{"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/urlnet-learning-a-url-representation-with","title":"URLNet: Learning a URL Representation with Deep Learning for Malicious URL Detection","arxiv_id":"1802.03162","date":"2018-02-09","proceeding":null,"authors":["Hung Le","Quang Pham","Doyen Sahoo","Steven C. H. Hoi"],"abstract":"Malicious URLs host unsolicited content and are used to perpetrate\ncybercrimes. It is imperative to detect them in a timely manner. Traditionally,\nthis is done through the usage of blacklists, which cannot be exhaustive, and\ncannot detect newly generated malicious URLs. To address this, recent years\nhave witnessed several efforts to perform Malicious URL Detection using Machine\nLearning. The most popular and scalable approaches use lexical properties of\nthe URL string by extracting Bag-of-words like features, followed by applying\nmachine learning models such as SVMs. There are also other features designed by\nexperts to improve the prediction performance of the model. These approaches\nsuffer from several limitations: (i) Inability to effectively capture semantic\nmeaning and sequential patterns in URL strings; (ii) Requiring substantial\nmanual feature engineering; and (iii) Inability to handle unseen features and\ngeneralize to test data. To address these challenges, we propose URLNet, an\nend-to-end deep learning framework to learn a nonlinear URL embedding for\nMalicious URL Detection directly from the URL. Specifically, we apply\nConvolutional Neural Networks to both characters and words of the URL String to\nlearn the URL embedding in a jointly optimized framework. This approach allows\nthe model to capture several types of semantic information, which was not\npossible by the existing models. We also propose advanced word-embeddings to\nsolve the problem of too many rare words observed in this task. We conduct\nextensive experiments on a large-scale dataset and show a significant\nperformance gain over existing methods. We also conduct ablation studies to\nevaluate the performance of various components of URLNet.","url_abs":"http://arxiv.org/abs/1802.03162v2","url_pdf":"http://arxiv.org/pdf/1802.03162v2.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":"urlnet-learning-a-url-representation-with","repo_url":"https://github.com/Antimalweb/URLNet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"urlnet-learning-a-url-representation-with","repo_url":"https://github.com/AdiCovaci/urlnet_3class","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"urlnet-learning-a-url-representation-with","repo_url":"https://github.com/MJafarMashhadi/Haplophysh","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"machine-learning","task_name":"BIG-bench Machine Learning"},{"task_slug":"feature-engineering","task_name":"Feature Engineering"},{"task_slug":"word-embeddings","task_name":"Word Embeddings"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1802.03162","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}