Blockchain-based technologies and the Web3 paradigm are becoming increasingly important in today’s world due to their decentralized, secure, and transparent nature and their potential to enable new forms of innovation and collaboration. Blockchain-based technologies are enabling new applications and use cases, such as decentralized finance (DeFi), non-fungible tokens (NFT), decentralized identity management, and decentralized autonomous organization (DAO), and they are playing a key role in the development of the metaverse. In the landscape of blockchain-based solutions, and even in the design of protocols and architectures of new blockchains, machine learning and data mining are becoming increasingly important to cope with issues such as preserving and increasing the security of blockchains networks, or optimizing the performance of smart contracts - building blocks of the most innovative applications. Blockchain-based technologies are appealing to the data mining and machine learning community for a further key aspect: data. Since they support transparency, immutability and validation by design, blockchains represent invaluable data sources for different application domains such as financial time series forecasting, mining and learning on large-scale graphs, malicious and fraudulent behavior in the transaction and relational data. In all contexts, researchers may access large collections of temporal and heterogeneous data, which capture different aspects of the interactions among the elements defining blockchain-based applications and platforms. Due to the richness, quality and high temporal resolution of blockchain data, different research areas of machine learning and data mining may benefit from data coming from blockchain networks and may impact the evolution of future blockchain-based solutions; to name a few, graph learning and mining on temporal, heterogeneous and feature-rich large scale graphs can massively exploit data from blockchains, as well as the communities on continual learning or reinforcement learning since a variety of data are continuously ingested into blockchains, asking for adaptive solutions for optimizing the efficiency of the network.

The LIMBO workshop aims at gathering researchers in different areas of machine learning and data mining to share their latest research on issues related to blockchain-based systems and/or based on data from blockchains, as well as experts and researchers in blockchain and blockchain-based technologies. Given its multidisciplinary nature, LIMBO will serve as a fertile ground where to discuss a wide variety of approaches, methods and issues related to different aspects of the blockchain landscape: cryptocurrencies, NFTs, token economics, DeFi, DAO, smart contracts, social media, security

and optimization. For these reasons, we expect submissions from experts in machine learning and data mining, blockchain and security, network science, and computational social science. The end result will be not only an overview of the state of the art in this growing field but also the construction of a seminal community and a common framework to help define its future development.

Topics for the workshops include, but are not limited to:

• Decentralized federated learning on blockchains 

• Anomaly detection in blockchains 

• Graph neural networks for blockchain data

•  Continual learning for blockchain data

• Machine learning and data mining for de-anonymization in blockchain

•  Address clustering in blockchain

•  Machine learning and data mining for identifying blockchain manipulation – 

• Machine learning for fraud detection in blockchains

•  Machine learning and data mining for wash trading and money laundering 

• Graph mining for networked blockchain data

• Graph mining for temporal and/or networked blockchain data

• Forecasting for temporal blockchain data

• Machine learning for IoT platforms on blockchain

• Analytics platforms for blockchain data

• Machine learning for smart contract optimization

• Machine learning and data mining for smart contract analysis

• Identification of security threats by machine learning and data mining

• Identification of malicious behaviors in blockchains

Organizers

• Sabrina Gaito, University of Milan, Milan, Italy - sabrina.gaito@unimi.it

• Roberto Interdonato, CIRAD, UMR Tetis, Montpellier, France - roberto.interdonato@cirad.fr

• Andrea Tagarelli, University of Calabria, Cosenza, Italy - andrea.tagarelli@unical.it

• Matteo Zignani, University of Milan, Milan, Italy - matteo.zignani@unimi.it