UvA Woo Woorkshop

30 juni. 13:00-16:30. Lab 42 Science Park Amsterdam, Zaal L.101

Live via Zoom: https://uva-live.zoom.us/j/89504741876

Programma

Deze middag presenteren 7 afstudeerders hun scriptiewerk gedaan binnen de Woogle/Wooverheid/NWO Access projecten, aan de UvA in de bachelor Informatiekunde en de Master Information Science. 

Sessie 1 13:00. De Woo als FAIR data

1. Femke Bakker
   • Hoe goed kan ChatGPT automatisch alle datums met hun gebeurtenissen uit een besluitbrief na een Woo-verzoek extraheren en op een tjdslijn plaatsen?
2. Lars Nelissen
   • Hoe toegankelijk zijn stukken vrijgegeven onder de Woo voor visueel gehandicapten? En kunnen we die stukken die dat niet zijn automatisch zo repareren zodat ze dat wel worden?
3. Anne van Musscher
   • Hoe kunnen we aan gegevens over de andere kant van een Woo-verzoek, vanuit het perspectief van de verzoeker, komen? 
4. Ramon Duursma
   • Hoe volledig, accuraat en “vers” (up to date) zijn de bereikbaarheidsgegevens beschikbaar gemaakt op organisaties.overheid.nl
5. Maurice Silverio
   • Is automatische spraakherkenning al goed genoeg om automatisch geschreven notulen te maken uit video-opnames van zittingen van de gemeenteraad? En is daarmee dan een handig zoeksysteem mee te maken? En, als we toch bezig zijn, kunnen we dan ook niet met ChatGPT achtige technieken vragen over zo’n vergadering automatisch laten beantwoorden.

Sessie 2 15:30 ParlaMint, zoeken in Handelingen uit 17 landen tegelijk.

1. Asher de Jong
   • Kunnen we snel diachronisch comperatief onderzoek op al die Handelingen mogelijk maken?
2. Noah Horwitz
   • Wordt een cross lingual zoeksysteem over al die Handelingen fijner om mee te werken als we extra nadruk leggen op de entiteiten die genoemd worden?

---

De workshop is mede mogelijk gemaakt door een Humane AI grant van de Universiteit van Amsterdam, en door het NWO Access project CISC.CC.016.

Andere dingen op Science Park die (mid)dag

June 30th – End of year festival 2023

Get ready for a smashing celebration of the end of the academic year! Imagine it is 30 June 2023 and you have just completed your very last assignment, passed your last exam with flying colours or maybe even handed in your long-awaited thesis. And do you know what you deserve? An evening of unparalleled drinking pleasure, a culinary delight straight from the grill and the catchiest music you’ve ever heard! The barbecue committee has once again pulled out all the stops to end the college year spectacularly. Get ready for an unforgettable evening of endless enjoyment, with even more drinks, even more rousing music and, above all, even more fun. So don’t hesitate any longer and get your ticket now. We will see you on 30 June, ready to party like never before!

Included with your ticket: unlimited drinks (beer🍻 & soft drinks), delicious BBQ food🔥, amazing music🔥 and fun activities!

To get your ticket visit https://svia.nl/flux

Master AI Student Conference June 30

Celebrate the end of the academic year with the master AI students who participate in the courses Interpretability & Explainability in AI, Multimedia Analytics and Recommender Systems and present a poster, demo and/or other kind of presentation.

With keynote lectures of

• Mounia Lalmas of Spotify: One of Spotify’s missions is “to match fans and creators in a personal and relevant way”. This talk will share some of the research work aimed at achieving this, from using machine learning to metric validation, and illustrated through examples within the context of Spotify.
• Mostafa Dehghani of Google: Transformers at Scale as Universal Foundation Models

Check out the full program

Entrance is free for all students and staff!

Het programma Master AI Student Conference

About the Conference
The goal of the student conference is (1) to have a celebratory event at the end of the
academic year and (2) to have a conference in which students who participate in the
courses Interpretability & Explainability in AI, Multimedia Analytics and
Recommender Systems present their work in the course in the form of a poster, demo
and/or other kind of presentation.
In addition to these presentations, prof. dr. Mounia Lalmas will give a keynote lecture
on her work in AI at Spotify and dr. Mostafa Dehghani will close of the day with a
lecture on transformers.
The conference is open to anyone interested.
When?
June 30th, 2023, 10:00 – 17:30
Where?
Locations:
– Lecture hall H0.08, Science Park, Amsterdam
– Startup Village, Night and Day rooms, Science Park, Amsterdam
https://www.startupvillage.nl/
Schedule:
When What Where
10:00 – 11:30 1st Poster/Demo round Startup Village
11:45 – 12:45 Keynote Lecture by Mounia Lalmas* Lecture Hall H0.08
12:45 – 13:30 Lunch (buffet is provided) Startup Village
13:30 – 15:00 2nd Poster/Demo round Startup Village
15:00 – 16:30 3rd Poster/Demo round Startup Village
16:30 – 17:30 Award Ceremony + Closing speech by Mostafa Dehghani** Startup Village

* =
Abstract:
One of Spotify’s missions is “to match fans and creators in a personal and relevant way”.
This talk will share some of the research work aimed at achieving this, from using
machine learning to metric validation, and illustrated through examples within the
context of Spotify. An important aspect will focus on illustrating that when aiming to
create personalized listening experiences, it is important to consider the following three
angles when developing machine learning solutions for personalization: (1) How do we
know where the user is in their journey?; (2) How do we know what the user may want
to listen in the future?; and (3) How do we know that we are looking at the right metrics?.
Bio
Mounia Lalmas is a Senior Director of Research at Spotify, and the Head of Tech Research
in Personalisation, where she leads an interdisciplinary team of research scientists,
working on personalization. Mounia also holds an honorary professorship at University
College London. She also holds an additional appointment as a Distinguished Research
Fellow at the University of Amsterdam. Before that, she was a Director of Research at
Yahoo, where she led a team of researchers working on advertising quality. She also
worked with various teams at Yahoo on topics related to user engagement in the context
of news, search, and user-generated content. Prior to this, she held a Microsoft
Research/RAEng Research Chair at the School of Computing Science, University of
Glasgow. Before that, she was Professor of Information Retrieval at the Department of
Computer Science at Queen Mary, University of London. She is regularly a senior
programme committee member at conferences such as WSDM, KDD, WWW and SIGIR.
She was programme co-chair for SIGIR 2015, WWW 2018 and WSDM 2020, and is one of
CIKM 2023 programme co-chairs.
** =
Title
Transformers at Scale as Universal Foundation Models
Abstract
Transformers have revolutionized the field of machine learning emerging as powerful
and versatile models for various tasks in different domains. However, their true
potential is unlocked when they are scaled up. This talk focuses on the scaling of
transformers as universal foundation models, exploring their capabilities and
implications. We delve into the process of pre-training these models on massive
datasets using advanced computational resources. We talk about the emergent
capabilities and new ideas unleashed at scale like in context learning. Moreover, we

address challenges encountered when scaling, including stability issues and efficient
deployment of these large models. We discuss potential solutions to make these models
more accessible to the broader research and development community. By leveraging the
power of transformers at scale as universal foundation models, we can push the
boundaries of machine learning and foster the development of more robust and
sophisticated AI systems. This talk aims to shed light on the potential, challenges, and
future directions of this exciting research area.
Bio
Mostafa Dehghani is a Research Scientist at Google Brain, working on scaling up
language and vision algorithms, in particular attention-based models. Before Google, he
was doing a PhD in Machine Learning at the University of Amsterdam. During his PhD
he focused on data-efficient deep learning and training neural networks with weak
supervision