I have been featured at The Science Museum's Robot Exhibition. The picture shows the iCub humanoid robot built by IIT, Italy. The bottom of the picture shows myself around 8 years ago when I was starting PhD and was training the iCub robot to learn various actions using the GPU-accelerated Multiple Time-scales Recurrent Neural Network (MTRNN). Back then my work with GPUs and their applications to machine learning was one of the earliest attempts to apply CUDA parallel computing in cognitive robotics. Since then the field has come a long way and I have played a key role in establishing a CUDA teaching centre in Plymouth where I also taught parallel computing. After that point, myself and Davide Marocco were working with NVIDIA to provide 2-3 day courses across UK.
NVIDIA features my work again
NVIDIA has recently published a cool article about my life during the past 10 years and how GPUs played a major role in my career. 5 years ago, NVIDIA featured my work on their site for the first time. Back then I was still doing my PhD in GPU computing for cognitive robotics with the iCub humanoid robot. I hope you will enjoy these posts and thank you for popping over to my site from time to time :)
Last Updated on Tuesday, 30 December 2014 14:24
The first TED talk on GPU Computing and Robotics
Here is the TEDx video with English subtitles. I hope you will enjoy it :) To enable subtitles, click CC on the bottom right part of the video.
Last Updated on Wednesday, 08 October 2014 21:06
My TEDx Experience
I have recently done a TEDx talk and now I finally have got some time to write about my experience. In my previous post, I have mentioned that the TEDx talk was going to be in the Slovak National Theatre, Bratislava. I was really happy to return back to my home country after so many years not being there. I have to admit that doing this TEDx talk was one of the most rewarding things I have done in my life...and I have done plenty of crazy stuff already. The talk required lots of preparation just like with everything else worth doing. The presentation style was completely different and focused on getting the main points across to a general and not necessarily technical audience.
Initially, I was going to speak about my research with the iCub humanoid robot and how I applied parallel GPU computing to make him learn actions and language. I prepared my talk and reviewed it multiple times with the TEDx team. The slides were cool and all was good, however, I felt that something was missing. I was thinking: "I have this awesome opportunity to speak to hundreds of people and I should make it worthwhile". I am very passionate about GPU computing and AI but another side of me is very passionate about inspiring and encouraging people that everything is possible with passion, perseverance and hard work. My life story was featured several times and tells why I believe this so strongly. Doing a TEDx talk without being able to express this and try to encourage others through this story would really be a shame. Therefore, couple of hours before the talk, I changed my plan. I decided to improvise and start by talking about my life story of coming to UK with nothing and achieving my dreams step-by-step. It was the best thing I could have done. I felt like I could connect with the audience and transfer my passion of dreaming big and believing that everything is possible if we work hard for it!
The audience was massive, somewhere around 600 people. I kept making people laugh by telling this story. I never thought it that funny but it might have been the way I presented it. I continued the talk and got back to the original plan of talking about the power of human imagination and how it helps creating the future. When it comes to AI, I mentioned that this imagination reflected in many movies but we have still a long way to go to get to that level. The presentation carried on describing classical and new approaches for developing AI. Then the focus slowly shifted to new AI, artificial neural networks and computing power.
I am a great believer in GPU computing and I had to play the Mythbusters video comparing CPUs and GPUs. I love this video despite it being slightly biased towards GPU. The last part of the talk focused on how I used GPUs to develop complex artificial neural network that were able to control the iCub and allow it to acquire complex actions and basic language understanding.
Here are few pictures from the event. Many others are available on the internet, just search for TEDx Bratislava 2014. The videos will be also online very soon so stay tuned!
Last Updated on Wednesday, 03 September 2014 14:29
GPU Computing for Cognitive Robotics
This is probably the last post regarding my PhD. I have printed four copies of my thesis. Two for my parents, one for the best supervisor in this world Angelo Cangelosi and one for myself :)
This PhD has ended but I am just starting...
Last Updated on Thursday, 03 July 2014 19:08
TED talk on GPU computing and cognitive robotics
I am happy to announce that I will be doing a TED talk in the Slovak National Theatre, Bratislava. This will be the very first TED talk on GPU computing and its application to cognitive robotics. I am preparing something really cool and inspirational as this will be a great opportunity to reach approximately 1,000 people who will attend the event and later thousands more people online. The talk will be in my mother tongue (Slovakian) but the video will be later subtitled in different languages. If you happen to understand Slovakian then see the picture below, which was used on TED Bratislava Facebook page to introduce my profile. If you do not speak Slovakian then the picture basically says that I am specialising in parallel computing and that I have previously collaborated with European Space Agency and work at NVIDIA. It also mentions that I am currently working at Cortexica at Imperial College London where we are developing biologically inspired visual search system. The profile is now also published on TEDx Bratislava. Stay tuned... :)
Last Updated on Thursday, 05 June 2014 13:07
Submitting PhD Thesis
After 5 years of hard work and postponing PhD two times (European Space Agency and NVIDIA) I am going to be submitting my PhD thesis on 28 March, 2014. Once submitted, I am going to write a blog post that will summarize the journey and highlight what has been achieved throughout this time :)
This website was created at the time I started PhD. Initially, it was getting around 10 visitors per day and it took few years to reach the first 100, 000 visitors. Over the last couple of years, however, the number of daily visitors increased to approximately 700 and a few days ago reached the total number of 700, 000 visits.
Finishing the PhD and a couple of on-going collaborations with NVIDIA have made it quite challenging to keep this website updated with interesting articles. However, very soon, I am going to have more time and looking forward to writing cool articles on cool topics!