A fairly long time ago in an office actually rather close by…

Koen van der Blom joined the ADA research group as a post-doctoral researcher. From March 2019 onward he started working with Holger Hoos in the area of meta-algorithmics. One of the things he works on is a tool called Sparkle. Sparkle aims to make meta-algorithmics such as algorithm selection and configuration easier to use for a wide audience. Besides this, he is also interested in performance analysis and prediction. What can be said about the expected performance of a new instance, based on previously seen instances? And how can you compare performance in a fair way?

Before this, during his PhD, Koen worked on multi-objective mixed-integer evolutionary algorithms applied to early-stage building design under the supervision of Michael Emmerich, Hèrm Hofmeyer, and Thomas Bäck. He continues to be interested in these problems, and particularly when it comes to optimisation in mixed-integer spaces. Who knows, perhaps combining aspects from the old and new will yet lead to other exciting work.

Laurens Arp has joined the ADA research group to perform his Master thesis project

How could AI techniques be used to best improve the living standards of people around the world? This is the main question of interest for Laurens Arp, a Master student who joined the ADA research group in November 2019. He is currently working on his Master Thesis under the supervision of Mitra Baratchi and Holger Hoos. The project is about the data-driven evaluation and optimization methods of geographical regions.

The main focus of the project will be on (spatial) representation learning. Current methods would not sufficiently address the problem yet, as most approaches will either focus too much on spatial structure instead of how this structure affects the features of a neighborhood, or are aimed too much at encoding similarity rather than interaction. Once a suitable representation has been found, machine learning and deep learning could be used to automatically learn the relationship between geographical features and the measures one might like to use to evaluate a region. If the resulting model is sufficiently accurate, it could then be used to rate the quality of region configurations generated by an optimization algorithm, allowing for the optimization of the development of the region.


The ADA Research Group welcomes a new visitor!

Zhou joined the ADA Research Group in September 2019 as a visiting PhD student for a period of one year. He has started his PhD in March 2017 under the supervision of dr. Gangquan Si at School of Electrical Engineering, Xi’an Jiaotong University.

In his research he focuses on time series data mining techniques, including pre-processing, representation, classification, and prediction. His work has been successfully applied to a project titled “Research of Data Mining Technique and Development of Intelligent Data Management System for Electrical Equipment”, supported by China Southern Power Grid. In this project, he tries to make full use of massive data collected from various tests and online monitoring systems, aimed at providing accurate evaluation and prediction of electrical equipment status.

Zhou is currently working on online time series segmentation with the purpose of dimension reduction, especially on how to apply Automated Machine Learning methods for this task. During his visit, he will be closely supervised by dr. Mitra Baratchi and prof. Holger Hoos.

Image from iOS

Double win at ICT.OPEN 2019: Anna Louise and Can take home pitch prize and poster prize

This year, the ADA group was well represented at ICT.OPEN, the conference for ICT professionals in The Netherlands, which took place on Tuesday 19 and Wednesday 20 March.

Cooperation is key

Prof. Holger H. Hoos gave the keynote lecture on Tuesday morning, talking about how machine learning is transforming the way we do science. He argued that, while competition is important, cooperation is what ultimately makes us conquer the problems of our times.

Impression of Prof. Dr. Holger Hoos’s keynote lecture by Flatland.

Bob the builder

That evening Anna Louise Latour won the elevator pitch competition. In the final round, four PhD students competed to explain their research in under three minutes to a jury and the audience, using only a single prop to illustrate their story.

Anna Louise described how make optimal decisions under constraints and uncertainty, using an optimisation version of the Powergrid Reliability Problem as an example. During her pitch, she used a yellow Bob-the-builder helmet to distinguish between the user with the problem, and the Computer Scientist with the solution.

She impressed jury and audience with the clarity and content of her pitch, and with her charisma, ranking highest in the jury report and receiving three times as many audience votes as the runner-up.

Predictive power

The next day, we had another win: Can Wang was awarded second prize in the Commit2Data poster competition. Her poster explained how Automated Machine Learning can be used to predict the energy consumption of a household. “Pitching my poster to the jury was a bit out of my comfort zone, but I’m very happy that I’ve won!”, Can said afterwards.

Missed it?

