{"abstract":"A parameter absent from a microbatch is counted as a zero update.","category":"Gradient accumulation weighting","checks":7,"contract":"Rows are microbatch coordinate sums; None means a parameter was absent and contributes zero to the mean over all microbatches. active_counts documents participation but is not a divisor.","contract_signature":"rows, active_counts","evaluation_group":"gradient-accumulation-parameter_participation","failed_approach":"Dividing only by active participation omits the zero contribution of absent parameters.","family":"w3-gradient-accumulation-weighting-parameter-participation","id":"FA-98886","implementations":{"attempt":{"sha256":"1a9d2715a2c4215a4d3f225539e3f46a8c83e2919ffaf5aa9fed89d781f34312","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(rows, active_counts):\n    return [sum(row[j] for row in rows if row[j] is not None) / max(1,sum(row[j] is not None for row in rows)) for j in range(len(rows[0]))] if rows else []\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ncheck('missing treated as zero',solve([[2*N,None],[4*N,6*N]],[1,2]),[3*N,3*N])\ncheck('all present',solve([[2*N,4*N],[6*N,8*N]],[2,2]),[4*N,6*N])\ncheck('single missing',solve([[None,5*N]],[0,1]),[0,5*N])\ncheck('empty rows',solve([],[]),[])\ncheck('zero active count',solve([[0,None],[0,None]],[0,0]),[0,0])\ncheck('late participation',solve([[None,2*N],[9*N,None],[3*N,4*N]],[1,1,2]),[4*N,2*N])\ncheck('zero gradient',solve([[0,None],[4*N,2*N]],[1,1]),[2*N,N])\nprint(json.dumps({\"observations\": observations, \"passed\": all(x[\"passed\"] for x in observations)}, ensure_ascii=False))\nraise SystemExit(0 if all(x[\"passed\"] for x in observations) else 1)\n"},"broken":{"sha256":"0d02e4e185a14d0abe6a78c97453c415215147b7e1e137b771b44d785053713e","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(rows, active_counts):\n    return [sum((row[j] or 0) for row in rows) / max(1,sum(active_counts)) for j in range(len(rows[0]))] if rows else []\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ncheck('missing treated as zero',solve([[2*N,None],[4*N,6*N]],[1,2]),[3*N,3*N])\ncheck('all present',solve([[2*N,4*N],[6*N,8*N]],[2,2]),[4*N,6*N])\ncheck('single missing',solve([[None,5*N]],[0,1]),[0,5*N])\ncheck('empty rows',solve([],[]),[])\ncheck('zero active count',solve([[0,None],[0,None]],[0,0]),[0,0])\ncheck('late participation',solve([[None,2*N],[9*N,None],[3*N,4*N]],[1,1,2]),[4*N,2*N])\ncheck('zero gradient',solve([[0,None],[4*N,2*N]],[1,1]),[2*N,N])\nprint(json.dumps({\"observations\": observations, \"passed\": all(x[\"passed\"] for x in observations)}, ensure_ascii=False))\nraise SystemExit(0 if all(x[\"passed\"] for x in observations) else 1)\n"}},"limitations":"Inputs are scalar or short-list summaries and do not model autodiff, devices, or an optimizer. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.","method":"Deterministic executable model with adversarial boundary fixtures.","provenance":{"created_by":"Failure Map","dependencies":"Python standard library","family":"w3-gradient-accumulation-weighting-parameter-participation","generated_at":"2026-09-30T16:48:18.065493+00:00","license":"CC0-1.0","python":"3.12.14","seed":1,"split":"open-access"},"relevance":"A deterministic toy model isolates one local accumulation contract.","root_cause":"The broken divisor sums activity counters rather than counting microbatches; absent parameters contribute zero in the contract.","sha256":"50867ada14df70af7b437561368252e3fac7073de9b016a339466a3ed562e47d","title":"A parameter absent from a microbatch is counted as a zero update · case 01","variant":1,"variant_policy":"The fixture code references N, supplied as 1 through 5. Fixed boundary checks may repeat. These records share one mechanism and must remain together in evaluation splits.","verified":true,"visibility":"public","verification":{"attempt":{"elapsed_ms":32.522,"exit_code":1,"observations":[{"actual":[3.0,6.0],"check":"missing treated as zero","expected":[3,3],"passed":false},{"actual":[4.0,6.0],"check":"all present","expected":[4,6],"passed":true},{"actual":[0.0,5.0],"check":"single missing","expected":[0,5],"passed":true},{"actual":[],"check":"empty rows","expected":[],"passed":true},{"actual":[0.0,0.0],"check":"zero active count","expected":[0,0],"passed":true},{"actual":[6.0,3.0],"check":"late participation","expected":[4,2],"passed":false},{"actual":[2.0,2.0],"check":"zero gradient","expected":[2,1],"passed":false}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"missing treated as zero\", \"actual\": [3.0, 6.0], \"expected\": [3, 3], \"passed\": false}, {\"check\": \"all present\", \"actual\": [4.0, 6.0], \"expected\": [4, 6], \"passed\": true}, {\"check\": \"single missing\", \"actual\": [0.0, 5.0], \"expected\": [0, 5], \"passed\": true}, {\"check\": \"empty rows\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"zero active count\", \"actual\": [0.0, 0.0], \"expected\": [0, 0], \"passed\": true}, {\"check\": \"late participation\", \"actual\": [6.0, 3.0], \"expected\": [4, 2], \"passed\": false}, {\"check\": \"zero gradient\", \"actual\": [2.0, 2.0], \"expected\": [2, 1], \"passed\": false}], \"passed\": false}\n"},"broken":{"elapsed_ms":29.901,"exit_code":1,"observations":[{"actual":[2.0,2.0],"check":"missing treated as zero","expected":[3,3],"passed":false},{"actual":[2.0,3.0],"check":"all present","expected":[4,6],"passed":false},{"actual":[0.0,5.0],"check":"single missing","expected":[0,5],"passed":true},{"actual":[],"check":"empty rows","expected":[],"passed":true},{"actual":[0.0,0.0],"check":"zero active count","expected":[0,0],"passed":true},{"actual":[3.0,1.5],"check":"late participation","expected":[4,2],"passed":false},{"actual":[2.0,1.0],"check":"zero gradient","expected":[2,1],"passed":true}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"missing treated as zero\", \"actual\": [2.0, 2.0], \"expected\": [3, 3], \"passed\": false}, {\"check\": \"all present\", \"actual\": [2.0, 3.0], \"expected\": [4, 6], \"passed\": false}, {\"check\": \"single missing\", \"actual\": [0.0, 5.0], \"expected\": [0, 5], \"passed\": true}, {\"check\": \"empty rows\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"zero active count\", \"actual\": [0.0, 0.0], \"expected\": [0, 0], \"passed\": true}, {\"check\": \"late participation\", \"actual\": [3.0, 1.5], \"expected\": [4, 2], \"passed\": false}, {\"check\": \"zero gradient\", \"actual\": [2.0, 1.0], \"expected\": [2, 1], \"passed\": true}], \"passed\": false}\n"}},"member_only":{"stages":["fixed"],"fields":["implementations.fixed","verification.fixed","harness","repair"],"note":"The verified repair, its recorded checks, the repair description, and the scoring harness are available to members."}}