We are looking for up to 2 students who are interested in creative research and media technology, and want to do a project in this area. We would welcome students who have ideas themselves already but can also meet and make suggestions based on background and interests. The project should fit the requirements for a Media Technology project, but at Bachelor level with a clear computer science component. From animal robot interaction to augmented creativity, anything that fits the criteria.

For reference, see http://mediatechnology.leiden.edu/ and http://mediatechnology.leiden.edu/research/publications but also note the additional criteria (BSc level, CS component).

Peter van der Putten and Maarten Lamers

p.w.h.van.der.putten@liacs.leidenuniv.nl

In this project, an existing Brain-Computer Interface (BCI) implemented in software will be extended with biosignal compression/decompression software tasks in order to reduce the amount of information sent by bio sensors to the BCI, thereby reducing the power needed for data transmission. The extended BCI will be analysed in order to identify an operation mode with reduced energy consumption that still guarantees proper operation of the system.

Todor Stefanov

Todor Stefanov (t.p.stefanov@liacs.leidenuniv.nl))

Step1: Design, optimize, and test a simple embedded processor/controller by hand; Step2: Design, optimize, and test the same processor/controller using Hardware Description Language (VHDL) Step3: Implement (synthesize and map) both designs and compare them in terms of: Performance/Speed and Resource/Hardware requirements

Todor Stefanov

Todor Stefanov (t.p.stefanov@liacs.leidenuniv.nl)

Step1: Design and optimize a robot controller as Finite State Machine; Step2: Implement the robot controller in C; Step3: Analyze/Verify the robot controller by simulating the robot behavior in different scenarios : (1) Can the robot accomplish successfully a given task in a given scenario? (2) How fast can the robot accomplish a given task in a given scenario?

Todor Stefanov

Todor Stefanov (t.p.stefanov@liacs.leidenuniv.nl)

At the Nationaal Lucht en Ruimtevaart Lab in Amsterdam, work with PhD student on modeling adaptive behavior of simulated fighters.

https://askeplaat.files.wordpress.com/2013/01/paper-transfer-learning-2015-08-06.pdf

Armon Toubman (NLR), Aske Plaat (LIACS)

Aske Plaat

Create a system to interactively visualize large search spaces, for example for games such as chess, checkers, hex, or go

https://askeplaat.wordpress.com/teaching-2/masters-thesis-topics/smooth-or-peaked-mapping-state-spaces/

Aske Plaat

Aske Plaat

Information theory can be used to compress data. It can also be argued that this is a way of generalizing knowledge. Try this principle on end game databases

https://askeplaat.wordpress.com/teaching-2/masters-thesis-topics/856-2/

Aske Plaat, Matthijs van Leeuwen

Aske Plaat

On small go boards the game is trivially solved. On large boards it is challenging and exciting. At what board size does this phase transition occur, and what is its nature?

https://en.wikipedia.org/wiki/Solved_game

Aske Plaat, Walter Kosters

Aske Plaat

Excel is used by many ordinary users. Supercomputers can compute large datasets. Create a backend for OpenOffice to link Calc to the Leiden DAS

Come talk to me

Aske Plaat

aske.plaat@gmail.com

Many simulations are used in complex systems where small differences give diverging behavior. Arbitrary precision systems allow such behavior to be studied. This project is about the creation and use of a system for arbitrary precision computation

http://news.leiden.edu/news-2014/leiden-astronomers-carry-out-the-most-precise-measurements-ever-made.html

Aske Plaat

aske.plaat@gmail.com

FORM is a symbolic manipulation toolkit that specializes in processing enormous amounts of terms (intermediate expressions can take up terabyes of disk space). It is mainly used by physicists, to compute analytic predictions for collision events at CERN. Some features are missing or need to be improved. Examples are introducing graph isomorphism functions, improving compression of temporary files, improving polynomial GCD algorithms, etc. The challenge is to write code that is very fast.

https://github.com/vermaseren/form

Ben Ruijl, Jos Vermaseren

benruyl@gmail.com

Data mining researchers and practitioners often refer to many rules of thumb and urban legends such as: 'data preparation is more important than the modeling algorithm used' or 'non linear models typically perform better', however large scale experimental evidence is sometimes lacking. We will leverage OpenML, a repository with over 2500 data sets and close to 0.5 million run results to try and bust or prove some of these data mining urban myths.

