{"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/graph-networks-as-learnable-physics-engines","title":"Graph networks as learnable physics engines for inference and control","arxiv_id":"1806.01242","date":"2018-06-04","proceeding":"ICML 2018 7","authors":["Alvaro Sanchez-Gonzalez","Nicolas Heess","Jost Tobias Springenberg","Josh Merel","Martin Riedmiller","Raia Hadsell","Peter Battaglia"],"abstract":"Understanding and interacting with everyday physical scenes requires rich\nknowledge about the structure of the world, represented either implicitly in a\nvalue or policy function, or explicitly in a transition model. Here we\nintroduce a new class of learnable models--based on graph networks--which\nimplement an inductive bias for object- and relation-centric representations of\ncomplex, dynamical systems. Our results show that as a forward model, our\napproach supports accurate predictions from real and simulated data, and\nsurprisingly strong and efficient generalization, across eight distinct\nphysical systems which we varied parametrically and structurally. We also found\nthat our inference model can perform system identification. Our models are also\ndifferentiable, and support online planning via gradient-based trajectory\noptimization, as well as offline policy optimization. Our framework offers new\nopportunities for harnessing and exploiting rich knowledge about the world, and\ntakes a key step toward building machines with more human-like representations\nof the world.","url_abs":"http://arxiv.org/abs/1806.01242v1","url_pdf":"http://arxiv.org/pdf/1806.01242v1.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":"graph-networks-as-learnable-physics-engines","repo_url":"https://github.com/fxia22/gn.pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"inductive-bias","task_name":"Inductive Bias"},{"task_slug":"weather-forecasting","task_name":"Weather Forecasting"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/weather-forecasting-on-sd","task":"Weather Forecasting","dataset":"SD","model":"GN-skip","rank_in_archive_order":9,"of":10,"metrics":{"MSE (t+1)":"0.6543 ± 0.1195","MSE (t+6)":"0.9872 ± 0.2425"},"uses_additional_data":false},{"leaderboard":"/sota/weather-forecasting-on-sd","task":"Weather Forecasting","dataset":"SD","model":"GN-only","rank_in_archive_order":10,"of":10,"metrics":{"MSE (t+1)":"0.7007 ± 0.0848","MSE (t+6)":"1.0422 ± 0.0673"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.01242","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1806.01242"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+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/fxia22/gn.pytorch","reach":null}],"summary":{"ran_draft_wrong":1},"by_repo_kind":{"listed":{"samples":1,"ran":1,"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":1,"samples":[{"code_sha256_prefix":"55cdceb4f1fdcd41","entry":"evaluate_graph_loss","repo":"fxia22/gn.pytorch","repo_kind":"listed","path":"evaluate_gn.py","file_url":"https://github.com/fxia22/gn.pytorch/blob/HEAD/evaluate_gn.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"55cdceb4f1fdcd41"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}