Euskal Encounter 28 'The Unexpected Data Science; let's have a laugh!'

Under that name more worthy of an Internet clickbait ad than anything else (The Unexpected Data Science; let’s have a laugh!), I gave a talk at Euskal Encounter 28 at the end of July; specifically, on 07/25/2020, at the Urretxu fronton. Here you can find my promotional space on the main website.

Doing a quick research in Wikipedia [ES], for those who do not know the Euskal Encounter, it is the most veteran LAN Party in Spain, which has been celebrated since 1994. In fact, this Euskal Encounter 28 (#EE28), “Revival – Beyond AI “, had as its main topic AI and its impact, in a world that has been affected by COVID-19.

And I have to say that writing these lines is somewhat surprising to me. I had never been to the Euskal Encounter, although it was an event that I always wanted to go to for various reasons. And the fact of being there as a speaker in my first (although different, due to the online format) attendance at the event was a motivation boost for me. And this is even funnier since I was having some good vermouth with a good friend just a week before my participation was announced, who already told me: “They will call you to give talks, I’m sure about that …”. It is curious to see how fate works.

But … why was I there?

The truth is that I wasn’t there by chance. The point is that for several months we have been organizing between Joxemari Gallastegi, Jesus Angel Bravo and myself the introduction of an AI community called AI Saturdays to Euskadi – to both Donostia and Bilbao. And due to the obvious relationship between AI Saturdays and the theme of this edition of the Euskal Encounter, they decided to contact me.

And … what was the talk about?

I wanted to make a somewhat different talk, which was aimed to be something more curious to see / hear. The objective was to understand non-productive applications of the Data Science process, which used techniques and algorithms typical of data analysis and AI. With this goal in mind, I analyzed 3 different examples:

  1. El Odiómetro: A measure of the level of hate on Twitter using sentiment analysis.
  2. Meme Generator: A meme generator using Convolutional Neural Networks.
  3. Jukebox: A musical style transferor using a combination of a VQ-VAE + Autoregressive Transformers model.

And finally I gave a few strokes of the benefits that AI Saturdays could bring to the participants of the program. Although I liked the content (which could be improved) I delivered on the live event, it could not be appreciated well due to the audiovisual quality (which could also be improved). So I decided to re-record myself at home and post it on Galde‘s YouTube channel.

I leave both videos so that you can appreciate the one you prefer (SPOILER: I much prefer the new recording; in the live there are a couple of concept errors).

VIDEO 1: EXTENDED RECORDING

 

VIDEO 2: LIVE RECORDING

 

Which was my impression?

It was a great experience that I would like to repeat again (although if possible, in face-to-face format, not online). This type of community that brings together computer fans, both professionally and amateurishly, is the type of community that is necessary and that makes a social good of hardly tangible value. I would have loved if the first recording had been the final one, but it is not something that could not be helped.

I would also like to thank.