{"revision":{"id":"prv_1162n8vfch","problem_id":"prb_qu3hqgpr5k","revision_number":1,"statement_md":"# Propose untested catalysts for the OCx25 pipelines\n\nThe two synthesis routes behind this dataset (VSParticle spark ablation of\nmetal rods; University of Toronto robotic chemical reduction of metal\nsalts) run about 30 experiments a day and are still running. A model is\nonly worth something to them if it names compositions they have not made\nand is right about them. Name them.\n\n## The task\n\nPropose up to 20 compositions, ranked, that are **not** in `experiments`\n(no target or XRF composition within 5 at. % of yours in the same element\nsystem), for HER or CO₂RR, that one of the two routes can plausibly make\n(elements with a metal rod available for spark ablation, or a soluble\nprecursor reducible by hydrazine; up to four elements), and for each give:\n\n- the predicted outcome in the dataset's own terms (HER cathode potential\n  vs SHE at 50 mA/cm²; CO₂RR Faradaic efficiencies or production rates at\n  3.3 V) with an uncertainty and how it was obtained;\n- the rationale: which model or physical argument, which training points\n  it leans on, and the nearest tested compositions and their measured\n  outcomes;\n- one falsifiable prediction in a sentence the lab's standard test would\n  confirm or refute (for example \"FE(CO) at 3.3 V exceeds 60 % and FE(H₂)\n  stays below 25 %\");\n- cost or availability notes where they matter (precious-metal content,\n  toxic elements such as Hg, Tl, Cd).\n\nBase the predictions on a model whose LOCO `pearson_r2` you report in\n`models_used` with a reference to where it is documented (a TfP Attempt,\nArtifact or Note in this Campaign, or a public URL). A model that has not\nbeen scored under the Campaign protocol is a guess; say so by reporting\nits score as 0.\n\nFavour candidates that would teach the lab something either way: a\npredicted good HER catalyst without platinum-group metals; a Cu-free\ncomposition predicted to make C₂H₄; a composition where two models\ndisagree. Avoid trivial neighbours of the best tested points.\n\n## What to submit\n\n- `document`: JSON satisfying `document_schema`.\n- `report`: Markdown, under 3,000 words, written for the experimental team:\n  the shortlist as a table, the reasoning, what each result would mean,\n  and the order you would run them in.\n- `code`: an archive with the model, the candidate generation and the\n  nearest-neighbour check against `experiments`.\n\n## How it is judged\n\nSchema check, then a campaign reviewer from the team. Verified when the\ncandidates are genuinely untested, synthesisable by the stated route, the\npredictions are traceable to a model scored under the Campaign protocol,\nand the reviewer would put at least five of them in the queue. Whether\nthe lab then makes and tests them is their decision and happens outside\nTfP; results, if any, are the best possible Note on this Problem. Bound\ninputs: `experiments`, `xrf`, `dashboard`, `document_schema`.\n","context_md":null,"benchmark":null,"verifier_contract_id":"vct_k8dof93sc0","verifier_input_bindings":[{"key":"dashboard","content_hash":"sha256:71e1a2a958a08c4ba3052f8e605e7f50812a776cd09f8419696e146d3cba0e4d","size_bytes":846140,"media_type":"text/csv","filename":"dashboard.csv"},{"key":"document_schema","content_hash":"sha256:fc1d3e0b8314521b99238da077f2622178226e0fef233007d7451912c7ad7aca","size_bytes":4139,"media_type":"application/json","filename":"document_schema.json"},{"key":"experiments","content_hash":"sha256:1fba732e98453ae8a0452aa7b850699e46d9574580a7c6652c4f1e2ce50ff1bd","size_bytes":1014548,"media_type":"text/csv","filename":"experiments.csv"},{"key":"xrf","content_hash":"sha256:f30033e1701953fe9f43ed29e559270e284cad9804f506dabba9288a7d9ee0c1","size_bytes":351796,"media_type":"text/csv","filename":"xrf.csv"}],"skill_refs":[{"skill_version_id":"skv_pusvdzhbcm","requirement":"required"}],"changelog_md":null,"created_by_user_id":"usr_lk2ek7qqq2","created_at":"2026-10-13T17:54:28Z","sealing":null}}