{"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/apk2vec-semi-supervised-multi-view","title":"apk2vec: Semi-supervised multi-view representation learning for profiling Android applications","arxiv_id":"1809.05693","date":"2018-09-15","proceeding":null,"authors":["Annamalai Narayanan","Charlie Soh","Lihui Chen","Yang Liu","Lipo Wang"],"abstract":"Building behavior profiles of Android applications (apps) with holistic, rich\nand multi-view information (e.g., incorporating several semantic views of an\napp such as API sequences, system calls, etc.) would help catering downstream\nanalytics tasks such as app categorization, recommendation and malware analysis\nsignificantly better. Towards this goal, we design a semi-supervised\nRepresentation Learning (RL) framework named apk2vec to automatically generate\na compact representation (aka profile/embedding) for a given app. More\nspecifically, apk2vec has the three following unique characteristics which make\nit an excellent choice for largescale app profiling: (1) it encompasses\ninformation from multiple semantic views such as API sequences, permissions,\netc., (2) being a semi-supervised embedding technique, it can make use of\nlabels associated with apps (e.g., malware family or app category labels) to\nbuild high quality app profiles, and (3) it combines RL and feature hashing\nwhich allows it to efficiently build profiles of apps that stream over time\n(i.e., online learning). The resulting semi-supervised multi-view hash\nembeddings of apps could then be used for a wide variety of downstream tasks\nsuch as the ones mentioned above. Our extensive evaluations with more than\n42,000 apps demonstrate that apk2vec's app profiles could significantly\noutperform state-of-the-art techniques in four app analytics tasks namely,\nmalware detection, familial clustering, app clone detection and app\nrecommendation.","url_abs":"http://arxiv.org/abs/1809.05693v1","url_pdf":"http://arxiv.org/pdf/1809.05693v1.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":[],"tasks":[{"task_slug":"clone-detection","task_name":"Clone Detection"},{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":"malware-analysis","task_name":"Malware Analysis"},{"task_slug":"malware-detection","task_name":"Malware Detection"},{"task_slug":"representation-learning","task_name":"Representation Learning"}],"methods":[],"datasets_introduced":[{"slug":"the-manifest-and-store-data-of-870515-android","name":"The manifest and store data of 870,515 Android mobile applications","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}