For those of you who missed ICT.OPEN this year: make sure to join next year! In the mean time, enjoy NWO’s video of this year’s event:

Cooperation is key, revisited

Of course this outcome was also a team effort. ADA group members, as well as other people from LIACS, helped Can and Anna Louise by providing feedback on their posters and presentations, or being there to cheer.  Therefore, we celebrated as a team, with cakes 🙂

New master student joining ADA

Daniël Fokkinga joined the ADA research group as a Master’s student in September of 2018. His research is jointly supervised by Anna Louise Latour and Marie Anastacio and applies Automated Algorithm Configuration (AAC) on a pipeline for solving Stochastic Constraint Optimisation Problems (SCOPs).

An example of such a SCOP is the viral marketing problem. We are given a social network where an edge between two nodes represents a probabilistic communication relationship between two people in the network. We want to launch a new product by handing out a limited number of free samples to people in the network. We hope that they will like our product and spread the word to turn their acquaintances into buyers of our product. To whom in the network do we give the free samples, in order to maximise the expected number of people that will buy our product after they’ve heard about it from others? Who are the most influential people in our network?

Recently, a new pipeline for solving SCOPs was proposed. While it has shown its merit in a proof of principle, the different steps in the pipeline have not been optimised yet. While alternative design choices for elements in the pipeline have been proposed, their influence on the performance of the pipeline over a wide range of problems has not been extensively explored.

Daniël’s focus is on exposing these alternatives as parameters to optimise the pipeline for different applications, using Automated Algorithm Configuration.

Automated Algorithm Configuration allows a user to leave the tuning of parameters and thus the decision of the best approach to follow for parts of the full pipeline to a computer. Given a set of example instances, a configurator finds the combination of parameter values that is most likely to perform well on a new instance.

With this research Daniël hopes to show another successful application of Algorithm Configuration in a new field and improve the performance of previously developed algorithms.


New guest researcher

Ada research group welcomes another guest researcher, Yi Chu. She will stay with Ada research group for a year.

Yi is a PhD candidate at the Institute of Computing Technology, Chinese Academy of Sciences. She received her master’s degree in computer technology from Beijing University on Posts and Telecommunications in 2014.
Yi’s research interests include heuristic algorithms for NP-hard problems and automatic algorithm design using optimization and learning techniques. Currently, she is working on using programming by optimization to improve the performance of heuristic algorithms for solving the maximum clique problem.Yi Chu-1709

ADA research group has a new visiting scholar!

ADA research group welcomes Yanyan Xu, a new visiting scholar!

Yanyan Xu is a visiting scholar at LIACS. She joined ADA Research Group in September 2018. Yanyan holds M.Sc. and B.Sc. degrees from Sun Yat-sen University. She also received her Ph.D. degree from the Institute of Software, Chinese Academy of Sciences. Since then, she has been working at the School of Information Science and Technology of Beijing Forestry University. Currently, she holds an associate professor position there.

Yanyan Xu has a broad interest in the areas of artificial intelligence and algorithm design. She is particularly interested in pattern recognition and deep learning, as well as, heuristic algorithms in robotics and formal methods. Currently, she is working on combining formal methods with deep learning.


ADA research group welcomes a new PhD student

Photo credit: Hélène Verhaeghe

Anna Louise Latour started her PhD research in January 2017 at the Institute of Information and Communication Technologies, Electronics and Applied Mathematics (ICTEAM) of Université catholique de Louvain (UC Louvain) in Belgium, under supervision of dr. Siegfried Nijssen (UC Louvain) and prof. dr. Joost Kok (Universiteit Twente). She came to Leiden in February 2018 and joined the ADA research group in November 2018.

Her research is funded by an NWO TOP grant awarded to dr. Nijssen for his PROFIDDS (PRObabilistic Features for Intelligent Declarative Data Science) project. Within this project, she focuses on developing solving methods for problems in which users have to make optimal decisions under constraints and uncertainties.

An example of such a problem is a Power Grid Reliability Problem. Suppose that we want to help a user who runs a maintenance project on an electric grid to make it more robust against natural disasters (like earthquakes or hurricanes). During a disaster, each powerline in the grid has a certain probability of breakage (uncertainty). If too many of them break, important buildings like hospitals may become disconnected from the grid and lose power. By reinforcing powerlines, the user can make them stronger and less likely to break during an earthquake. However, reinforcing lines is an expensive task, and the user only has a limited budget (a constraint). On which powerlines should the user spend their budget, such that they maximise (optimal decision making) the expected number of important buildings that are still connected to a powerplant?

Her aim is to develop solving methods for these kinds of problems such that they are a) generic, and therefore applicable to a wide range of problems, and b) accessible, even to people who are not programmers.

