FA-99451 / Vector retrieval ranking / Open access

Vector retrieval: Retrieval channel weights disappear during fusion · case 01

Retrieval channel weights disappear during fusion.

Verified by executionVariant 1 · 6 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

Each channel contributes the same reciprocal-rank value regardless of its configured reliability weight.

THE FAILURE

Each channel contributes the same reciprocal-rank value regardless of its configured reliability weight.

Unsuccessful approach: Applying the weight but starting at zero assigns an extra reciprocal contribution to the first-ranked item.

Case contract

Lists contain unique IDs in ranked order and have one corresponding positive integer weight each. Offset is a positive integer. Sum weight/(offset+one-based rank); missing IDs contribute zero. Return IDs by descending exact Fraction total then ID.

Why this case matters

An offline vector retrieval toy model isolates one ranking contract from embedding generation, storage, and serving.

1 / The failure

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(lists, weights, offset):
    from fractions import Fraction
    s={}
    for xs in lists:
     for rank,(ident,score) in enumerate(xs,1): s[ident]=s.get(ident,0)+Fraction(1,offset+rank)
    return sorted(s,key=lambda x:(-s[x],x))
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('trusted channel weight',solve([[['a',9],['x',0]],[['b',8],['x',0]]],[5,1],1),['a','x','b'])
check('weight reverses winner',solve([[['a',1]],[['b',1]]],[1,4],2),['b','a'])
check('missing candidate contribution',solve([[['a',1],['x',0]],[['b',1],['x',0]]],[3,2],1),['x','a','b'])
check('deep rank weighted',solve([[['a',1],['b',0],['x',0]],[['b',1]]],[1,4],1),['b','a','x'])
check('one list weight',solve([[['z',0],['a',0]]],[N+2],N),['z','a'])
check('several unequal weights',solve([[['a',1],['b',0]],[['c',1],['b',0]]],[N+1,2],N),['b','a','c'])
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
trusted channel weight	`['x', 'a', 'b']`	`['a', 'x', 'b']`	Failed
weight reverses winner	`['a', 'b']`	`['b', 'a']`	Failed
missing candidate contribution	`['x', 'a', 'b']`	`['x', 'a', 'b']`	Passed
deep rank weighted	`['b', 'a', 'x']`	`['b', 'a', 'x']`	Passed
one list weight	`['z', 'a']`	`['z', 'a']`	Passed
several unequal weights	`['b', 'a', 'c']`	`['b', 'a', 'c']`	Passed

SHA-256 / 692f467371588a36cde09d7a0b2eaf0eec5bce4563c0d09218bc890acc16c4c5

2 / The unsuccessful fix

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(lists, weights, offset):
    from fractions import Fraction
    s={}
    for xs,w in zip(lists,weights):
     for rank,(ident,score) in enumerate(xs): s[ident]=s.get(ident,0)+Fraction(w,offset+rank)
    return sorted(s,key=lambda x:(-s[x],x))
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('trusted channel weight',solve([[['a',9],['x',0]],[['b',8],['x',0]]],[5,1],1),['a','x','b'])
check('weight reverses winner',solve([[['a',1]],[['b',1]]],[1,4],2),['b','a'])
check('missing candidate contribution',solve([[['a',1],['x',0]],[['b',1],['x',0]]],[3,2],1),['x','a','b'])
check('deep rank weighted',solve([[['a',1],['b',0],['x',0]],[['b',1]]],[1,4],1),['b','a','x'])
check('one list weight',solve([[['z',0],['a',0]]],[N+2],N),['z','a'])
check('several unequal weights',solve([[['a',1],['b',0]],[['c',1],['b',0]]],[N+1,2],N),['b','a','c'])
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
trusted channel weight	`['a', 'x', 'b']`	`['a', 'x', 'b']`	Passed
weight reverses winner	`['b', 'a']`	`['b', 'a']`	Passed
missing candidate contribution	`['a', 'x', 'b']`	`['x', 'a', 'b']`	Failed
deep rank weighted	`['b', 'a', 'x']`	`['b', 'a', 'x']`	Passed
one list weight	`['z', 'a']`	`['z', 'a']`	Passed
several unequal weights	`['a', 'b', 'c']`	`['b', 'a', 'c']`	Failed

SHA-256 / e8308b67883a735806780f0d28dddb54a2a8180c84ec67711a5a6c6e46ddedf4

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.

Verification & scope

Inputs are small exact toy vectors and scores. No production index, ANN recall, embedding quality, or latency claim is implied. 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 2026-09-30T16:48:24.475676+00:00.

Case digest / e5c8607a784be95d7f873613826dab12dbee6109d733fd90f6f0f7d29f14446c