Analysis of Cryptocurrency Transactions from a Network Perspective
Survey of cryptocurrency-transaction research organized around three approaches: network modeling, network profiling, and network-based detection.
Cross-Domain / General Corpus
43 curated papers.
Analysis of Ego Multi-Token Transfer Networks: A Multilayer Approach
Introduces the Multilayer Token Network (MLTN) and a PageRank-CheiRank Trade Balance measure, applied to 2018-2024 Ethereum data to trace Alameda Research's fund interdependencies and token accumulation/distribution shifts.
Analyzing Voting Power in Decentralized Governance: Who Controls DAOs?
Empirically measures voting-power distribution in the Compound, Uniswap, and ENS DAOs from token-holder, delegate, proposal, and voting-record data.
Boltzmann-Gibbs Distribution on Connected Graphs
Proves that money distribution among locally-interacting agents on a connected graph converges to an exponential distribution (Poisson in a wealth-proportional-purchase variant).
ChainNet: Learning on Blockchain Graphs with Topological Features
Uses persistent homology to extract topological features from Bitcoin transaction graphs for price-fluctuation prediction, in place of standard degree-distribution graph metrics.
Compound/Uniswap Governance Analysis
Empirical study of 370+ Compound and Uniswap governance proposals finding as few as 3-5 voters sufficient to sway a majority of outcomes.
Crypto as Investable Asset Class
Identifies seven stylized facts about cryptocurrency returns via empirical asset-pricing analysis; risk-adjusted performance is broadly comparable to traditional assets despite frequent large price jumps.
Crypto Portfolios via Stable Clustering
Combines Louvain consensus clustering of price-correlation networks with ARIMA forecasts to build cryptocurrency portfolios showing positive returns up to 14 days and favorable tail-risk management.
Cryptocurrencies Activity as a Complex Network
Introduces DiLeNA, a DLT transaction-graph analysis tool; finds Bitcoin, Dogecoin, Ethereum, and Ripple transaction networks all exhibit small-world properties.
DAO Voting Resistant to Whale and Collusion
Shows Quadratic Voting alone lacks collusion resistance and proposes combining it with vote-escrowed (veToken) locking to mitigate the whale problem while gaining collusion resilience.
Decentralized Governance of AI Agents (ETHOS)
Proposes ETHOS, a Web3 governance framework for autonomous AI agents combining a risk-classified registry, soulbound-token/zero-knowledge compliance monitoring, and decentralized dispute resolution.
Decentralized Peer Review in Open Science
Proposes a blockchain-based peer-review mechanism that pays reviewers, publishes anonymized reports, tracks reviewer reputation, and issues digital certificates.
Econophysics and the Entropic Foundations of Economics
Distinguishes ontological from metaphorical entropy in economics and surveys how entropic and anti-entropic forces interact across growth theory, business cycles, and income/wealth distribution.
Engineering Token Economy with System Modeling
Develops a dynamical-systems framework, drawing on differential games and control engineering, for modeling and simulating token economies.
Evaluating Multivariate Network Visualization
Crowdsourced comparison finds node-link diagrams best for topology-plus-few-attribute tasks and adjacency matrices best for cluster tasks involving many attributes.
Evaluating Voting Design for Retroactive Funding
Formal and simulation-based analysis of quadratic, mean, and median voting in Optimism's RetroPGF program, identifying vulnerabilities in each and recommending design improvements.
Financial Contagion: Complexity Modeling
Reviews network models of financial contagion and the counterparty-default and shared-portfolio channels through which shocks propagate across interconnected institutions.
From Interacting Agents to Boltzmann-Gibbs
Proves a propagation-of-chaos result for a stochastic agent-based money-exchange model, showing convergence to a geometric wealth distribution at an almost-exponential rate in relative entropy.
GNN Financial Fraud Detection Review
Reviews 100+ studies on GNN-based financial fraud detection, presenting a unified categorization framework and identifying open limitations.
Herding in Cryptocurrency Markets
Tests for herding behavior in cryptocurrency markets under asymmetric and symmetric conditions using a Markov-Switching regime approach.
Hybrid-DAOs
Proposes 'Hybrid-DAOs' — decentralized governance paired with traditional legal frameworks — to address DAO scalability, governance, compliance, and Sybil-attack vulnerabilities.
Inside Alameda Research: A Multi-token Network Analysis
Multi-token network analysis of Ethereum accounts tied to Alameda Research, tracing token accumulation/distribution shifts leading up to its November 2022 bankruptcy.