In order to achieve this goal, she combines modelling and solving techniques from both Constraint Programming and Probabilistic Logic Programming.

Together with Marie Anastacio, Anna Louise also supervises Master student Daniël Fokkinga in his Master’s research.

Sharing our passion with the next generation of students

Within our ADA group, we are passionate about the cutting-edge research we do, but we are also always keen on sharing our passion with others – most recently, with a group of talented high-school students, some of whom we hope will become part of the next generation of advanced computer science researchers.


Students from several VWOs (secondary school for future university students) of Leiden and neighbouring cities came to the Leiden Institute of Advanced Computer Science (LIACS) to get a hands-on experience of computer science and to see if they would like to join our university as bachelor students in 2019.

15 top students participated in this introductory course. First, they got an introduction to networks and shortest path algorithms from Dr. Michael Emmerich (Associate Professor at LIACS). Then, the ADA group took the torch: Dr. Holger Hoos (Professor of Machine Learning), Marie Anastacio and Can Wang (PhD candidates) set out to give them a taste of computational complexity and some of the techniques used to deal with complex problems.

To do so, we built on the shortest path idea and worked with out pre-university students on the travelling salesman problem (TSP). The TSP involves finding a shortest, quickest or cheapest round trip visiting a number of locations – e.g., cities or villages. It is one of the most widely studied problems in computer science and has a broad range of important applications in logistics, manufacturing and even biology. We helped the students to discover effective techniques for finding shortest round trips visiting various locations in the Netherlands and introduced them to slightly simplified versions of the best known algorithms.

More precisely, endeavoured to find the shortest tour visiting every provincial capital of the continental Netherlands (by bicycle, of course). As we bet on the length of the resulting “grand tour of the Netherlands”, almost all of our ADA group members overestimated the size of the country, and we were all surprised by the result. Somewhat embarrassingly for our Dutch group members, the winner of our little game was Chuan, our Chinese postdoc. To give Dutch national pride a second chance, we asked each of our high-school students to place a bet on the length of the optimal tour. The best of these bets came from Jonas, who precisely tied with Chuan – estimating the shortest bicycle your through the 12 provincial capitals at 900km. We decided to declare him the winner of our game, breaking the tie in favour of the younger and less experienced student.

The actual shortest tour is 1013.45 km long and would take around 52h to cycle, according to Google Maps:


Working with the pre-university students was a very interesting and highly enjoyable experience for all of us, and we look forward to doing it again next year. As for our college students, we hope to meet them again soon. And who know, perhaps one day, we’ll cycle the tour …

1M € DACCOMPLI Project Launched


We are very excited to be a key contributor to the DACCOMPLI (Dynamic Data Analytics through automatically Constructed Machine Learning Pipelines) Project, which was officially kicked off on 9 February 2018, with a meeting of all consortium members at LIACS, Leiden University. The project aims at developing a platform for dynamic data analytics based on techniques for automatically constructing machine learning pipelines for a broad range of real-world challenges, including dynamic management of energy stored in fleets of electric cars and prediction of Parkinson’s disease; it received one million euros in funding from the Netherlands Organisation for Scientific Research (NWO) and private partners.

The people involved in the project include Prof. Dr. Thomas Bäck (PI and adjunct member of ADA), Prof. Dr. Holger hoos (co-PI and head of ADA) and Can Wang (PhD student and member of ADA) from LIACS, Leiden University, as well as colleagues from Delft University of Technology, Eindhoven University of Technology, Honda Research Institute Europe GmbH and the Leiden University Medical Centre (LUMC).

The project will develop algorithm configuration approaches for composing, configuring, and parameterising data analytics pipelines from scratch – thereby automatically generating the best solutions for a broad range of data analytics tasks. To demonstrate the approach, two complementary practical application tasks have been selected: The early detection and treatment optimization for Parkinson’s disease, and the cost-effective and environmentally optimised management of energy for private households with electric vehicles. The first case deals with video recordings and slow dynamics over time (analyzing the progression of the disease over a series of diagnostic observations), while the second addresses numerical data with fast dynamics and the need for optimal decision making in real time. Our group is focussed on the second application, in cooperation with Honda Research Institute Europe GmbH, as well as on foundational work together with TU Eindhoven on the automatic construction of data analytics pipelines (an application of AutoML), while work at TU Delft and is focussed on the Parkinson application.

Links: ADA Research Group ∙ LIACS ∙ Leiden University