Interview with Dr. Piotr Warchoł, Lead Data Scientist at Arteïa
Can you tell us a little bit about yourself and why you have decided to take part in Arteïa’s adventure?
Sure. Until recently I was a physicist studying random matrix theory and its applications to complex systems. As such, I worked in several research institutions in Poland, France and the US. It was around the time I was finishing my Ph.D., in 2014, that I got interested in data science. Since then, I participated in the ‘Science to Data Science’ programme in London, worked as a data science freelancer and advisor, taught machine learning at the Jagiellonian University and incorporated its tools into my research. Finally, this July, I moved from academia into a full-time research and development position at Arteïa.
I decided to get involved because the company is facing a diverse and quite unique set of data related challenges. To give you a sense of the scope, we are using machine vision for feature extraction, segmentation, search and to aid authentication. There is a recommendation system being developed, one tailor made for artworks. Moreover, we will need reasonably advanced statistics to produce artwork price estimates and make predictions of future artwork prices. These are all extremely interesting and the list goes on. The point is also that many of the platform features will be data driven, something that is important for me as a data scientist — it means that I can have an impact within the company.
What do you think are the biggest challenges that Arteïa is or will be facing in terms of data?
The fundamental challenges for companies in this area and in this stage of development are data acquisition and user privacy. As to the former, the problem as well as an opportunity is to design the product so that it incorporates a sustainable data ingestion framework. One which can scale, doesn’t compromise on the quality of the data and benefits from the network effect of the user base. This is both a technical as well as a business challenge. The later, at minimum means GDPR (General Data Protection Regulation) compliance, however for Arteïa and the art world in general, it is more subtle than that. There is a significant segment of players in the art market who treat privacy as a priority. The challenge here is to design solutions that boast important, often machine learning based functionalities but which are developed in such a way that user data is protected. And this can mean, to some extent, not accessible even to Arteïa. On first sight, solving this might seem impossible, but this is how such technologies as blockchain, homomorphic encryption and federated learning come into play. Interestingly, the healthcare industry faces a similar set of problems with personal medical data.
What can we expect from the combination of machine learning / AI and blockchain technology when it comes to platforms such as Arteïa?
In short, an increase in privacy, security and efficiency.
But before I elaborate, let me make a general comment as one has to be careful here. Indeed data is central to both of those technologies. Machine learning uses data to learn how to answer, often difficult, questions in complicated settings, and Blockchain is the technology by which data can be stored (and sometimes processed — via smart contracts) in a decentralised manner, providing security and trust. You have to realise however, that many companies won’t have a valid use case for a real fusion of the two. The scope of blockchain applications is already limited by law. Some ways of utilization of machine learning will soon be. Most importantly, there needs to be a business justification. Thus, I think more often the not, AI and distributed ledgers will act as complementary technologies, solving different, sometimes connected problems a business tackles.
As for Arteïa, we strive to first ask the question of what is the customer need - what will make our clients interaction with art more fulfilling and secure. We start from there. In our case, the distributed ledger technology will increase transparency in the market while preserving a level of privacy via the provenance tracker. Machine learning will be used for example to crunch that data to increase the relevance and trustworthiness of the artworks recommended on Arteïa’s Marketplace.
—
To find out more about Arteïa and keep up to date with the news, please go to our website and sign up for Telegram.
