Blog

I’ve been thinking about this post for a while now: I’ve collected a handful of great mini tools which I use day by day to make my life easier. These are various small onefunctional web or offline tools, some of which you may find useful as well.
more...

I’ve had a very interesting start for the autumn: not only gave a nice keynote at IEEE CogInfoCom about the Neuroscience of Virtual Reality, but also the track that we organized with my father Zsolt Török received the Best Session Award. It is really good to feel that our work invested pays off. And there are some interesting things coming in November.
more...

I’ve finally managed to find some time to write a short post about our winning solution at the Telekom Leading Data Hackathon. The aim of this post is to give an example of how to use deep learning in a practical business use case.
more...

The prestigious journal Psychophysiology has just published our new paper on the neural background of navigation. For the impatient (like myself) I’m summarising the interesting parts of the results and methods here.
more...

About five years ago, we have come up with a idea about grid cells that challenged mainstream scientific thinking. While grid cells in rodent form a context independent coordinate system for spatial navigation, we thought that their activity actually does depend on the context in humans. To verify this we designed an experiment where epliepsy patients with implanted electrodes had to navigate in virtual environments of different sizes.
more...

In recent times the popularity of Bayesian statistics has greatly increased, thanks to the large computing power of modern computers. As a result, there is an ongoing debate on whether the Bayesian or frequentist approach is more suitable for statistical and scientific purposes. A great number of introductory papers on these two schools of statistical inference are available online, therefore I will not spend time reiterating the basic definitions here (for those who are interested I recommend these two documents, which contain excellent introductions to the topic). Instead, my goal is to take a look at which of these two approaches is worth following in order to reach the right statistical conclusions – whatever the definition of ‘right’ may be.
more...

I’ve just returned from pilot testing in the VeME lab directed by Elisa Ferré at Royal Holloway, University of London thanks to my EPS grant. We are studying how vestibular and visual informations interact in the perception of gravity. We focus the current experiment on the dynamic nature of gravity, namely the perception of moving objects.
more...

Good start for 2017: our paper is in PLoSONE!
more...

So here we are: another great year is ending tomorrow. During the Christmas period I looked back to what happened in 2016, despite the phenomenal John Oliver wrap up, I’m pretty happy with my achievements this year. First I wanted to write a long post about what I am proud of and thankful for, then I realized that in this case a picture indeed tells more than 10^2 words. So here is my 2016 infographic.
more...

While I was preparing for the PhD defense, I figured these experiments would be more easy to understand if one can even be there. Luckily, this can now be easily solved with the Google Cardboard SDK and Unity 3D. So here is my Doctoral dissertation in VR :)
more...

I gave a talk at the Budapest BI Forum 2016. This was one the best events I’ve been to this year, respect to the organizers. The talk was about how to communicate data science results to the management. This has been an ongoing challenge for me, so I was happy to share some tips based on my experiences. You can have a look at the slides yourself here.
more...

I’m giving a talk at the CogInfoCom 2016. My slides on what I see as the future of HCI can be seen here made in Slides of loci. The paper which the presentation is based on can be read here on your request.
more...

So we have won the Qusp Prize at the largest Brain Hackathon :)
more...

I gave a talk to a brilliant audience yesterday at the NLP meetup. My slides on what we do at Synetiq are available here made in Slides of loci. If you are wondering what that is: it a 3D capable presentation tool I made to design functional and effective 3D enabled presentations. You can read more about the philosophy here.
more...

Next week is going to be a little bit busy in research, we can meet at several events:
more...

My talk on using hidden Markov modelling to uncover emotions is available online.
more...

I’m traveling to Barcelona to visit the ECVP2016, where I continue my mission on AR evangelism. I learnt a couple things from the ICOM6 that made this poster a better shot.
more...

I’m lucky to give a talk at the satRday conference on September 3 2016. I will tell about time series analysis based on our recent work at Synetiq lab. The most interesting part of this will be about how Hidden Markov Modeling can help identifying emotions based on data from EEG, heartrate, and GSR recordings.
more...

Our team, YetAnotherDirtyCelticGod, finished at 28th place in a very strong field of data enthousiasts at the Senior Data Science challenge on drivendata.org.
more...

This week, I’m at the ICOM6 conference, this is probably the largest conference I have so far been to. I have just submitted my PhD thesis on virtual reality and spatial cognition and brought here a slice of that.
more...

I was playing basketball in high school, and those were the times when everybody started to be crazy about Kobe here. So I did feel that an era has ended on 13 April. Celebrating his career, Kaggle started a competition where you have to predict missing shot info in his stats.
more...

I gave a talk about how illusions could save your life at Pszinapszis. I was surprised to see the room being so crowded for such a cognitive topic, there were even people sitting on the floor. It was a great experience to talk about my research in front of such a good audience. Thanks to the organizers for inviting me.
more...

Yesterday, Prof. Endre Szemerédi came to visit us in the Synetiq lab. We had one serious discussion on graph similarity and clustering. It started as a half hour meeting and ended up 11.30pm and we were still there :)
more...

We are very lucky to give a talk at the next event of the Budapest Data Projects meetup series.
more...