Latest News
-
[March 2023] We released an open source data science course for the UvA Bachelor Informatiekunde Program.
-
[February 2023] A paper ("MATTE: Multi-task multi-scale attention") has been published by Gjorgji Strezoski in the Computer Vision and Image Understanding journal.
-
[February 2023] A paper ("Meta Learning to Bridge Vision and Language Models for Multimodal Few-Shot Learning") has been published by Ivona Najdenkoska in the ICLR conference.
-
[January 2023] A paper ("Expert-Defined Keywords Improve Interpretability of Retinal Image Captioning"), has been published by Jia-Hong Huang in the WACV conference.
-
[January 2023] A paper ("Probabilistic Integration of Object Level Annotations in Chest X-ray Classification") has been published by Tom van Sonsbeek in the WACV conference.
-
[January 2023] A paper ("Association Graph Learning for Multi-Task Classification with Category Shifts") has been published by Jiayi Shen in the NeurIPS conference.
-
[October 2022] A paper ("Uncertainty-aware report generation for chest X-rays by variational topic inference") published by Ivona Najdenkoska has received runner-up for the best paper award in the Medical Image Analysis journal.
-
[September 2022] Marcel Worring received a grant (€1.5M) from NWO about the project of "AI4Intelligence: from Multimodal Data to Trustworthy Evidence in Court". More information here.
-
[September 2022] Our group has moved from Science Park 904 to the new LAB42 building nearby (Science Park 900).
-
[September 2022] Nanne van Noord received a grant (€387K) from ClickNL about the AI4FILM project to develop novel AI techniques that are tailored to film by learning from analysis, production practice, and theory.
-
[September 2022] Yen-Chia Hsu participated in the 2022 Heidelberg Laureate Forum.
-
[September 2022] Yen-Chia Hsu was invited to a panel discussion at the 2022 Internet Governance Forum, where he will participate in the discussion of AI transparency and communities.
Vacancies
-
[Due 4 June 2023] PhD position in AI Tools for Analyzing Multimodal Data in Law Enforcement
-
[Due 21 May 2023] Seven PhD positions in AI for Fintech
Team
Teng Long
Eleni Konstantina
Alumni
Paticipation in ICAI Labs
MultiX participates in the following labs of the Innovation Center for Artificial Intelligence:
-
The AIM Lab (AI for Medical Imaging) is a collaborative initiative of the Inception Institute of Artificial Intelligence from the United Arab Emirates and the University of Amsterdam. It focuses on using artificial intelligence for medical image recognition.
-
The National Police Lab AI is a collaborative initiative of the Dutch Police, Utrecht University, University of Amsterdam, and TU Delft. They aim to develop state-of-the-art AI techniques to improve safety in the Netherlands in a socially, legally, and ethically responsible way.
-
The AI4Forensics Lab is a collaborative initiative of the Netherlands Forensic Institute and the University of Amsterdam. Our focus is on the study and development of artificial intelligence for forensic purposes.
List of Funded Grants
AI4Intelligence: from Multimodal Data to Trustworthy Evidence in Court
- Primarily funded by NWO and co-funded by others (€1.5M), 09/2022 - 09/2028
- Law enforcement has to process huge amounts of data derived from online platforms, digital marketplaces, or communication services, where AI can serve as a solution. In AI4Intelligence we allow AI tool development, the use of these tools by investigators, and legal regulations to go hand in hand so that investigations can lead to trustworthy evidence that is admissible in court.
- Partners: Nationale Politie, NFI, Sustainable Rescue Foundation, TNO, Microsoft, SynerScope BV, CFLW Cyber Strategies, DuckDuck0Goose, ZiuZ Forensics BV, BG.legal, The Hague Court of Appeal, Innovation Team Testlab OM
- Link to the News
AI4FILM
- Funded by ClickNL (€387K), 09/2022 - 09/2026
- With this project, we aim to develop novel AI techniques that are tailored to film by learning from analysis, production practice, and theory.
- Partners: Kaspar
VisXP: Interactive Visual Exploration of Media Archives
- Funded by ClickNL (€391K), 02/2021 - 12/2023
- This project aims to make AI technology widely applicable within media archives based on interactive learning interfaces that enable users to explore the data visually. This requires research into combining the different types of data sources within archives, both for analysis and for displaying and visualizing with an interface.
- Partners: The Netherlands Institute for Sound & Vision, RTL Nederland
- Project Link
List of Demos
PanorAMS: Automatic Annotation for Detecting Objects in Urban Context
The PanorAMS framework involves a method to automatically generate bounding box annotations in geo-referenced panoramic images based on geospatial context information.
We acquire large-scale (albeit noisy) annotations from open data sources.
(website link, paper link)
TindART: A Personal Visual Arts Recommender
TindART is a web-based visual artwork reccomendation system.
The system has visual analytics controls that allow users to gain a deeper understanding of their art taste and refine their personal recommendation.
(website link, paper link)
OmniArt: A Large-scale Artistic Benchmark
OmniArt is a large scale artistic benchmark dataset aggregated from multiple collections around the world.
It is designed for easy data handling and fast integration with popular deep learning frameworks.
(website link, paper link)
GCNIllustrator: Illustrating the Effect of Hyperparameters on Graph Convolutional Networks
GCNIllustrator is a visual analytics tool for illustrating the effect of hyperparameters on graph convolutional networks (GCNs).
It addresses one of the most tedious steps in training GCNs: the choice of hyperparameters and their influence on performance.
(video demo and paper link)
List of Open Source Courses
Data Science (Third Year of the UvA Bachelor Informatiekunde Program)
This course teaches data science pipelines with three modules in processing structured data, text, and images.
Course materials involve Jupyter Notebooks for hands-on experiences and lectures that explain theories behind the implementations.
(website link)