Open Library

Using blockchain principles for improving AI research and security

  • Kresimir Kalafatic profile
    Kresimir Kalafatic
    20 June 2018
  • 7 comments
    Total votes: 2
Document(s): 
PDF icon Using blockchain principles for improving AI research and security
Author(s): 
Kresimir Kalafatic
Year of publication: 
2018
Tags: 
blockchainAIsecuritylinking research data and code

Distributed ledger (blockchain) is a distributed database whose every written record is signed by private key, every record
insertion and change can be traced back to specific public/private key pair and whose atomic operation is irrecoverably committed after
the databases mathematically prove their validity and distributed consistency. The mathematical principle of blockchains can be used in
improving AI research and security.

Comments

Log in to post comments
Kresimir Kalafatic's picture
Kresimir Kalafatic,
22/06/2018 16:23

Thank you for the comments. You raised some of the interesting questions.

Do you think that blockchain security will be as strong in the post-quantum computer era as it is nowadays?

At the beginnging of 2017 NIST has initiated a process to solicit, evaluate, and standardize one or more quantum-resistant public-key cryptographic algorithm. Here is the link to the 1 round of submissions which was updated few days ago.https://csrc.nist.gov/Projects/Post-Quantum-Cryptography/Round-1-Submiss...

In my published paper I have put a reference 11 to one interesting conceptual design for a quantum blockchain "Quantum Blockchain using entanglement in time".

Security is based on mathematical algorithms, and as technology advances the encryption which couldn't be breached in past, could be breached in the future (SHA1 is already obsolete). For example wallet address is made using hash algorithm of a public key.This enables hiding public key (is hash algorithm is secure), until the first transaction is initiated from the wallet.This hiding additionally preserves for small period the owner in the case of a error in cryptographic algorithm or technical implementation, enabling owner to move the assets to new wallet with different private/public key based on different algorithm and implementation if network allows it.

Few years ago there was a published research which showed that some quantum communication could be eavesdropped without detection. The problem was not in the physics, but in the technology used. The implementation used error code correction and partial eavesdropping was possible (in the amount which was corrected by ECC).  

Blockchain and ethereum was mentioned in the paper because they are best known to the general public. I am glad you mentioned IOTA which uses graph theory and DAG as you mentioned. The hashgraph is also based on graph theory and mathematics is simple to understand, but it is patented.

Total votes: 1
Norbert JASTROCH's picture
Norbert JASTROCH,
22/06/2018 10:59

The interesting question is if and how a concept like this can be applied to manage and track permits of individual persons for the use of their private data by third parties.

Total votes: 0
Vladimiros Peilivanidis's picture
Vladimiros Peil...,
22/06/2018 12:33

There are several 3d generation blockchain platforms that offer secure data storage, distribution, and even monetization of personal data. Take Zenome for example, a platform that uses blockchain technology to store, analyze/test, distribute, set permissions, and/or levels of accessibility, and again monetization of your personal genomic or other medical data, always with your consent and awareness. 

The question is not if blockchain could help us store and distributed data in an encrypted fashion, but whether that concept would last for long enough in order to invest in it and start using it. As I mention below, IOTA Foundation is currently the only DLT-powered organization that offers reliable, fast, and feeless transactions of data and micropayments, and that is why it has been chosen by governmental and banking institutions to "do the job".

Total votes: 1
Norbert JASTROCH's picture
Norbert JASTROCH,
22/06/2018 14:59

From my point of view, that situation calls for European Union funding (Horizon 2020). 

Total votes: 0
Kresimir Kalafatic's picture
Kresimir Kalafatic,
22/06/2018 16:16

Thank you for the comments. You raised some of the interesting questions.

Do you think that blockchain security will be as strong in the post-quantum computer era as it is nowadays?

At the beginnging of 2017 NIST has initiated a process to solicit, evaluate, and standardize one or more quantum-resistant public-key cryptographic algorithm. Here is the link to the 1 round of submissions which was updated few days ago.https://csrc.nist.gov/Projects/Post-Quantum-Cryptography/Round-1-Submiss...

