{"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/playing-hard-exploration-games-by-watching","title":"Playing hard exploration games by watching YouTube","arxiv_id":"1805.11592","date":"2018-05-29","proceeding":"NeurIPS 2018 12","authors":["Yusuf Aytar","Tobias Pfaff","David Budden","Tom Le Paine","Ziyu Wang","Nando de Freitas"],"abstract":"Deep reinforcement learning methods traditionally struggle with tasks where\nenvironment rewards are particularly sparse. One successful method of guiding\nexploration in these domains is to imitate trajectories provided by a human\ndemonstrator. However, these demonstrations are typically collected under\nartificial conditions, i.e. with access to the agent's exact environment setup\nand the demonstrator's action and reward trajectories. Here we propose a\ntwo-stage method that overcomes these limitations by relying on noisy,\nunaligned footage without access to such data. First, we learn to map unaligned\nvideos from multiple sources to a common representation using self-supervised\nobjectives constructed over both time and modality (i.e. vision and sound).\nSecond, we embed a single YouTube video in this representation to construct a\nreward function that encourages an agent to imitate human gameplay. This method\nof one-shot imitation allows our agent to convincingly exceed human-level\nperformance on the infamously hard exploration games Montezuma's Revenge,\nPitfall! and Private Eye for the first time, even if the agent is not presented\nwith any environment rewards.","url_abs":"http://arxiv.org/abs/1805.11592v2","url_pdf":"http://arxiv.org/pdf/1805.11592v2.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":"playing-hard-exploration-games-by-watching","repo_url":"https://github.com/MaxSobolMark/HardRLWithYoutube","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"deep-reinforcement-learning","task_name":"Deep Reinforcement Learning"},{"task_slug":"montezumas-revenge","task_name":"Montezuma's Revenge"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1805.11592","atlas_url":"https://app.syntology.ai/?focus=1805.11592","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1805.11592"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/MaxSobolMark/HardRLWithYoutube","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":3},"by_repo_kind":{"listed":{"samples":3,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"d3321eaba7ac3a3c","entry":"generate_dataset","repo":"MaxSobolMark/HardRLWithYoutube","repo_kind":"listed","path":"train_featurizer.py","file_url":"https://github.com/MaxSobolMark/HardRLWithYoutube/blob/HEAD/train_featurizer.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d3321eaba7ac3a3c"}},{"code_sha256_prefix":"1de52ac5b00a3380","entry":"preprocess_image","repo":"MaxSobolMark/HardRLWithYoutube","repo_kind":"listed","path":"train_featurizer.py","file_url":"https://github.com/MaxSobolMark/HardRLWithYoutube/blob/HEAD/train_featurizer.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1de52ac5b00a3380"}},{"code_sha256_prefix":"adfcb20d4eed00b4","entry":"residual_block","repo":"MaxSobolMark/HardRLWithYoutube","repo_kind":"listed","path":"residual_block.py","file_url":"https://github.com/MaxSobolMark/HardRLWithYoutube/blob/HEAD/residual_block.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"adfcb20d4eed00b4"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}