• Home
  • Content recommendations at jodel.com & spiking neural networks (In Person event)

Content recommendations at jodel.com & spiking neural networks (In Person event)

 Content recommendations at jodel.com & spiking neural networks (In Person event)
  • LOCATION

    Dataworkz offices, Tractieweg 41 - Studio N, 3534AO, Utrecht

  • DATUM

    22 Sep, 2021

  • TIJD

    18:00

  • KOSTEN

NOTE: This event will be hybrid and with this link you register for the In Person part.

Register for the live event : https://www.meetup.com/DataWorkz/events/280334091

If you’d rather like to join virtually: https://www.meetup.com/DataWorkz/events/280440191/

18:00 - Pizza’s
18:30 - Talk 1 (start livestream)
19:15 - Talk 2
20:00 - Borrel/Drinks (bij live event).

Talk 1: Towards More Energy-Efficient Neural Networks? Use Your Brain

In the last decade many different types of neural networks have been developed. They showed us the amazing power and opportunities of machine learning. On the other hand the whole training process is not that efficient and costs a lot of energy. With that in mind we do know that neural networks are sort of based on the way humans learn. Except that the human brain is much more energy efficient. Could we achieve that same energy-efficient level in artificial neural networks? The answer is yes! I will show you what is often called the third generation neural networks: Spiking Neural Networks.
Let’s go for great accuracy and major energy savings!

By Olaf de Leeuw, Data Scientist at Dataworkz

Talk 2: Content Recommendations in Hyperlocal Communities

Jodel is a hyperlocal community app operating around the entire world
that allows users to share text, images, and videos with other users in
their area in real-time. Among other things like content moderation and
image detection, one goal of our Data Science team is to find the best
content for every user. We are currently moving away from a strictly
(reverse) chronological feed towards a more personalized feed. One core
challenge is to infer content preferences from hundreds of millions of
user interactions every day. In this talk, I will give an overview on
different content recommendation approaches that we have been testing
lately at Jodel, how we cope with the challenge rising from the
hyperlocal setting, how we develop different approaches offline, and
then eventually test them in experiments in our app.

By Fabian Hadiji, Head of Data Science at Jodel

Event sprekers

Andere events