See http://www.openml.org/ and http://liacs.leidenuniv.nl/~puttenpwhvander/

Peter van der Putten and Jan van Rijn

p.w.h.van.der.putten@liacs.leidenuniv.nl

Design and Development of Smart services architectures for particular domain such as traffic control, healthcare, emergency etc.

to be defined

a.m.t.ali-eldin@liacs.leidenuniv.nl

developing an effective trust propagation model for Social networks

TBD

a.m.t.ali-eldin@liacs.leidenuniv.nl

developing solutions for privacy protection in Linked Open Data domains

TBD

a.m.t.ali-eldin@liacs.leidenuniv.nl

developing a secure way to transfer electric data on the smart grid.

Amr Ali-eldin + TBD

a.m.t.ali-eldin@liacs.leidenuniv.nl

Browsers use HTML, an internal DOM Model and Javascript to script webpages. A similar structure is desirable for software that reads in node-edge oriented data structures (trees / graphs) into an internal representation that is scriptable with custom algorithms (walkthroughs, shortest path, etc). Both the scripting language, representation and file structure are wanted, either through new development, literature study or a combination. A prototype is also required.

Tim Cocx

t.k.cocx@liacs.leidenuniv.nl

Machine Learning algorithms attempt to learn from past data and make predictions for the future. There are many of such algorithms (see Weka). Choosing the right algorithm influences the quality of these predictions. However, recent insights suggest that choosing the right data representation influences this even more. We will use tools like KNIME and RapidMiner to learn from past experiments which algorithm, parameter setting and data representation should be used for a given dataset.

-

Jan van Rijn

j.n.van.rijn@liacs.leidenuniv.nl

A common data mining problem is given a dataset, predict what algorithm will work best on it. This task is known as the algorithm selection problem. One approach is to use landmarkers: first see how some fast algorithms work and based on that select a good algorithm. Many variants of landmarkers have been proposed in the past. However, most of these can be enhanced by choosing the right representation, which we will explore in this project.

-

Jan van Rijn

j.n.van.rijn@liacs.leidenuniv.nl

Optimization of algorithms for use on GPUs (for instance using CUDA) is still largely a manual task. In this project, we will select an algorithm (or set of closely related algorithms) that you will be mapping and optimizing on GPUs. Example algorithms are for instance: (sparse) matrix computations, graph algorithms, sorting algorithms.

(In case multiple students select this project, each student will work on a different algorithm).

Rietveld

krietvel@liacs.nl

Because vectorization capabilities of many optimizing compilers are weak, the development of vectorized software is still largely a manual task. In this project, we will select an algorithm (or set of closely related algorithms) that you will be mapping and optimizing using modern SIMD instructions (AVX2). Example algorithms are for instance: (sparse) matrix computations, graph algorithms, sorting algorithms.

(In case of multiple students, each will work on a different algorithm)

Rietveld

krietvel@liacs.nl

In a cluster computer, we can take advantage of many different levels of parallelism: vectorization (per-core), multi-core (per CPU), multi-CPU (per node), multi-node. What sort algorithm works best at which level and how do these sort algorithms work together? This leads to the development of hybrid sort algorithms. In this project you will implement two or three distributed hybrid sort algorithms and benchmark these algorithms on the DAS4 and DAS5 cluster computers.

Wijshoff, Rietveld

krietvel@liacs.nl

In this project you will be working on an implementation of distributed matrix multiplication by starting with an elementary specification of matrix multiplication. Through a series of transformations you will be deriving distributed variants of matrix multiplication which you will be implementing, testing and benchmarking on the DAS4 and DAS5 cluster computers.

