To kick off our interview series with each of our team members, we are glad to present the first interview with Dr. Anton Kolonin who has taken the lead in AI research and development at Autonio. During the last week members of our DAO and our community were asked to send questions about artificial intelligence (AI) and its role for Autonios development.
We are happy that Anton took his time to give us his answers. Enjoy!
Good day, Anton!
First of all, we are very glad to see you in our team and we hope that your experience and skills will not only benefit the Autonio project, its community and the ecosystem as a whole, but also hope it will open up completely new, hitherto unseen ideas and directions that will be useful for the entire crypto industry.
But let’s start from the beginning.
- Can you, please, tell us about yourself, namely, how you first became interested in artificial intelligence, what problems you wanted to solve with it (and still want to)?
In 1981 when in school, me and my friend wrote chat-bot using ancient Hewlett-Packard terminals with storage based on paper tapes with punched holes and printing input/output systems. Honestly, the level of intelligence of that chatbot 39 years ago was quite comparable to modern chatbots available in public.
Since then, I was occasionally involved in various AI projects including images, video and audio recognition, time series prediction, text mining and natural language processing. Brief description may be found here.
At this point, I am primarily focused on the development of Artificial General Intelligence also targeting democratization of AI use. In particular, I am working on my own open-source project called Aigents which is expected to provide intelligent personal assistants for people who rely on online information in their everyday lives and businesses.
At the same time, I am actively contributing to SingularityNET and Autonio projects. The former provides a decentralized and democratic way of using AI across applications and developers while the latter brings this to the financial ecosystem.
- How do you think the crypto / blockchain space will evolve with the use of AI?
I have a somewhat complex relationship to this idea. I have written some concerns about this last year in a separate post. On one hand, having everything, starting with personal records and ending with personal savings, stored in democrating decentralized way using peer-to-peer networks and distributed ledgers, such as blockchains is more fair and more protected from totalitarian actions from governments and global corporations, compared to centralized processing of the same information by conventional social networks and banks. On the other hand, such democratic way of doing things is much less efficient in terms of resource consumptions and brings other risks due to complexity of resistance to decentralized crime.
In short, there are two trends continuously fighting with each other - “decentralization” and “centralization” while both of them have pros and cons. Being a fan of the former, I clearly see the benefits of the latter, so advancing the decentralized technologies need more investments in performance and security. That is the reason why me and my colleagues at SingularityNET are pushing the concept of “liquid reputation” to the industry - here is one of my latest writings on the matter.
- Thank you for such a detailed introduction. Now let's get down to more project-specific questions. When did you first hear about Autonio and why you got interested in what Autonio is doing?
The idea of collaborating with Autonio has been brought up by Ben Goertzel who I work with on SingularityNET project - in January this year. It seemed like a good chance to advance the use of AI for decentralization in the financial domain.
- What global tasks do you, as an Autonio AI researcher, set for the next 3-5 years?
I am a rather agile person, so having long-term goals is important but solving short-term tasks is rather critical :-)
For 3-5 years I would say the ultimate goal is to make the world population less financially dependent from financial pyramids introduced by world fiat currencies. At the same time, it also means protection of the same popluation from scams introduced by some cryptocurrencies.
On the practical side, I am focusing on building AI-powered instruments for better organization of individual investment portfolios in crypto markets, we are at the very beginning of the R&D phase so it is hard to promise something firmly. Still, releasing initial versions of “AI oracles” suggesting the portfolio structures, trading strategies and trading signals during the current year is a primary objective.
- Can you tell us some details, what you are working on right now for Autonio and what can we expect from AI in the nearest future?
Specifically, we are doing many things at one time for this. First, we are searching ways to simulate the existing markets “in vitro'' so we can develop optimal trading strategies “in the lab” before going live on real markets “in vivo”. Second, we are exploring which kinds of technical and fundamental data can be obtained to provide more reliable AI-based predictions and recommendations, like trying to use non only the OHLCV data, but also use historical trades and order books. Third, we are exploring with AI/ML techniques that are more reliable and efficient to provide predictions based on that data.
- Can you tell us about AI agents - what vision you have for it and how does it align with Autonio?
There are two kinds of AI agents that you may be asking for, and each of the two are highly complementary to each other and the Autonio :-)
First, the concept of SingularityNET project implies the networks of AI agents serving each other via the SingularityNET platform and so the Autonio “AI Oracles” may be one of such agents.
Second, my own Aigents are the AI agents being representatives of individual users on the Internet, including existing social networks and blockchains as well as SingularityNET. That means that such “personal AI agent” might get services from any AI agent on SingularityNET platform, including Autonio AI Oracle.
Practically, the short term goal is to reuse the Aigents capabilities for sentiment analysis in news feeds and social media channels for Autonio AI Oracle so it can use sentiment in the news feeds and social media as extra input for its predictions and recommendations.
- Do you think we'll reach a point where our AI agents will be smart enough to analyze the markets in real time and adjust strategies accordingly (for example, respond quickly to volatile swings in the markets with BTC that will trickle down and affect altcoins)?
I think we are already there in conventional finance where the majority of the trading is algorithmic and we are getting to that point in crypto-trading as well.
Given my knowledge in the domain, there is no algorithmic way to foresee “black swans” or resist human traders possessing insider information. Also, the speed of response in many cases is more based on technical (physical) constraints imposed by the exchange than on the ability of an algorithm to suggest this response. I mean, no matter how smart your algorithm is, it has no value in case a less smarter competitor can just throttle your transaction, right?
In particular, given the volatility and highly manipulative nature of the crypto market I am more optimistic in regard to having “winning strategies” operating on a market long term rather than making occasional profits on handling unexpected ups and downs.
- You are also a part of SingularityNET project. For those who don’t know, SingularityNET is a blockchain-based platform that allows to create and even monetize AI services. How can Autonio and SingularityNET help each other to build a strong foundation & ecosystem?
Sure and I think I have mentioned that above.
SingularityNET is a generic purpose platform for AI democratization, Autonio is a platform for financial democratization employing and deploying its services on SingularityNET. And the Aigents may be a connector of all this to the consumer users and provider of the sentiment from the online and social media. That’s the plan!
Well, that was an amazing, thank you for your time! We are confident that our community will love this interview and will follow Autonio AI development with even greater attention.
The next interview with another team member is coming soon! Once our DAO members voted on who they want to get interviewed next (via fair voting), all our community will have an opportunity to send us the questions.
Stay tuned and never miss any updates!