{"about":{"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.","site":"https://codewithpapers.app","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","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/safe-mutations-for-deep-and-recurrent-neural","title":"Safe Mutations for Deep and Recurrent Neural Networks through Output Gradients","arxiv_id":"1712.06563","date":"2017-12-18","proceeding":null,"authors":["Joel Lehman","Jay Chen","Jeff Clune","Kenneth O. Stanley"],"abstract":"While neuroevolution (evolving neural networks) has a successful track record\nacross a variety of domains from reinforcement learning to artificial life, it\nis rarely applied to large, deep neural networks. A central reason is that\nwhile random mutation generally works in low dimensions, a random perturbation\nof thousands or millions of weights is likely to break existing functionality,\nproviding no learning signal even if some individual weight changes were\nbeneficial. This paper proposes a solution by introducing a family of safe\nmutation (SM) operators that aim within the mutation operator itself to find a\ndegree of change that does not alter network behavior too much, but still\nfacilitates exploration. Importantly, these SM operators do not require any\nadditional interactions with the environment. The most effective SM variant\ncapitalizes on the intriguing opportunity to scale the degree of mutation of\neach individual weight according to the sensitivity of the network's outputs to\nthat weight, which requires computing the gradient of outputs with respect to\nthe weights (instead of the gradient of error, as in conventional deep\nlearning). This safe mutation through gradients (SM-G) operator dramatically\nincreases the ability of a simple genetic algorithm-based neuroevolution method\nto find solutions in high-dimensional domains that require deep and/or\nrecurrent neural networks (which tend to be particularly brittle to mutation),\nincluding domains that require processing raw pixels. By improving our ability\nto evolve deep neural networks, this new safer approach to mutation expands the\nscope of domains amenable to neuroevolution.","url_abs":"http://arxiv.org/abs/1712.06563v3","url_pdf":"http://arxiv.org/pdf/1712.06563v3.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":"safe-mutations-for-deep-and-recurrent-neural","repo_url":"https://github.com/uber-common/safemutations","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"artificial-life","task_name":"Artificial Life"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1712.06563","atlas_url":"https://app.syntology.ai/?focus=1712.06563","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"1712.06563"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"1712.06563\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"1712.06563"},"call":"get_concepts_for_paper(arxiv_id=\"1712.06563\")"}],"by_repo":[{"repo":"uber-common/safemutations","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0}],"by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"5180922c0b4103d797d000df273f170fe81213fb3c890af02ceec6896a0527ab","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}