Wijshoff, Rietveld

krietvel@liacs.nl

In previous work we have devised methods to optimize database (SQL) queries through the use of compiler transformations. Key to this approach is to express queries as loops in which iteration is controlled by "index sets" ("forelem" loops). During code generation, code is generated to compute these index sets. In this project, you will be writing a code generator that can generate vectorized codes (using SIMD instructions) instead. Experimental evaluation is also part of this project.

Previous publications: http://link.springer.com/chapter/10.1007/978-3-319-17473-0_20 http://dl.acm.org/citation.cfm?id=2818180

Rietveld

krietvel@liacs.nl

As the first part of this project you will be investigating the latest compiler technology (latest gcc, clang and Intel compiler) and see whether the compilers are capable of vectorizing entire functions. You will investigate the limitations of the compilers: when does the compiler's vectorizer break down? For a number of the identified cases, you will be investigating whether it is possible to vectorize these cases by hand and if so whether a generic method for doing so can be devised.

Rietveld

krietvel@liacs.nl

New algorithms to browse, recommend or search image and video databases or collections

Dr. Michael S. Lew

mlew@liacs.nl

Data has been collected from children playing on a playground. Obvious data analysis questions could be, e.g., in which 'communities' the children play and how this evolves over time.

http://liacs.leidenuniv.nl/~nijssensgr/bachelorklas-2015-2016/topic-slides/vanleeuwen.pdf

Matthijs van Leeuwen

m.van.leeuwen@liacs.leidenuniv.nl

Many algorithms for community detection, but most of them only consider the network. In reality, much more data is often available, i.e., we often have an attributed graph rather than just a graph. How can we use this information to search for communities that have a description?

http://liacs.leidenuniv.nl/~nijssensgr/bachelorklas-2015-2016/topic-slides/vanleeuwen.pdf

Matthijs van Leeuwen

m.van.leeuwen@liacs.leidenuniv.nl

Mining patterns from data is easy, but how do you present these to a domain expert who is not a data mining expert? One idea is to connect patterns and data through metro maps, which have also been used to visualise other types of information.

http://liacs.leidenuniv.nl/~nijssensgr/bachelorklas-2015-2016/topic-slides/vanleeuwen.pdf

Matthijs van Leeuwen

m.van.leeuwen@liacs.leidenuniv.nl

Data mining algorithms do not always return models or patterns that a user finds interesting, but in some cases this can be learned from the user through interaction. In this project we investigate new instances of this overall approach.

http://liacs.leidenuniv.nl/~nijssensgr/bachelorklas-2015-2016/topic-slides/vanleeuwen.pdf

Matthijs van Leeuwen

m.van.leeuwen@liacs.leidenuniv.nl

Recently, various machine learning algorithms were successfully deployed to increase security of computer systems, networks, or industrial control systems. The goal of this project is to provide an overview of the state-of-the-art of existing solutions in one, well defined area (for example, mining system event logs, network access logs, monitoring network traffic, detection of DDoS attacks, malware detection, etc.), and to design, implement and experimentally validate new solutions.

Dr. W. Kowalczyk

w.j.kowalczyk@liacs.leidenuniv.nl

LSH is a powerful (approximate) algorithm for finding similar objects in huge collections of texts, images, sound recordings, etc. LSH reduces the search time from O(N) to O(1). However, in order to achieve such performance, the algorithm must be tuned - there are several parameters that highly influence the algorithm's speed and accuracy. The objective of this project is to study existing heuristics for tuning LSH and to invent new heuristics and experimentally demonstrate their properties.

Dr. W. Kowalczyk

w.j.kowalczyk@liacs.leidenuniv.nl

Wolters Kluwer (Alphen aan den Rijn) has an interesting data mining project (or projects) in the field of direct marketing. They are looking for (Dutch-speaking) students who would like to dig into their marketing data a produce some predictive models with help of R/RStudio or Python.

Dr. W. Kowalczyk

w.j.kowalczyk@liacs.leidenuniv.nl

Functions of RNA molecules depend on their structures. Using the algorithms predicting structures folded by RNA sequences, it is possible to identify functional RNA patterns and mine RNA genes in the sequence databases.

