Back to Blogs
DLT WILL ENHANCE ARTIFICIAL INTELLIGENCE, MACHINE LEARNING & RESPONSIBLE INNOVATION
By Gowthaman Ragothaman, CEO of Aqilliz
Published on February 13, 2019
We are going to live in a world, where increasingly, there is going to be more information outside than it is possible to “hoard” them inside. While the industry is running in different directions to invest and build separate capabilities (thrice over!) across AI, ML and Innovation, perhaps it is a better idea to take a pause, connect the dots and prudently invest in one common Distributed Ledger Technology, one that can efficiently address all these three challenges as a baseline.

Artificial Intelligence
Artificial Intelligence (AI) is all about the ability to correctly interpret external data, to learn from such data, and to use those learnings to achieve specific goals and tasks through flexible adaptation. Increasingly, within the gardens of Search, Social, Commerce and amongst the Internet-of-Things, while there is widespread attention and optimism on the application of AI, often, there are now, worries about the “correct interpretation” of data. AI has two components to it – one is an algorithm (that helps in automating repeatable procedures) and the other is the inference (where information is applied with knowledge). While the industry has excelled in building algorithms, we are still struggling to get the inference right. There is a need to “clean” the information from the “chaos”, remove junk and get to the core of it. This process of curation requires some form of human intervention for AI to be effective. This also requires some form of decentralised control to ensure that AI does not go “off-hand” by giving us a Skynet, where humans fall into the trap of self-fulfilling prophecy of wrong interpretations.

Machine Learning
Machine Learning (ML) is all about the scientific study of algorithms and statistical models used by computer systems to effectively perform a specific task without using explicit instructions, relying on models and inference instead. Some of the common challenges with increased adoption of ML is that it is becoming so “data hungry” that the rising computing power that is required by the computer systems is exponentially increasing, not to ignore the related talent deficit that comes with it while building the models. ML again has two components to it – one is the algorithm (like what an AI will require, that helps in automating repeatable procedures) and the other is the technology (skills, methods or processes required for development). Again, while the industry has excelled itself in building algorithms, we are still struggling to match the computing requirements both in terms of speed and talent. There is a need to leverage the existing infrastructure to provide “network effects”, essentially to cast the net far and wide and optimise existing capabilities across the world, that can address this challenge. We have pockets of computers working overtime while the others that do not. In essence the internet as a peer-to-peer network can be leveraged lot better for ML to become efficient.
And there is another interesting angle to this. Best innovations did not come from the algorithms. It is the spirit of human intelligence, the inherent ability to perceive or infer information and to retain it as knowledge, apply it within an environment or context that has led to all the innovations that we see today, including the gardens of Search, Social and Commerce and the Internet-of-things. Human intelligence also needed a collection of techniques, skills, methods and processes to make it scalable for large quantity production or development of goods, services, tools or platforms. Innovation therefore has two components to it – one is the spirit of human intelligence and other is the technology that can provide the tools to scale it up. The spirit of entrepreneurship has driven innovations, all across the world, from the Silicon Valley and similar such environments, largely funded by the private sector and fuelled by venture capitalists with the hope to become the next millionaire in the shortest time. However, this has opened up lots of concerns and questions about “profit” vs “security” and if I have to be specific about “security”, I would drill it down to the most important need for humanity today i.e., privacy. This “Privacy” vs “Profit” debate is just the tip of the iceberg of the larger question about unchecked innovations.
These developments are NOT disconnected. In fact, this is a closely knit ecosystem where human intelligence is mixed with algorithms to create artificial intelligence (AI), algorithms are mixed with technology to form machine learning (ML) and technology is mixed with intelligence to enable scalable innovations.
Curated inference makes AI effective; Scaled up computing makes ML efficient and Responsible innovation balances privacy with profit.
  • Curating inferences requires incentives (mining with smart contracts) for humans to participate
  • Scaled up computing needs a more connected peer-to-peer network (network)
  • Responsible innovation will come from state level security (consensus protocol)
So if we have a peer-to-peer network that has a state level security, incentivised by humans to curate, we should be in a better position to address all three challenges.
Distributed Ledger Technology is one such protocol that has the potential to offer such an integrated solution and Industry level consortiums will become a necessity to make this shift.
It is definitely a better idea to take a pause, connect the dots and prudently invest in one common Distributed Ledger Technology, one that can help scale up Artificial Intelligence, Machine Learning and Responsible Innovation, all three, at the same time. After all there is only a finite amount of money that is available, in any case!