Machine learning (including deep and reinforcement learning) and blockchain are two of the most noticeable technologies in recent years. The first one is the foundation of artificial intelligence and big data, and the second one has significantly disrupted the financial industry. Both technologies are data-driven, and thus there are rapidly growing interests in integrating them for more secure and efficient data sharing and analysis. In this paper, we review the research on combining blockchain and machine learning technologies and demonstrate that they can collaborate efficiently and effectively. In the end, we point out some future directions and expect more researches on deeper integration of the two promising technologies.

1. Machine Learning in/for
Blockchain: Future and Challenges
*Fang Chen, ***Hong Wan, *Hui Cai, and **Guang Cheng,
*School of Industrial Engineering, Purdue University
**Department of Statistics, Purdue University
***Edward P. Fitts Department of Industrial and Systems Engineering
Oct 2019

2. In this talk:
• Incorporating machine learning into blockchain
algorithms or frameworks
• Analyzing attributes or applications of blockchain
using machine learning methods
• Blockchain-based learning systems
• Blockchain beyond cryptocurrency
• Design and characterize the blockchain systems

4. Learning as a Prediction and
Analysis Tool
https://www.google.com/url?sa=i&source=images&cd=&ved=2ahUKEwj
G4oqbp4nlAhUCQq0KHS3-
CiYQjRx6BAgBEAQ&url=https%3A%2F%2Fcointelegraph.com%2Fnews%2
Fbtc-price-rebounds-from-7-700-crypto-traders-hope-for-relief-
rally&psig=AOvVaw2Nx-Z1btIKD3iRAf5Rrmxp&ust=1570508630461698
https://blog.emsisoft.com/wp-content/uploads/2019/09/ransomware-bitcoin.png

5. Transaction Categorization
• Supervised learning to detect cybercriminal
activities
• Categorizing all transactions into: Exchange,
Service, Gambling, and Mining Pool.
• Topology-based methods to detect abnormal
behaviors

6. Cryptocurrency Price Prediction
• Topology based methods: chainlet
• Bayesian regression for “latent source model"
• Cryptocurrency price prediction and digital portfolio
management using Recurrent Neural Network
(RNN) and Long-Short Term Memory
• Reinforcement-learning based portfolio optimization

7. Computational Resource Allocation
• Deep-learning based auction algorithm for edge
computing resources allocation to support mobile
mining activities.
• Mobile device miners will submit their bid valuation
profiles to one Edge Computing Service Provider
(ECSP) for buying additional computing power.
• ECSP want to maximize its revenue.

8. Blockchain for Machine Learning
• Decentralized privacy-preserving and secure
machine learning system without a trusted central
server.
• IOT sensors with reinforcement learning abilities
share data through a blockchain network.
• Is public chain the right choice?
Video Simulation

9. Futures and Challenges
Machine learning to help design better blockchain
systems; and blockchain to facilitate data sharing and
verification

10. Blockchain Design Questions
• Who can access data: Private vs. Public
• Who can validate data/add block: Permissioned
vs Permissionless
• Consensus to be used: Trade-off among security
and efficiency.
https://www.google.com/url?sa=i&rct=j&q=&esrc=s&source=images&cd=&ved=2ahUKEwinjN2s7_DkAhXlmeAKHXxhAIUQjRx6BAgBEAQ&url=ht
tps%3A%2F%2F101blockchains.com%2Fconsensus-algorithms-blockchain%2F&psig=AOvVaw23pKh4qS8W_xgyajJ3aFl9&ust=1569669093339830

11. Challenges for Each Type
• Permissionless-public chain: scalability and privacy
• Permissioned-public chain: centralization and
privacy
• Permissionless-private chain: scalability and
consensus
• Permissioned-private chain: centralization and
consensus
https://medium.com/ltonetwork/the-rise-of-private-permissionless-blockchains-
part-1-4c39bea2e2be

12. Bad News First
• “Private blockchains are completely uninteresting…
-- the only reason to operate one is to ride on the
blockchain hype…” Bruce Schneier
https://www.schneier.com/blog/archives/2019/02/blockchain_and_.html
http://shorturl.at/ahsRU, shorturl.at/gETV2
Tonight we will talk about
cryptocurrencies… .everything
you don’t understand money
combined by everything you
don’t understand about
computers….
Cryptocurrencies: Last Week
Tonight with John Oliver (HBO)

13. https://www.google.com/url?sa=i&rct=j&q=&esrc=s&source=images&cd=&ved=2ahUKEwj-
p72L7vDkAhVjQt8KHU18CjsQjRx6BAgBEAQ&url=https%3A%2F%2Fwww.wbur.org%2Fonpoint%2F2017%2F11%2F20%2Fwho-can-you-
trust&psig=AOvVaw2WVera9lB95ujTubez6lPJ&ust=1569668688845079
https://d279m997dpfwgl.cloudfront.net/wp/2017/11/Trust-cropped.jpg

14. Blockchain for Bitcoin and Business
• Assets and records vs. Coins
• Identity vs anonymity: different trust level
• Data privacy vs open ledger
• Efficiency vs security
https://ibm.co/2xYSoGy

15. How Machine Learning can Help?
• Deciding the system dynamics and behaviors
Game theory modeling
Input modeling of various simulation models
utility function of smart agents
Stress test

16. How Blockchain can Help
• Connecting multiple blockchains.
• Data fusion
• Multi-level network: Centralization in some levels
and decentralized in other levels.

17. Journal of Blockchain Research
• System design and analysis
• Distributed data storage
• Security and privacy
• Trust and certifications
• Mathematics of blockchain
• Data science of blockchain

18. Journal of Blockchain Research
• Economics, finance and social science of
blockchain
• Manufacturing and blockchain
• Cognitive and behavioral science of blockchain
• Supply chain and service operations with
blockchain
• Case studies of blockchain
• Reviews and Discussions of blockchain
technology’s challenges, criticism and futures.

19. Discussions

20. https://medium.com/ltonetwork/the-rise-of-private-permissionless-blockchains-part-1-4c39bea2e2be

21. Blockchain vs Shared Database
• Private chain is a shared database with one or
multiple parties with administrative power.
• Private chain still make use of cryptography and
data structures (like Merkle trees) to ensure non-
valid transactions aren’t added to the blockchain
and data are immutable
• Higher levels of error checking and transaction
validity than regular shared databases.

22. Blockchain As a System of Record
• Digital identity: all four
kinds.
• Tokenization: private or
public, permissioned
• Inter-organizational data
management:
permissioned, private
• Usage:
government, financial
institutes, auditing...
https://www.coindesk.com/information/applications-use-cases-blockchains
https://i.ytimg.com/vi/1bVczDqh7_A/maxresdefault.jpg

23. Blockchain as A Platform
• Smart contracting
• automated governance
• Market without middleman
• automating regulatory compliance
https://static.coindesk.com/wp-content/uploads/2017/03/Screen-Shot-2017-03-28-at-5.43.08-PM.png
https://storefoundry.s3.amazonaws.com/store_items/images/000/013/039/large/7.jpg?1561546577

24. Ripple Coin

25. Hybrid Chain
• Connecting multiple blockchains.
• Data fusion
• Multi-level network: Centralization in some levels
and decentralized in other levels.