{"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/emergence-of-communication-in-an-interactive","title":"Emergence of Communication in an Interactive World with Consistent Speakers","arxiv_id":"1809.00549","date":"2018-09-03","proceeding":null,"authors":["Ben Bogin","Mor Geva","Jonathan Berant"],"abstract":"Training agents to communicate with one another given task-based supervision\nonly has attracted considerable attention recently, due to the growing interest\nin developing models for human-agent interaction. Prior work on the topic\nfocused on simple environments, where training using policy gradient was\nfeasible despite the non-stationarity of the agents during training. In this\npaper, we present a more challenging environment for testing the emergence of\ncommunication from raw pixels, where training using policy gradient fails. We\npropose a new model and training algorithm, that utilizes the structure of a\nlearned representation space to produce more consistent speakers at the initial\nphases of training, which stabilizes learning. We empirically show that our\nalgorithm substantially improves performance compared to policy gradient. We\nalso propose a new alignment-based metric for measuring context-independence in\nemerged communication and find our method increases context-independence\ncompared to policy gradient and other competitive baselines.","url_abs":"http://arxiv.org/abs/1809.00549v2","url_pdf":"http://arxiv.org/pdf/1809.00549v2.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":"emergence-of-communication-in-an-interactive","repo_url":"https://github.com/benbogin/emergence-communication-cco","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1809.00549","atlas_url":"https://app.syntology.ai/?focus=1809.00549","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}