the unseen classes in the testing\ndata. Then, in real-time, DeROL maps the current state of the one-shot learning\nprocess to operational actions based on the trained deep-Q network, to maximize\nthe objective function. Second, we develop the first open-source software for\npractical artificially intelligent one-shot classification systems with limited\nresources for the benefit of researchers in related fields. Third, we present\nan extensive experimental study using the OMNIGLOT dataset for computer vision\ntasks and the UNSW-NB15 dataset for intrusion detection tasks that demonstrates\nthe versatility and efficiency of the DeROL framework.","url_abs":"http://arxiv.org/abs/1808.01527v1","url_pdf":"http://arxiv.org/pdf/1808.01527v1.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":"deep-reinforcement-one-shot-learning-for","repo_url":"https://github.com/antonpuz/DeROL","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"deep-reinforcement-one-shot-learning-for","repo_url":"https://github.com/MillendRoy/one-shot-reinforcement-learning-classification","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"intrusion-detection","task_name":"Intrusion Detection"},{"task_slug":"one-shot-learning","task_name":"One-Shot Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}