Alexander Gultyaev

a.p.goultiaev@liacs.leidenuniv.nl

UML is state-of-the-art modelling language for object-oriented systems; a well-known weakness of UML is its lacking in support for consistency, particularly dynamic consistency. Paradigm is a non-standard coordination modelling language, also suited for modelling (unforeseen) self-adaptation. The Paradigm language syntactically guarantees vertical dynamic consistency for general Paradigm models. For a bachelor project, given Paradigm models should be translated into UML 2.0 models.

General UML documentation; for Paradigm papers, ask Luuk Groenewegen

dr. Luuk Groenewegen

luuk.and.liesje@gmail.com; office 132, on Mon-Wednes-Fri only

At the Oostvaardersplassen Staatsbosbeheer will conduct an experiment to control biodiversity in certain areas by influencing the water level. Many measurements for biodiversity will be taken, and need to be analysed

TBA

Joost Kok, Siegfried Nijssen, others

Joost Kok

At the LUMC, large amounts of data are collected regarding patients on the intensive care. In this data, researchers are interested in identifying different types of sepsis, a whole-body inflammatory response to an infection, as different types of sepsis may require different types of treatment. In this project, you will use predictive clustering techniques to cluster the patients and characterize the different forms of sepsis.

http://liacs.leidenuniv.nl/~nijssensgr/bachelorklas-2015-2016/intro.pdf

Siegfried Nijssen, Aske Plaat

s.nijssen@liacs.leidenuniv.nl

Pharmaceutical chemists have collected large databases in which the 2D structure of small molecules is associated with properties of these molecules, such as whether the molecule is toxic or carcinogenic. Based on the 2D structures of these molecules, they wish to predict the properties of the molecules. In this project, you will develop and evaluate new algorithms for this task. Possible algorithms include deep learning algorithms and pattern mining algorithms.

http://liacs.leidenuniv.nl/~nijssensgr/bachelorklas-2015-2016/intro.pdf

Siegfried Nijssen

s.nijssen@liacs.leidenuniv.nl

The Dutch association of mental health and addiction care collects data from all over the Netherlands that reflects the conditions of patients before and after they receive treatment. They are interested in discovering relationships between the conditions of patients and the effectiveness of treatments. You will develop machine learning algorithms that combine linear models and tree-based models to discover these relationships.

http://www.ggznederland.nl/

Siegfried Nijssen, Aske Plaat

s.nijssen@liacs.leidenuniv.nl

Nowadays, large amounts of data are collected that reflect the expression of genes in patients or organisms, as well as the mutations present in the genome of these patients or organisms. In this project, you will develop a tool to analyse such data. Possibilities include to search for an evolutionary tree that is consistent with the mutations observed in organisms, or the discovery of different clusters of patients in terms of gene expression profiles.

http://liacs.leidenuniv.nl/~nijssensgr/bachelorklas-2015-2016/intro.pdf

Siegfried Nijssen

s.nijssen@liacs.leidenuniv.nl

Probabilistic models, such as Bayesian networks, can be used to encode uncertain data, such as for instance social networks in which the strength of connections between individuals is not clear. On this model, new problems can then be solved such as: how likely is it that three people know each other? Which people should receive a targeted advertisement to influence the largest number of people in a social network? In this project you will develop a generic program to solve such problems.

http://liacs.leidenuniv.nl/~nijssensgr/bachelorklas-2015-2016/intro.pdf

Siegfried Nijssen

s.nijssen@liacs.leidenuniv.nl

While for databases, SQL is the commonly accepted declarative query language, in data mining there is no such standard. However, in artificial intelligence numerous declarative programming systems have been developed that could allow to solve data mining problems in a declarative manner as well. In this project, you will study how a pattern mining problem can be implemented in a declarative programming system.

http://liacs.leidenuniv.nl/~nijssensgr/bachelorklas-2015-2016/intro.pdf

Siegfried Nijssen

s.nijssen@liacs.leidenuniv.nl

The study of large graphs (networks) requires large volumes of data to be processed and analyzed in all kinds of different settings. Over the past 20 years, different toolkits, packages and other pieces of software have been developed to analyze large graphs. The student is expected to survey these software packages, and compare the most popular ones, paying specific attention to how well they are suitable for the analysis of large networks of millions of nodes and possibly billions of edges.

https://groups.google.com/forum/#!topic/liacs-thesis-projects/ouas3ZOTISA

dr. F.W. Takes

f.w.takes@liacs.leidenuniv.nl

Several physiological variables, such as heart rate, were measured while students were giving presentations. The research question is whether the resulting time series can be clustered in a meaningful way, such that different responses to presenting can be discovered.

