Learn Before
Method Part 2: Bidirectional Prerequisite Diffusion with Role-Aware (Auditable Strict-Parity Graph-RAG Paper)
Analysis And Limitations in Auditable Strict-Parity Evaluation of Prerequisite-Graph Retrieval for RAG under Leakage Controls
Method Part 2: MOOC-CS Prerequisite Benchmark (Auditable Strict-Parity Graph-RAG Paper)
Language-Matched Seeding as a Prerequisite for Graph-Expansion Gains
The Analysis-section operational lesson drawn from MOOC-CS is not that adaptive gating always helps, but rather that language-matched seeding is a prerequisite to benefiting from graph expansion. Concretely, on MOOC-CS the contrast gate adds almost nothing on top of diffusion because the dense seed pool is already weakened by bilingual aliasing and sparse/noisy edges, so graph traversal has no high-quality seeds to expand from. The lesson reframes graph-RAG advice for prerequisite retrieval: invest in matching the encoder/query language to the corpus before attributing wins or losses to traversal-policy choices.
0
1
Tags
Science
Auditable Strict-Parity Evaluation of Prerequisite-Graph Retrieval for RAG under Leakage Controls
Related
LectureBank-Full R@10 Gain from Diffusion and Role-Aware Quotas
LectureBank-Full Target-Disjoint R@10 Result (n=164): Diffusion Gain Survives, Adaptive Tied
Language-Matched Seeding as a Prerequisite for Graph-Expansion Gains
Hop-Penalized Path Score in Bidirectional Diffusion
Max-Aggregation Final Node Score in Bidirectional Diffusion
LectureBank-Full Decomposition: Diffusion+Quotas Drive ~18 R@10 Points; Contrast Gating Adds At Most ~1 Point (Statistically Tied)
Bounded Benchmark Validity: Two Question Families and 21/18 Unique Held-Out Targets Cap Statistical Power
Restrained Claim Scope: No Automatic-Judge Validation, No Semantic-Evasion Claim, No Robust End-to-End QA Gains
Language-Matched Seeding as a Prerequisite for Graph-Expansion Gains
HotpotQA External-Validity Probe: Adaptive Depth Does Not Transfer to a Denser Non-Prerequisite Graph (FullWiki-1k: Flat 93.4 / Hier 92.9 / Adaptive 94.0 R@10)
Multilingual Encoder Alone Does Not Improve MOOC-CS Recall (Hierarchical R@10 = 22.3 vs 23.1)
Graph Effect on MOOC-CS Is Conditional on Dense Seed Pool Quality
MOOC-CS Error Taxonomy: Residual Failures Dominated by Distant Misses and Bilingual Aliasing
Language-Matched Seeding as a Prerequisite for Graph-Expansion Gains