Mapping Bitcoin Network Structures 2009-2023
Traces Bitcoin's network evolution across three phases (Exploration, Adaptation, Maturity); finds network centralization and wealth concentration rose from the earliest years via a richer-get-richer mechanism.
MeritRank: Sybil Tolerant Reputation
Proposes MeritRank, a decentralized Sybil-resistant reputation system using transitivity, connectivity, and epoch decay, validated on MakerDAO participation data.
Multilayer Network Science: Theory, Methods, and Applications
Review of multilayer-network theory, methods (community detection, dynamical processes, temporal and higher-order structures, machine learning), and applications across infrastructure, epidemiology, social science, and finance.
Multilayer Networks in a Nutshell
Overview of multilayer-network methodology and the dynamical processes and applications that single-layer network models cannot capture.
Multiplex Dependency Structure of Financial Markets
Builds a four-layer multiplex network of financial markets (linear, non-linear, tail, and partial correlations); multiplex-specific structural changes track periods of financial stress that are invisible in any single layer.
Network Embedding for AML
Applies node2vec embeddings plus a new 'spread number' metric to over a million anonymized bank accounts to surface anti-central nodes — structurally significant accounts organized to evade detection.
On Modeling Blockchain-Enabled Economic Networks as Stochastic Dynamical Systems
Establishes a dynamical-systems framework for modeling blockchain token economies, illustrated with a miner/platform/user model tested under two block-reward strategies.
Phase Transitions via Ising Model
Monte Carlo simulation of the Ising model against S&P 500 stylized facts (volatility clustering, negative skewness, heavy tails, return-autocorrelation structure); reproduces most of the observed features.
Predicting Financial Market Crises using Multilayer Network + LSTM
Combines Granger causality, Random Forest interlayer spillover features, and LSTM forecasting on multilayer network features; reports improved market-crisis-prediction accuracy over traditional models.
Reputation in Blockchain Consensus: Mathematical Model
Proposes a reputation-based consensus mechanism that forms a validating group from the highest-reputation nodes to guard against malicious behavior in open peer-to-peer blockchain networks.
Reputation System for Artificial Societies
Proposes a reputation-based consensus mechanism for decentralized AI-agent networks intended to be more resistant to reputation gaming than existing systems.
Reputation-Weighted Governance in Decentralized Systems
Proposes an approval-voting-inspired mechanism for expert-weighted update selection in decentralized governance; proves it admits approximate pure Nash equilibria, including under repeated reweighting.
Shifting Crypto Influence High-Resolution
WITHDRAWN by the authors (arXiv, Jan 2024) due to a code implementation error affecting the results — not treated as an established finding.
Sybil Attack Resistant Voting Mechanism
Uses Bayesian mechanism design (wealth-deposit voting with outcome-based payouts) to construct a voting mechanism that is simultaneously Sybil-resistant and efficient.
Designing a Token Economy: Incentives, Governance, and Tokenomics
Presents TEDM, the Token Economy Design Method — stepwise design propositions spanning incentives, governance, and tokenomics — synthesized from prior literature and validated against the Currynomics real-estate-backed-stablecoin case study.
Temporal GNN for Financial Anomaly Detection
Benchmarks Temporal Graph Networks against static GNN and hypergraph baselines on the DGraph financial-fraud dataset; TGN significantly outperforms on AUC.
The Multiplex Nature of Global Financial Contagions
Builds a multiplex network of cross-country financial exposures by asset type; a multiplex contagion model yields up to twice as many systemically important countries as an aggregated single-layer network.
Time-Weighted Snapshot Framework
Examines flash-loan-enabled manipulation of token-weighted DAO voting and proposes a time-weighted snapshot defense mechanism, with applied case studies.
Token Economics in Real-Life: Insolar Network
Case study using systems modeling and simulation to design Insolar's token-incentive mechanisms; finds developer subsidy pools had a positive effect on network adoption.
Weighted Voting on the Blockchain
Proposes weighted-voting validator committees, updated by a multiplicative-weights algorithm, to make Proof-of-Stake consensus more robust to validator abstention.
When Crypto Economics Meet Graph Analytics and Learning
Applies graph analytics and graph transfer learning — including a 'graph of graphs' method — across thousands of cryptocurrency projects to study network effects, decentralization, tokenomics, and fraud detection.