Blog 5: 80% done, still lot to do!

There are two BIG transitions happening at the moment: Energy Transition and Artificial Intelligence. The development, utilisation, growth and market potential of these are incredible. However, depending from which side you, dear reader, are coming from, it might be difficult to fully understand what is happening on the other side.

AI related topics are full of vocabulary, which are close to gibberish to many people from renewable industry. However, the tools and solutions that AI can provide in order to accelerate Energy Transition are currently sometimes even beyond our imagination.

Good-Bad Artificial Intelligence

If you have watched many SciFi movies, you might be scared of AI. To be fair, some application are actually scary. Just open Netflix and check out The Social Dilemma and you might realise some nasty truths (“If you are not paying for the product, YOU are the product.”). However, AI can also outperform people in tasks that are not interesting for humans to do. For repetitive tasks that include a lot of information and analytically objective work, machines are optimal to use. They do it faster, cheaper and just all together better.

Data – BIG Data

In order to categorise research papers and other documents related to Renewable Energy, we first need to give a lot of examples for machine learning. For Proof-of-Concept we have 3 different categories: STEER aspects, value-chain position and geographical location. These 3 categories have total of 15 subcategories in which the papers are divided to. It’s a hell of a job, but we are doing it! One paper at a time, in sets of 25.

So far eWEning star has collected around 700 papers on Wind energy alone. Each of these papers are manually assessed by two different people. The results are compared and combined to one result. Currently about 80% of the job is done and the work continues.

Naturally, the manual categorisation is only for the machine learning. The real job is to feed thousands and thousand of documents to the AI that then categorises them according to the examples. By doing so, we are able to provide the user more relevant document results, according to his/her interest.

Collaboration?

Does your company, research organisation or university have a lot of scientific documents about wind energy? Contact us so that we can test the AI with uncategorised papers next year.

Are you a student in the University of Groningen and got interested in the development eWEning star is doing? Get in contact and let’s see if you can make your thesis related to Data Science, Back-End Development, Front-End Development or other “Wild Card” that comes to your mind.

PS.

If you got scared of AI, check out TedX talk of Prof. Malvina Nissim, where she makes great (positive!) points related to AI.

Leave a Reply