A Hybrid PageRank Impact Evaluator

Existing funding systems often struggle to incentivize contributions to public and common goods because they lack mechanisms for coordinating diverse agents toward shared objectives. The Generalized Impact Evaluators framework proposes modular systems that retrospectively reward impact by measuring, evaluating, and distributing rewards to contributors. Building on this framework, we propose a graph-based mechanism and impact evaluator that models contributors, the intermediate results of their work such as code, papers, or designs, and the eventual outcomes like funding or citations as a directed, weighted graph. Our model introduces a hybrid attribution algorithm combining forward PageRank and reverse personalized PageRank to estimate each agent’s contribution to observed outcomes. A tunable configuration object allows communities to adjust per-edge and per-node weights, making the model flexible and auditable. We describe how the graph, algorithm, and reward functions implement the GIE tuple IE \= {r, e, m, S}, present illustrative use cases, and discuss extensions such as temporal decay, confidence scores, and decentralized attestation.