Shengfa Miao, Matthijs van Leeuwen

m.van.leeuwen@liacs.leidenuniv.nl

A modern responsive web platform for the visualization of (social) networks using for example sigma.js or D3.js, and an option to upload a network network dataset and do online computation of various network measures, algorithms and statistics. Result is a service-oriented architecture that runs on specialized high performance hardware (16-core, 1.5TB RAM, 12TB SSD). Challenges in usability, scalability and performance.

http://liacs.leidenuniv.nl/~takesfw/pdf/bachelor-projects2015.pdf

dr. F.W. Takes

f.w.takes@liacs.leidenuniv.nl

IMOD is a software package for analyzing electron tomographs that we would like to run on a cluster computer. This project aims at deploying IMOD in a parallel fashion on the available cluster. You will study the source code of IMOD's batch execution facility (written in Python), make a design how to best map IMOD's workflow to the available cluster, implement the design and run benchmarks.

http://liacs.leidenuniv.nl/assets/Bachelorscripties/Inf-studiejaar-2014-2015/2014-2015SimonRKlaver.pdf

Verbeek, Rietveld

fverbeek@liacs.nl

This project must be done together with the "Tuning IMOD for use in a distributed environment" project. The aim to is provide a clear webinterface for scientists who want to use IMOD on the LLSC. In this webinterface they can upload the source data and configure the computation that must be performed. Upon completion of the computation, the webinterface shows the result of the computation and has the option of downloading the resulting dataset.

http://liacs.leidenuniv.nl/assets/Bachelorscripties/Inf-studiejaar-2014-2015/2014-2015SimonRKlaver.pdf

Verbeek, Rietveld

fverbeek@liacs.nl

PEET is a software package for analyzing particle electron tomographs that we would like to run on a cluster computer. This project aims at deploying PEET in a parallel fashion on the available cluster.

Verbeek, Rietveld

fverbeek@liacs.nl

DIPLIB is a collection of image filters for use in image processing. Currently, DIPLIB is integrated in Matlab. In the last few years, the combination of iPython, NumPy and matplotlib has been gaining a lot of traction as an alternative to Matlab. In this project, we will look into writing Python bindings for DIPLIB, such that DIPLIB can be used from within Python and iPython.

Verbeek, Rietveld

fverbeek@liacs.nl

Design and development of infrastructure and interface for scientists in Biology to easily submit sequencing jobs. The sequencing jobs will be automatically parallelized and distributed over multiple machines in the LLSC.

Collaboration will be sought with the department of Biology to determine the job characteristics and user interface requirements.

Verbeek, Rietveld

fverbeek@liacs.nl

The goal is to develop L-systems for finding structures that resemble real plants. The objective function is the human designer. Research objectives are to develop a genetic algorithm representation of production rules, to implement a software to visualize results and steer the evolution, and to do practical performance assessment.

http://liacs.leidenuniv.nl/~emmerich/BSCThesisTopics2015.pdf

Michael Emmerich

Michael Emmerich

Tensor Flow is the newest machine learning environment launched by Google. As a platform it is supporting the definition and combination of workflows consisting of coupled individual learning methods. The goal of the project is to assess the capabilities if Tensor Flow and compare it to other packages such as WEKA. Besides it has to be evaluated how to couple external optimization software with tensor flow in order to optimize workflows and their parameterization.

