17 Literature Bridge
Author: Todd B. Adams Reinforces: Proposal §2 — Theoretical Foundations · Reading order: doc 10 of the research-platform package
The claim of this doc. The proposal’s bibliography and the platform’s own reference library are one continuous body of work, not two disconnected reading lists. This doc reconciles them and states the single convention that keeps every cited paper captured.
17.1 The proposal’s §2 papers — bundled in docs/phd/articles/
The network / multilayer / GNN literature that grounds the proposal is downloaded (open-access) with per-paper summaries in docs/phd/articles/:
| Paper | Role in the proposal |
|---|---|
| Musmeci, Aste & Di Matteo (2015) | KCL information-filtering-network methodology (supervisor lineage) |
| Bardoscia, Battiston, Caccioli & Caldarelli (2017) | Mechanics of cascading failure / systemic risk |
| Boccaletti et al. (2014) | Definitive multilayer-network formulation |
| Kivelä et al. (2014) | Heterogeneous vs. multiplex terminology |
| Kipf & Welling (2016) | GCN message-passing foundation |
| Yu, Yin & Zhu (2017) | Spatio-temporal GCN, adaptable to financial series |
| Caldarelli (2007) | Scale-free network physics (monograph, citation-only) |
17.2 The platform’s library — docs/reference-papers/
The asset-pricing, cost, survivorship, overfitting and crowding literature that grounds the platform’s validation gauntlet lives in docs/reference-papers/README.md. The overlap with the proposal is the bridge:
- Lou & Polk (2012), Comomentum (row 74) — already bundled; grounds the platform’s crowding producer and the proposal’s Layer-3 ownership-crowding motivation (doc 05).
- Harvey, Liu & Zhu (2016); Bailey & López de Prado (2014); Benjamini–Yekutieli (2001) — the anti-overfitting core (doc 08) that would discipline the network claims.
17.3 One gap to close, and the convention that governs it
- Raddant & Di Matteo (2023), A look at financial dependencies by means of econophysics and financial economics — the supervisor’s survey the platform already stress-tested (findings) — is not yet captured in
docs/reference-papers/. It is open-access (CC-BY) and should be bundled asRaddant2023-financial-dependencies-econophysics.pdfwhen the first network/design-briefproceeds. Flagged here so it is not missed.
The convention (platform rule, [CLAUDE.md §7]): every academic paper cited anywhere in the repo must be captured in docs/reference-papers/ — an index row plus a bundled open-access PDF (or a “paywalled — citation only” note). The proposal’s docs/phd/articles/ bundle is the network-literature satellite of that same shelf. As the research proceeds, new network/GNN citations land in one of the two homes, never uncaptured.
17.4 How the two reading lists meet
The proposal contributes the network-physics + geometric-deep-learning half; the platform contributes the asset-pricing rigor + crowding half. The PhD sits exactly at their intersection — multiplex network structure, evaluated with asset-pricing-grade anti-overfitting discipline, on a live executable platform. That intersection is the whole argument of this folder.
Cross-links: proposal §2 · phd/articles · reference papers · crowding groundwork