FA-99996 / Evaluation metric aggregation / Open access
Multilabel subset accuracy counts matching labels · case 01
Multilabel subset accuracy counts matching labels.
ROOT CAUSE
Partial label overlap is reported as a correct whole example.
THE FAILURE
Partial label overlap is reported as a correct whole example.
Unsuccessful approach: Comparing raw lists makes equivalent label sets depend on their ordering.
Case contract
Return fraction of examples whose predicted label set equals its true label set; empty outer lists yield 0.
Why this case matters
A small explicit evaluation contract makes aggregation behavior reproducible across labeled examples.
1 / The failure
Exit 1"""Failure Map reference implementation. Python standard library only."""
import json
N = 1
observations = []
def solve(truth, pred):
return sum(bool(set(a)&set(b)) for a,b in zip(truth,pred))/len(truth) if truth else 0
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('partial overlap',solve([[1,2],[2]],[[1],[2]]),.5)
check('ordering irrelevant',solve([[1,2]],[[2,1]]),1)
check('empty label sets',solve([[],[]],[[],[1]]),.5)
check('empty examples',solve([],[]),0)
check('exact singleton',solve([[N]],[[N]]),1)
check('different singleton',solve([[N]],[[N+1]]),0)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
| Boundary fixture | Actual | Expected | Outcome |
|---|---|---|---|
| partial overlap | 1.0 | 0.5 | Failed |
| ordering irrelevant | 1.0 | 1 | Passed |
| empty label sets | 0.0 | 0.5 | Failed |
| empty examples | 0 | 0 | Passed |
| exact singleton | 1.0 | 1 | Passed |
| different singleton | 0.0 | 0 | Passed |
SHA-256 / 2b1c6b9c513eab3ea4d3292c4c64a111f9770a145419ee8128ca588113f6ef3b
2 / The unsuccessful fix
Exit 1"""Failure Map reference implementation. Python standard library only."""
import json
N = 1
observations = []
def solve(truth, pred):
return sum(a==b for a,b in zip(truth,pred))/len(truth) if truth else 0
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('partial overlap',solve([[1,2],[2]],[[1],[2]]),.5)
check('ordering irrelevant',solve([[1,2]],[[2,1]]),1)
check('empty label sets',solve([[],[]],[[],[1]]),.5)
check('empty examples',solve([],[]),0)
check('exact singleton',solve([[N]],[[N]]),1)
check('different singleton',solve([[N]],[[N+1]]),0)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
| Boundary fixture | Actual | Expected | Outcome |
|---|---|---|---|
| partial overlap | 0.5 | 0.5 | Passed |
| ordering irrelevant | 0.0 | 1 | Failed |
| empty label sets | 0.5 | 0.5 | Passed |
| empty examples | 0 | 0 | Passed |
| exact singleton | 1.0 | 1 | Passed |
| different singleton | 0.0 | 0 | Passed |
SHA-256 / a3b093171b296a7e4a2e613890cf0196039b04a0f6c250f11b7184a7c4b91b21
HELD IN THE MEMBER ARCHIVE
The verified repair and its recorded checks are member-only.
This mechanism has 6 recorded checks per implementation. The open-access tier publishes the failure and the unsuccessful fix; the repaired source that passes every check, and the observations that prove it, are available to members.
Every case sharing this mechanism uses the same contract and the same repair, so this one record is held back for all of them.
Member access is invitation-based. Sign in with your invited account to inspect the repair.
Sign in to the archive ↗Verification & scope
Inputs are finite toy labels and scores; this defines no benchmark protocol, production metric library, or model quality claim. 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.
Observations recorded using Python 3.12.14 at 2026-09-30T16:48:30.387883+00:00.
Case digest / 99103ba6cf2d055d0b77804cbbc360a53dbd873b7a6e7fccca3c856e7d49cc96