In my published paper I have put a reference 11 to one interesting conceptual design for a quantum blockchain "Quantum Blockchain using entanglement in time".

Security is based on mathematical algorithms, and as technology advances the encryption which couldn't be breached in past, could be breached in the future (SHA1 is already obsolete). For example wallet address is made using hash algorithm of a public key.This enables hiding public key (is hash algorithm is secure), until the first transaction is initiated from the wallet.This hiding additionally preserves for small period the owner in the case of a error in cryptographic algorithm or technical implementation, enabling owner to move the assets to new wallet with different private/public key based on different algorithm and implementation if network allows it.

Few years ago there was a published research which showed that some quantum communication could be eavesdropped without detection. The problem was not in the physics, but in the technology used. The implementation used error code correction and partial eavesdropping was possible (in the amount which was corrected by ECC).  

Blockchain and ethereum was mentioned in the paper because they are best known to the general public. I am glad you mentioned IOTA which uses graph theory and DAG as you mentioned. The hashgraph is also based on graph theory and mathematics is simple to understand, but it is patented.

Total votes: 0
Kresimir Kalafatic's picture
Kresimir Kalafatic,
22/06/2018 16:54

Thank you for the question.The usage of wallets (private/public) keys enables anonymization or pseudo-anonymization of the data required by GDPR. If the data is encrypted with secure algorithm and valid implementation, data is safe for the time the used technology is secure.

One theoretical possibility for managing and tracking permits could be by using different encryption key for every access and some kind of digital watermarking. In large quantities of data, before granting access to the data, some companies add unique ghost elements. If the data is resold, by identifying ghost elements, the source of the data can be tracked.

If you look at Spectre and Meltdown CPU problems, the HW leaks the data. Homeomorphic encryption mentioned in the paper would reduce the usage of unencrypted data.

Some samples of data leakage was described in appendix  A.

Here is one research paper about Cloud data leakage.

http://v.wpi.edu/wp-content/uploads/Papers/Publications/seriously.pdf

Total votes: 0
Vladimiros Peilivanidis's picture
Vladimiros Peil...,
21/06/2018 13:23

Really interesting posit, Mr. Kalafatic. 

Do you think that blockchain security will be as strong in the post-quantum computer era as it is nowadays?

Discovering a block might be "hard" if you're not ready to throw a couple million dollars in mining operations (when applicable - as many DLTs use pre minted tokens), but what happens when we introduce quantum computing to the mainstream public? 

sha256 and similar protocols have no chance of defending a network against a quantum attack. 

Correct me if I am wrong, but I am almost certain than Serguei Popov's "Tangle" is the only DLT network that is also quantum-proof as the organization behind the Tangle, the German non-profit IOTA Foundation, is also behind JINN microprocessors, currently working on a series of quantum computers for the DLT scene.

In opposition to traditional blockchain platforms, the Tangle uses Directed Acyclic Graph as its backbone, or simply DAG which allows the network to possess abilities like infinite scalability (an other aspect that no other DLT possesses at the moment), quantum-proof security, and literally 0 transaction fees, as the network is not relying on blockchain and "miners".

The Dutch government among many other industrial, banking, academic, and governmental institutions are already using the Tangle to distribute public documents regarding social housing on an encrypted fashion with lighting speed, reliability, and integrity. 

Other use-cases on the Tangle include Taiwan's move to upgrade Taipei city into a modern smart city, with the help of BiiLab and IOTA Foundation. Julie Maupin from G20, who is also a member of IOTA Foundation, was recently visiting Taipei and showcased how the Tangle could help the city to achieve its goals. 

Deutsche Bank stated on an official paper that IOTA is the future of the economy and explains in details the reasons why. Major firms like Volkswagen, Porsche, BOSCH, to mention a few have already invested both financially and strategically to IOTA, with the foundation's board including key members of each respective organization to help them promote what we refer to as the "future" with the most efficient way possible. 

Blockchain technology was and still is very important, but I think that it lacks the technological and scientific background, other sophisticated networks can already provide. 

Nevertheless, I am looking forward to seeing how blockchain and DLT, in general, would support AI platforms and systems to store, distribute and even monetize data, without human interference.  

Total votes: 1