Home     Books & Chapters     Publications     Video Lectures     PyHubs     Typing Dynamics    


News

Video lecture on person identification based on keystroke dynamics
21 July 2016

We have just uploaded a short video lecture about our work on person identification based on keystroke dynamics. See our page containing resources related to typing dynamics.

A model for classification based on the functional connectivity pattern dynamics of the brain
15 July 2016

We are happy to announce that our recent work titled A model for classification based on the functional connectivity pattern dynamics of the brain has been accepted for publication in the proceedings of the Third European Network Intelligence Conference (ENIC 2016) and it will be presented at the conference. See you there!

Challenge - person identification based on keystroke dynamics
10 June 2016

We are happy to announce an open challenge on person identification based on keystroke dynamics! Person identification is essential in various online services such as banking or online exams. Typing dynamics is a simple biometrics that allows user identification. In order to successfully apply user identification based on keystroke dynamics, machine learning is required. In order to encourage research in this important domain, we announce an open challenge.



Collaboration with researchers of the Babes-Bolyai University
10 June 2016

Together with researchers from the Babes-Bolyai University in Cluj Napoca, we aim to explore how community detection techniques can be used to find communities of brain regions. We analyse the community structure under various conditions, such as different phenotypes, presence or absence of a disease, addiction, etc. On the long term, such an analysis may contribute to better understanding the phenomena and mechanisms underlying various diseases and phenotypes. First results of this collaboration will be presented at the 14th International Conference on Parallel Problem Solving from Nature and they will appear in the proceedings of the conference.



Classification of fMRI data using Dynamic Time Warping based functional connectivity analysis
10 June 2016

We are happy to announce that our recent work on Classification of fMRI data using Dynamic Time Warping based functional connectivity analysis has been accepted for publication in the proceedings of the EUSIPCO 2016 conference.



Meet us at SACI!
09 May 2016

In just a few days, we will start off to Timisoara where IEEE's 11th International Symposium on Applied Computational Intelligence and Informatics will be organized. We aim to give two talks, one about the prediction of interactions between drugs and pharmacological targets and another one about person identification based on keystroke dynamics. We hope that you will find them interesting!



Ladislav's talk: recommender systems are much(!) more than matrix factorization
25 Feb 2016

Today, Mr. Ladislav Peska, our guest researcher gave an extraordinarily interesting talk about recommender systems. During the informal discussion in the cafeteria of our institute, we learned from him that real-world recommender systems are much more interesting than we would have ever thought: what matters is not another regularization term in the optimization objective of matrix factorization, but the surprisingly many different types of data that capture different aspects of users' behaviour. Please click here for the slides.



Our paper about classification of gene expression data is now available
25 Jan 2016

The published version of our recent work, titled Classification of gene expression data: A hubness-aware semi-supervised approach is now available at here. You can also watch the audio slides to get an extremely short summary of the results discussed in the paper.

Ladislav's visit
26 Jan 2016

Mr. Ladislav Peska, co-author of one of our recent papers, has recently started his research stay at the Brain Imaging Center of the Research Center of Natural Sciences. We wish successful work!



Dori's new project
26 Jan 2016

Dóra Neubrandt, who successfully participated at TUDOK 2015 conference with the Angel Personal Sleepometer is working on a new machine learning project. We hope, we will be able to report about new results soon!



Paper award at CORES 2015
30 May 2015

The 9th International Conference on Computer Recognition Systems (CORES) took place in Wroc³aw (Poland) a few days ago. Best papers and best presentations were kindly awarded by the Organizing Committee of the conference. It is an honour for us to announce that Krisztian Buza was one of the awarded participants with his work titled Semi-supervised Naive Hubness-Bayesian k-Nearest Neighbor for Gene Expression Data.



Mararu Vlad joined the BioIntelligence lab for an internship
30 May 2015

Mararu Vlad, MSc student, has recently joined our team for an internship with strong desire of studying machine learning and working with us on challenging research tasks in bioinformatics. In particular, Mr. Mararu will focus on the implementation of hubness-aware classifiers and regression techniques in Python and R. The relevance of his work in underlined by the fact that, despite the growing interest in hubness-aware classification, only Java-based implementations are publicly available at the moment. Therefore, his work is expected to extend the hubness-aware research community and to contribute to further applications of hubness-aware techniques.



Krisztian Buza and Noémi Ágnes Varga have just presented their work at DUCOG
22 May 2015

Krisztian Buza and Noémi Ágnes Varga have just presented their poster at the VII. Dubrovnik Conference on Cognitive Science. The Authors used machine learning to estimate the UPDRS score from speech data and previous UPDRS scores. The estimation of UPDRS score is relevant for patients suffering from Parkinson's Disease. As a potential visionary application, UPDRS scores could be estimated while patients make phone or Skype calls using theirs smartphones or tables. Our work is intended to be a small step toward this visionary application.
The work was performed in collaboration between Noémi Ágnes Varga, medical doctor and PhD student of the Semmelweis University and Krisztian Buza who has expertise in machine learning and data mining.
Click here to see the poster.


Krisztian Buza is an invited speaker of the 9th Hungarian-Japanese Symposium on Discrete Mathematics and Its Applications
2015

It is an honor to us that the Program Committee invited Krisztian Buza as speaker for the 9th Hungarian-Japanese Symposium on Discrete Mathematics and Its Applications. Most likely, Dr. Buza will talk about hubness, an interesting property of nearest neighbor graphs and its implications to machine learning techniques, such as classification, clustering and regression. The event takes place between the 2nd and 5th of June in Fukuoka, Japan. See you there!

Kristóf Marussy won the second prize at students' scientific conference
2015

With his study entitled The Curse of Intrinsic Dimensionality in Genome Expression Classification, Kristóf Marussy, BSc student of the Budapest University of Technology and Economics won the second prize of the university's students' scientific conference (TDK). Congratulations!

We opened the BioIntelligence site
2015

We opened the BioIntelligence site. Currently, a first prototype of the site is up, we will extend its content continuously.