Speaking hour: 13:30-16:00 Thursday

Michael Emmerich

emmerich@liacs.nl

What distinguishes successful trials from unsuccessful trials in bitcoin mining? What distinguished wildtype DNA from a DNA of a patient with a disease. Both questions can be solved with pattern mining. In this project advanced pattern mining paradigms are to be tried on gene data and on bitcoin data. If significant patterns are found, this can have profound implications on both fields. Moreover the trade-off between support and complexity of the pattern can be assessed.

Michael Emmerich and Matthijs van Leuwen

Michael Emmerich

This project is about developing and testing optimization procedures based on genetic algorithms and possibly other methods for finding dimensionings in air/water distribition. The project is in collaboration with Stabiplan (Bodegraven).

Michael Emmerich

emmerich@liacs.nl

A recently discussed problem in computer science and statistics is to find optimal sites for sequential experimental planning. A new approach that we are investigating in LIACS is to use ideas from portfolio selection in financial investment problems to plan experiments, or to "invest in experiments". The aim of the thesis is to investigate the design of different variants of such experimental portfolio solvers and to assess their performance. Collaboration with Newcastle U.

Speaking hours Thursday 13:30-16:00

Michael Emmerich, Andre Deutz, Iryna Yevseyeva (Newcastle U)

Michael Emmerich

LIACS and the Built Environment Department of TuEindhoven are running a long term project where we want to use optimization algorithms to automatically search for architectural of buildings that perform well in energy consumption as well as in stability objectives. Building simulation algorithms serve as objective functions, and black box optimization methods need to be found for finding efficient designs. The aim is to analyse the strength of some of these methods and possibly improve them.

Speaking hour Thursday 13:30-15:00 or email

Michael Emmerich

emmerich@liacs.nl

This is a generic project. Michael Emmerich's is a leading expert in optimization with multiple objective functions and constraints, using deterministic and stochastic methods. The idea is to find compromise solutions or to visualize trade-off, but also to find efficient (optimal) solutions. If you know a interesting problem in this domain, the task can be use and further develop advanced methods for solving problems. Both deterministic and heuristic algorithms can be investigated.

Group homepage http://moda.liacs.nl, Speaking hour: Thursday 13:30-16:00

Michael Emmerich

Michael Emmerich

The goal of this project is to develop an interactive visualization tool for large amounts of sensor data. The tool would visualize time series, and allow for zooming in and out, scrolling through the data, etc.

Thomas Bäck; Bas van Stein

b.van.stein@liacs.leidenuniv.nl

HTM is an online machine learning model that models some of the structural and algorithmic properties of the neocortex. It is related to recently developed deep learning algorithms. The project aims at understanding the value of this approach for online learning, and comparing it to at least one other algorithm. It will be applied within the thesis project to synthetical as well as real-world data.

Thomas Bäck, Hao Wang

Hao Wang

Collaboration with BMW.The goal of this thesis is the identification of possible fields of application of data mining approaches for production- and controlling- figures.

Thomas Bäck, Inga Reddig (BMW)

Thomas Bäck

The goal of this thesis is the analysis of possible data warehouse approaches for data mining systems. Collaboration with BMW.

Thomas Bäck, Marco Schönfelder (BMW)

Thomas Bäck

The goal of this thesis is a benchmark of data mining tools.

Tasks: Comparison of the tools: CVA vs R vs Rapidminer vs Hadoop (further can be added) in terms of: Field of application, analysis performance, suitability and data mining purposes in an automotive environment. Connectivity for data in- and out-put into the respective systems (e.g. File Bases, SQL,…). Supported environments and integration possibilities of the results in 3rd party software.

Thomas Bäck, Marco Schönfelder

Thomas Bäck

In this project the student will have to implement a monitor for Java program base on context free grammar or logic

Marcello Bonsangue and Frank de Boer

m.m.bonsangue@liacs.leidenuniv.nl

In this project the student will have to give a graphic representation of a (set of) Java Classes

Marcello Bonsangue

m.m.bonsangue@liacs.leidenuniv.nl

In this project the student will have to study and give an implementation of some novel algorithms for (non-deterministic) automata.

