"blockchains for ai", trent mcconaghy, ai researcher, blockchain engineer. founder &...
TRANSCRIPT
![Page 1: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/1.jpg)
Blockchains for Artificial Intelligence
Trent McConaghy
@trentmc0
![Page 2: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/2.jpg)
Blue Ocean DatabasesRelational DBs: Oracle – 80s
Website DBs: MySQL – early 00sDist’d/NoSQL DBs: MongoDB -- late 00s
![Page 3: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/3.jpg)
How do distributed (“big data”) databases scale? Answer: Distribute storage across many machines, i.e. shardingAnd, a consensus algorithm keeps distributed nodes in sync.
0 50 100 150 200 250 300 350
0
200,000
400,000
600,000
800,000
1,200,000
175,000
367,000
537,000
1,100,000
Nodes
Writes/s
![Page 4: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/4.jpg)
![Page 5: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/5.jpg)
Blockchains are databases with“blue ocean” benefits
Decentralized / shared controlImmutability / audit trail
Tokens / exchanges
![Page 6: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/6.jpg)
How to build a scalable blockchain database (BigchainDB)1. Start with an enterprise-grade distributed DB, e.g. MongoDB2. Engineer in blockchain characteristics
• Each DB node is a federation nodeDecentralized / Shared Control
• Hash Previous Blocks• Append-only
Immutable / Audit Trails
• “Own” = have private key• Asset lives on the database
Native assets
![Page 7: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/7.jpg)
IPDB = a public global blockchain database
![Page 8: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/8.jpg)
E-GOLD/CASHBitcoin, zcash
FILE SYSTEMe.g. S3, HDFS, WWW,
IPFS
PROCESSINGe.g. EC2, Ethereum, Hyperledger,
Tendermint, Lisk
DATABASEe.g. MySQL, MongoDB
BigchainDB + IPDB
The Emerging Decentralized Stack
![Page 9: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/9.jpg)
Example real-world use: ascribe
![Page 10: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/10.jpg)
Artist creates piece
![Page 11: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/11.jpg)
Artist goes to ascribe
![Page 12: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/12.jpg)
Artist registers piece
![Page 13: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/13.jpg)
Unique ID for piece
![Page 14: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/14.jpg)
Piece for sale on marketplace
![Page 15: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/15.jpg)
Artist transfers to new owner
![Page 16: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/16.jpg)
Provenance of old -> new owners
![Page 17: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/17.jpg)
New owner gets a COA
![Page 18: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/18.jpg)
How can blockchains help AI?
![Page 19: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/19.jpg)
Work off of each of the benefits…Decentralized / shared control
Immutability / audit trailTokens / exchanges
![Page 20: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/20.jpg)
Decentralized / shared control encourages data sharingMore data better models
-10.3 + 7.08e-5 / id1 + 1.87 * ln( -1.95e+9 + 1.00e+10
/ (vsg1*vsg3) + 1.42e+9 *(vds2*vsd5) / (vsg1*vgs2*vsg5*id2) )
10^( 5.68 - 0.03 * vsg1 / vds2 -55.43 * id1+ 5.63e-6 / id1 )
90.5 + 190.6 * id1 / vsg1 + 22.2 * id2 / vds2
2.36e+7 + 1.95e+4 * id2 / id1 - 104.69 / id2 + 2.15e+9 * id2 + 4.63e+8 * id1
- 5.72e+7 -2.50e+11 * (id1*id2) / vgs2 + 5.53e+6 * vds2 / vgs2 + 109.72 /
id1
Merge
Build models Build model
Low accuracyHigh accuracy
![Page 21: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/21.jpg)
Decentralized / shared control encourages data sharing Qualitatively new ecosystem-level data qualitatively new models
Example: shared diamond certification houses data makes fraud id possible
Merge
All the diamond cert houses
All
dia
mo
nd
s fo
r ce
rt h
ou
se 1
Ce
rt h
ou
se 2
Ce
rt h
ou
se 3
Ce
rt h
ou
se 4
All the legit diamonds
Build 1-class classifier
Legit
FraudulentNo single cert
house has enough data to make an
accurate classifier
Build classifiers
![Page 22: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/22.jpg)
Decentralized / shared control encourages data sharing Qualitatively new planet-level data qualitatively new models
“IPDB is kibbles for AI”
--David Holtzman
![Page 23: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/23.jpg)
Immutability for An Audit Trail on Training/Testing Data & ModelsFor greater trustworthiness of the data & models (Avoid garbage-in, garbage-out)
Provenance in building models:
• Sensor / input stream data
• Training X/y data
• Model building convergence
Provenance in testing / in the field:
• Testing X data
• Model simulation
• Testing yhat data Tim
e-s
tam
p /
sto
re
Applications:• you can tell if a sensor is lying• you know the “story” of a model• catch leaks in the data chain
![Page 24: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/24.jpg)
Another Opportunity:A shared global registry of training data & models
All the Kaggle datasets
All the Kaggle models
All the ImageNet datasets
All the ImageNet models
….…
.“Models are owned
by the planet”
![Page 25: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/25.jpg)
Training/testing data & models as intellectual property assets Decentralized data & model exchanges
Your datasets or models…
…licensed to others
Others’ datasets & models
…licensed to you
….…
.“EMX – European
Model Exchange?”
