{"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/aviary-training-language-agents-on","title":"Aviary: training language agents on challenging scientific tasks","arxiv_id":"2412.21154","date":"2024-12-30","proceeding":null,"authors":["Siddharth Narayanan","James D. Braza","Ryan-Rhys Griffiths","Manu Ponnapati","Albert Bou","Jon Laurent","Ori Kabeli","Geemi Wellawatte","Sam Cox","Samuel G. Rodriques","Andrew D. White"],"abstract":"Solving complex real-world tasks requires cycles of actions and observations. This is particularly true in science, where tasks require many cycles of analysis, tool use, and experimentation. Language agents are promising for automating intellectual tasks in science because they can interact with tools via natural language or code. Yet their flexibility creates conceptual and practical challenges for software implementations, since agents may comprise non-standard components such as internal reasoning, planning, tool usage, as well as the inherent stochasticity of temperature-sampled language models. Here, we introduce Aviary, an extensible gymnasium for language agents. We formalize agents as policies solving language-grounded partially observable Markov decision processes, which we term language decision processes. We then implement five environments, including three challenging scientific environments: (1) manipulating DNA constructs for molecular cloning, (2) answering research questions by accessing scientific literature, and (3) engineering protein stability. These environments were selected for their focus on multi-step reasoning and their relevance to contemporary biology research. Finally, with online training and scaling inference-time compute, we show that language agents backed by open-source, non-frontier LLMs can match and exceed both frontier LLM agents and human experts on multiple tasks at up to 100x lower inference cost.","url_abs":"https://arxiv.org/abs/2412.21154v1","url_pdf":"https://arxiv.org/pdf/2412.21154v1.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":"aviary-training-language-agents-on","repo_url":"https://github.com/Future-House/aviary","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"aviary-training-language-agents-on","repo_url":"https://github.com/Future-House/ldp","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"aviary-training-language-agents-on","repo_url":"https://github.com/future-house/paper-qa","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[],"methods":[{"method_slug":"focus","method_name":"Focus"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2412.21154","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2412.21154"}},"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/Future-House/aviary","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/future-house/paper-qa","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Future-House/ldp","reach":null}],"summary":{"ran_violates":1,"unverified":6},"by_repo_kind":{"official":{"samples":7,"ran":1,"repositories":2}},"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":"ed4b4f0080141346","entry":"measure_consensus_proportion","repo":"Future-House/ldp","repo_kind":"official","path":"src/ldp/alg/algorithms.py","file_url":"https://github.com/Future-House/ldp/blob/HEAD/src/ldp/alg/algorithms.py","link_basis":"first_harvest_node","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"ed4b4f0080141346"}},{"code_sha256_prefix":"55531063786b8630","entry":"clean_str","repo":"Future-House/aviary","repo_kind":"official","path":"packages/hotpotqa/src/aviary/envs/hotpotqa/env.py","file_url":"https://github.com/Future-House/aviary/blob/HEAD/packages/hotpotqa/src/aviary/envs/hotpotqa/env.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"55531063786b8630"}},{"code_sha256_prefix":"a01f894e20b35ba3","entry":"encode_image_to_base64","repo":"Future-House/aviary","repo_kind":"official","path":"src/aviary/utils.py","file_url":"https://github.com/Future-House/aviary/blob/HEAD/src/aviary/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"a01f894e20b35ba3"}},{"code_sha256_prefix":"580da04192374581","entry":"fetch_with_retry","repo":"Future-House/aviary","repo_kind":"official","path":"packages/hotpotqa/src/aviary/envs/hotpotqa/env.py","file_url":"https://github.com/Future-House/aviary/blob/HEAD/packages/hotpotqa/src/aviary/envs/hotpotqa/env.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"580da04192374581"}},{"code_sha256_prefix":"4910db1a89176c6d","entry":"maybe_wait_for","repo":"Future-House/aviary","repo_kind":"official","path":"src/aviary/env.py","file_url":"https://github.com/Future-House/aviary/blob/HEAD/src/aviary/env.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"4910db1a89176c6d"}},{"code_sha256_prefix":"42be47bbb1a2ba5a","entry":"partial_format","repo":"Future-House/aviary","repo_kind":"official","path":"src/aviary/utils.py","file_url":"https://github.com/Future-House/aviary/blob/HEAD/src/aviary/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"42be47bbb1a2ba5a"}},{"code_sha256_prefix":"cb9e73e07be980af","entry":"validate_base64_image","repo":"Future-House/aviary","repo_kind":"official","path":"src/aviary/utils.py","file_url":"https://github.com/Future-House/aviary/blob/HEAD/src/aviary/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"cb9e73e07be980af"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}