Marcello Bonsangue

m.m.bonsangue@liacs.leidenuniv.nl

The Hallmarks of Cancer is a conceptual framework that describes which biological processes are disrupted in cancer. By developing methods to systematically integrate and map cancer data to the Hallmarks, we will enable new approaches to mining cancer data. We will 1) explore methods for structuring and integrating the data (e.g. NO-SQL databases verses traditional SQL), 2) develop strategies for clustering and mining and 3) develop strategies for visualising data for cancer researchers.

Katy Wolstencroft, Fons Verbeek

k.j.wolstencroft@liacs.leidenuniv.nl

Texts written by non-native speakers often include hidden clues that betray their native language. Writers are prone to misspellings, grammatical errors or unusual turns of phrase that are characteristic of their mother tongue. In this project you will leverage the power of machine learning and automated text analysis to uncover these clues and detect English language texts written by Dutch authors.

Dr Cor Veenman

c.j.veenman@liacs.leidenuniv.nl

Your cellphone's (and your) approximate location can be derived from the location of the cell tower you are currently using. This requires a database of active cell tower locations, but such a database may become inaccurate or incomplete as providers update their network over time. The goal of this project is to evaluate various location databases, given actual GPS measurements.

Dr. Cor Veenman

c.j.veenman@liacs.leidenuniv.nl

Together with the student(s) the exact topic and research question will be determined.

Jetty Kleijn

Jetty Kleijn

Petri nets provide a robust modelling framework for many different types of concurrent systems. All kind of features can be added to represent typical phenomena of the systems under consideration. In this project students investigate the effect of such an extension.

Jetty Kleijn

h.c.m.kleijn@liacs.leidenuniv.nl

In this project the student will investigate an approach for the verification of inter-organisational workflows (with an application to the financial markets domain).

Jetty Kleijn, Pieter Kwantes

h.c.m.kleijn@liacs.leidenuniv.nl

In this project the student gets acquainted with and extends a tool to simulate and analyse (certain) Petri nets.

Jetty Kleijn, Bas van Stein

h.c.m.kleijn@liacs.leidenuniv.nl

Problem G at BAPC 2015 was called `Mario'. In this project, you examine different solutions to this problem, possibly improving on them, and analyse their time complexities. The analysis includes designing worst case instances.

http://bapc.eu/problems/bapc2015.pdf

Rudy van Vliet

rvvliet@liacs.nl

Study a recent paper in the area and investigate its remaining open problems. For example Alzahov&Freund, Polarizationless P Systems with One Active Membrane (CMC16, Valencia, 2015)

http://users.dsic.upv.es/workshops/cmc16/proceedings.html

H.J. Hoogeboom

h.j.hoogeboom@liacs.leidenuniv.nl

De bedoeling is om enig inzicht te krijgen in de huidige stand van zaken betreffende quantum computing. Er bestaan weliswaar nog geen echte quantum computers, maar er is wel degelijk enige theorie ontwikkeld, alsmede enige algoritmen. Daarop ligt de focus van het project, niet op de fysische achtergrond.

Jeannette de Graaf en Andre Deutz

j.m.de.graaf@liacs.leidenuniv.nl

The purpose of this project is to research and develop an IIoT broker (i.e., server) set-up, which provides high availability to its clients. In such a set-up, several brokers operate in a redundant mode. We target the Mosquitto broker and the widely adopted MQTT communication protocol. A comparison between Mosquitto/MQTT and OPC-UA (industrial Machine-to-Machine communication protocol for interoperability) is also subject of this project.

Requirements: Some experience of SW development in C# or Python. Links: https://www.youtube.com/embed/_pRwZb3vGCs?list=PL1DoNZDb4gUFJGnJzcsfrvdp0Ae-JUHdD (IIoT by Honeywell introduction) http://www.hivemq.com/blog/mqtt-essentials-part-1-introducing-mqtt (MQTT Essentials)

Dr. Hristo Nikolov at Honeywell (Delft) and Todor Stefanov at LIACS

Todor Stefanov (t.p.stefanov@liacs.leidenuniv.nl)

designing adaptive cyber security systems for critical infrastructures

Amr Ali-Eldin

a.m.t.ali-eldin@liacs.leidenuniv.nl