5 Data takeaways from Etymo, Prowler and Ajit Joakar

Panel @ London AI: Weijan Zhang, CEO of Etymo; Aleksi Tukiainen, Co-founder and Engineer at Prowler and Ajit Jaokar; moderated by Mike Reiner, Co-founder of City AI and Venture Partner at Open Ocean.

DataSeries, in collaboration with the team at London AI, decided to bring together a panel of data scientists, to discuss the importance of data and specifically tackle the following question: “Is your Data ready for AI?”

Key Takeaways

1. Find alternative sources of data. Many early startups don’t have access to data. For Prowler.ai, in the logistics space, they went to have representative data sets that didn’t represent the exact problem that they were hoping to solve but was closely related. Kaggle is another source that can be useful.
2. Knowledge graphs can be useful to interconnect fragmented data. Data scientists often spend their time working on data cleaning. Weijan at Etymo feels that this is not as time efficient as knowledge graphs, that could hugely improve productivity. More importantly, it is important to start with the data points that you really care about and build the picture around it.
3. Be wary of data lakes! The most important question that should be asked before proceeding further is: What is the data for? Why am I collecting the data? And what can I do with the data? Organizations and companies will have to answer these questions, before they can figure out how to make their data useful. Ajit firmly believes that in addition, there is an importance on data labeling.
4. If you want to have the performance of your system at the level which you care about, we need to try and get the data that you need and not just vast amounts of data. For organisations that just have just kicked off, try and aggregate the most relevant data points and then slowly build on that.
5. Try and be multidisciplinary and use AI to solve hard problems. AI is automatic feature discovery, and presumably one is not tell the algorithms what the features are that are relevant to the decision but letting the algorithm filter out, group-up, what those decision are. We should aim to use AI to solve for hard problems (i.e drug discovery) in the long-run, but not discount smaller problems with tangible results in the short- to medium-term.

The panel comprised an impressive lineup, including…

Moderator

• Mike Reiner, Co-founder of City AI and Venture Partner at Open Ocean. Mike invests in innovative startups from Europe and beyond, in several areas and industries, including AI (which he is particularly fond of).

Panelists

• Weijan Zhang, CEO of Etymo, whose work focuses on building lighter knowledge graphs from highly sophisticated enterprise databases. Weijan is a PhD student in Mathematical Sciences (Numerical Analysis). He is interested in Numerical Linear Algebra and Machine Learning and the interplay between these two fields. Currently, he is focusing on problems related to Evolving Graphs. Weijan was also a visiting student (Spring 2016) at the MIT Julia Lab, which is a member of the bigdata@CSAIL MIT Big Data Initiative.
• Aleksi Tukiainen, Co-founder and Engineer at Prowler, the the provider of the world’s first AI decision-making platform, is taking AI beyond pattern recognition. PROWLER.io’s VUKU™ platform combines three distinct branches of mathematics — Probabilistic Modeling, Reinforcement Learning and Multi-Agent Systems. . Aleksi is from Finnish Lapland and moved to the UK with his wife-to-be in 2012. He completed his engineering studies at the University of Cambridge, with a self-learning unicycle project as his thesis. He is one of the co-founders of PROWLER.io and now heads the Release Management team.
• Ajit Jaokar, based in London, Ajit’s work spans research, entrepreneurship and academia relating to Artificial Intelligence (AI) and Internet of Things (IoT). Ajit works as a Data Scientist (Bioinformatics and IoT domains). He is the course director at Oxford University on “Data Science for Internet of Things”. Besides Oxford University, Ajit has also conducted AI courses in LSE, UPM and part of the Harvard Kennedy Future society research on AI. Ajit was recently(Oct 2017) listed in the list of top 30 influencers for IoT for 2017. Ajit publishes extensively on KDnuggets and Data Science Central and his book, Data Science for the Internet of Things, is included as a course book at Stanford University. He was recently included in top 16 influencers (Data Science Central), Top 100 blogs (KDnuggets), Top 50 (IoT central), No 19 among top 50 twitter IoT influencers (IoT Institute).