Learn Before
Dense-Seeded Local Diffusion with Additive Hop Penalty and Max Aggregation
Method Part 2: Bidirectional Prerequisite Diffusion with Role-Aware (Auditable Strict-Parity Graph-RAG Paper)
Bidirectional Diffusion (Method) in Auditable Strict-Parity Evaluation of Prerequisite-Graph Retrieval for RAG under Leakage Controls
Max-Aggregation Final Node Score in Bidirectional Diffusion
The final score of a candidate concept node under bidirectional diffusion is the maximum over its own dense score and the best upward and downward reaching paths:
where and mathcal{P}_downarrow(v) are the truncated ancestor and descendant path sets from dense seeds. Using rather than summing (as in random-walk propagation) keeps the score equal to the strongest single justification for — either being a strong seed itself or being reachable via a short, low-penalty path from one.
0
1
Tags
Science
Auditable Strict-Parity Evaluation of Prerequisite-Graph Retrieval for RAG under Leakage Controls
Related
Hop-Penalized Path Score in Bidirectional Diffusion
Max-Aggregation Final Node Score in Bidirectional Diffusion
Deterministic BFS Upward/Downward Traversal from Dense Seeds in Bidirectional Diffusion
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
Adaptive Ancestor/Descendant Depth Controller Formula
Bidirectional Diffusion Hyperparameter Settings (LectureBank-Full Main Runs)
Deterministic BFS Upward/Downward Traversal from Dense Seeds in Bidirectional Diffusion
Hop-Penalized Path Score in Bidirectional Diffusion
Max-Aggregation Final Node Score in Bidirectional Diffusion