× Dashboard Register Log in
ICT Open 2022

Robust & Interactive Artificial Intelligence

Artificial Intelligence (AI) is the major area of innovation in the digital revolution. AI is changing the way we live and work and has great potential for society’s grand challenges, e.g. in healthcare and well-being or digital twin development in robotics and industry. At the same time, developing interactive AI in a robust and reliable way is challenging and needs a combined effort from different research disciplines.

In recent years, Dutch applied mathematicians and computer scientists developed exciting results around deep reinforcement learning, online convex optimization, online statistical learning, and uncertainty quantification for deep learning based on techniques from numerics, optimal control, optimization, and statistics. Mathematical results in such interactive AI methods often provide valuable guarantees towards robustness and explainability towards increased reliability, especially in challenging cases like adversarial attacks or unseen environments.

This track aims to provide an open forum for connecting Dutch research by computer scientists and applied mathematicians on this exciting theme and showcase the Dutch NWO AIM (AI and Mathematics) initiative. The focal point of the track and poster session is the presentation of the work of young researchers. Preference will be given to abstracts by postdocs and PhD candidates.

Chair: Dutch NWO AIM Initiative

Invited speakers

Ciara Pike-Burke

Thursday 7 April
11:30 - 12:10 hour

Decebal Mocanu

Thursday 7 April
12:10 - 12:30 hour

Joris Mooij

Thursday 7 April
14:30 - 14:52 hour

Oral presentation round 2

Robust Deep Learning on Noisy Labels via Trusted Data

Amirmasoud Ghiassi (TU Delft) (e.a.)

Thursday 7 April
14:50 - 15:05 hour

Learning Representations in Deep Reinforcement Learning

Nicolò Botteghi (Universiteit Twente) (e.a.)

Thursday 7 April
15:05 - 15:15 hour

Training Generative Adversarial Networks via stochastic Nash games

Barbara Franci (Universiteit Maastricht) (e.a.)

Thursday 7 April
15:15 - 15:30 hour


Impact of data distribution on the performance of distributed machine learning

Saba Amiri (Universiteit van Amsterdam) (e.a.)

Exploring Explainability and Robustness of Point Cloud Segmentation DL Model by Visualization

Floris Verburg (Universiteit Twente) (e.a.)

Pipeline for the Reconstruction of 3D Models from Point Cloud-based Railway Scenes

Zino Vieth (Universiteit Twente) (e.a.)

Towards A Robust Meta-Reinforcement Learning-Based Scheduling for Time-Critical Task in Cloud Environments

Hongyun Liu (Universiteit van Amsterdam) (e.a.)

Automatic Curriculum Learning for Task-oriented Dialogue Policy Learning

Yangyang Zhao (Universiteit van Utrecht) (e.a.)

Towards Secure & Private AI on the Edge

Daphnee Chabal (Universiteit van Amsterdam)

A visual analytics framework for explainable deep learning beyond classification

Vidya Prasad (TU Eindhoven)

ECiDA: Building distributed data science pipelines using container technology

Frank Blaauw (Researchable B.V.) (e.a.)

Using machine learning for constructing complex phylogenetic networks

Esther Julien (TU Delft)

Model Stealing via Collaborative Generator-Substitute Networks

Chi Hong (TU Delft) (e.a.)

Discovering Missing Links in Scientific Knowledge

Dimitrios Alivanistos (Vrije Universiteit) (e.a.)

Demo stands

Stein Variational Gradient Descent to train ensembles of neural networks

Pascal van der Vaart (TU Delft)

Tackling Mavericks in Federated Learning via Adaptive Client Selection Strategy

Jiyue Huang (NIRICT) (e.a.)

naar boven