Graph-based Blockchain Transaction Networks Modeling

​Title:

​Graph-based Blockchain Transaction Networks Modeling

​Abstract:

​This presentation offers an overview of emerging methods for extracting insight from low-level blockchain transaction data, where transparency coexists with semantic opacity due to pseudonymity, composability, and the lack of structured supervision. Three interrelated research of on-chain analytics include address classification with limited labeled data, leveraging positive-unlabeled learning and scalable sampling in transaction networks; functional abstraction of DeFi protocols by decomposing execution traces into graph-based primitives; and modeling decentralized user engagement through the reconstruction of interaction graphs from heterogeneous smart contract behaviors.

​Bio:

​Junliang Luo is a PhD candidate at McGill University, with research centers on blockchain transaction graph‑modeling, such as incremental transaction-network modeling, Sybil‑resilient token economics in blockchains, etc. He holds publications in blockchain and DeFi conferences and journals, contributing to scalable transaction analytics, decentralized credential systems, and U.S Treasury RWA token profiling.

​Organizer:

​Decentralized AI

​https://www.linkedin.com/groups/10180003/

​By registering to attend the event, you agree to be recorded during the event.