![Page 26: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/26.jpg)
Sell your CARTS?
![Page 27: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/27.jpg)
AI DAOs by Example: The ArtDAOAlgorithm, running on decentralized compute substrate…
1. Run AI art engine to generate new image, using GP or deep
2. Claim attribution in blockchain, using ascribe
3. Create multiple editions, using ascribe
4. Post editions for sale onto a marketplace, using Getty (centralized), or OpenBazaar (decent.)
5. Sell the editions. $ goes to ArtDAO using built-in cryptocurrency like Ether. IP go from ArtDAO using ascribe.
6. Repeat!
![Page 28: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/28.jpg)
AI DAOs by Example: The ArtDAOAlgorithm…
1. Run AI art engine to generate new image, using GP or deep
2. Claim attribution in blockchain, using ascribe
3. Create multiple editions, using ascribe
4. Post editions for sale onto a marketplace, using Getty (centralized), or OpenBazaar (decent.)
5. Sell the editions. $ goes to ArtDAO using built-in cryptocurrency like Ether. IP go from ArtDAO using ascribe.
6. Repeat!
Over time, if ArtDAO makes more money from sales
than from generating new art, then
it will accumulate wealth. And, you can’t turn it off.
![Page 29: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/29.jpg)
Blockchains for Artificial Intelligence
A planetary-scale blockchain database (IPDB) unlocks opportunities:1. Data sharing Better models2. Data sharing Qualitatively new models3. Audit trails on data & models for more trustworthy predictions4. Shared global registry of training data & models5. Data & models as IP assets data & model exchange6. AI DAOs – AI that can accumulate wealth, that you can’t turn off
Trent McConaghy@trentmc0
bigchaindb.comipdb.foundation
![Page 30: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/30.jpg)
AppendixDetails of AI
DAOs
![Page 31: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/31.jpg)
Then you get decentralized processing.
aka “smart contracts”
What if you used a blockchain to store state of a state machine?
State
Virtual machine
![Page 32: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/32.jpg)
Then you get decentralized processing.
And you can build a
world computer
having decentralized processing, storage, and communications
(e.g. Ethereum vision)
What if you used a blockchain to store state of a state machine?
State
Virtual machine
Decentralized applications (dapps)
World computer
![Page 33: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/33.jpg)
DAO: a computational process that
• runs autonomously,
• on decentralized infrastructure,
• with resource manipulation.
It’s code that can own stuff!
DAO: Decentralized Autonomous Organization
State
Virtual machine
DAO Dapp
![Page 34: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/34.jpg)
AI entity is a feedback control system. That is, AGI.
Its feedback loop would continue on its own, taking inputs, updating its state, and actuating outputs, with the resources to do so continually.
AGI on a DAO?
AI DAO
World computer
![Page 35: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/35.jpg)
Example: The ArtDAOAlgorithm…
1. Run AI art engine to generate new image, using GP or deep
2. Claim attribution in blockchain, using ascribe
3. Create multiple editions, using ascribe
4. Post editions for sale onto a marketplace, using Getty (centralized), or OpenBazaar (decent.)
5. Sell the editions. $ goes to ArtDAO using built-in cryptocurrency like Ether. IP go from ArtDAO using ascribe.
6. Repeat! Create more art, sell it, get wealthier
![Page 36: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/36.jpg)
Example: The ArtDAOAlgorithm…
1. Run AI art engine to generate new image, using GP or deep
2. Claim attribution in blockchain, using ascribe
3. Create multiple editions, using ascribe
4. Post editions for sale onto a marketplace, using Getty (centralized), or OpenBazaar (decent.)
5. Sell the editions. $ goes to ArtDAO using built-in cryptocurrency like Ether. IP go from ArtDAO using ascribe.
6. Repeat! Create more art, sell it, get wealthier
Over time, if ArtDAO makes more money from sales
than from generating new art, then
it will accumulate wealth. And, you can’t turn it off.
![Page 37: "Blockchains for AI", Trent McConaghy, AI researcher, blockchain engineer. Founder & CTO of BigchainDB | IPDB | ascribe | Solido](https://reader034.vdocuments.net/reader034/viewer/2022051709/586fc0d11a28aba24c8b4853/html5/thumbnails/37.jpg)
Angles to Making AI DAOs
• DAO AI DAO. Start with DAO, add AI. E.g. Plantoid
• AI AI DAO. Start with AI, add DAO. E.g. numer.ai
• SaaS DAO AI DAO. Convert SaaS to DAO. Then add AI
• Physical service AI DAO. E.g. Uber self-owning cars