I am a second year PhD student working within the ATLAS Collaboration, which is one of the multi-purpose detectors installed at the Large Hadron Collider in CERN. I am currently based out at CERN on a long term attachment away from my home institution. This tumblr is a platform from which I can try to educate and inform others interested in particle physics and wonderful science, as well as provide some insight into the workings of a PhD. Click here to see the kinds of things I like. You can find my personal blog here.
I was fortunate this afternoon to be able to attend a colloquium given by Prof. Henry Markram who is founder of the Brain Mind Institute at EPFL and one of the leading figures of the Blue Brain Project.
I found out at short notice that he was giving a talk but made sure I was available because sometimes you get so bogged down, focussing on your own tiny piece of thread, you sometimes forget to take an interest in the outside world and some of the remarkable things it contains, and this project is such an ambitious one!
I had read a month or two back that this project was vying for some European funding, and by the way Prof. Henry Markram was describing their future plans I can only assume they were awarded it (although I’m not certain as I think it was due to be announced in May).
The Blue Brain Project, at its ultimate climax, intends to model a human brain, but not just in the macroscopic sense of having a model which approximates to a brain, but in an intricate level of detail Prof. Henry Markram put forward the strong case that in the field of neurobiology, there is a problem of fragmentation, much how I described my work. People in the field divesify intensely such that one group of people studying a specific gene, or a certain ion pathway, or a different molecule structure cannot see how this links into the bigger picture of “how the brain works”. It is the aim of the Blue Brain Project to knit together all these lost and mixed up threads by using as much physical data and scientific results to piece together a model at each level of the brain which agrees with all the current information known about the brain, and also interacts all together to produce, on the macroscopic level, a simulation which can learn as a brain does and that could maybe control a form of avatar in a closed loop scenario.
Clearly it is a project that could take a lifetime, but the group have made good progress at École Polytechnique Fédérale de Lausanne, in Switzerland and has become a multinational collaboration. Already the group have been able to demonstrate a simulation of the cortical column processing information and acting on it. A small video was shown of a virtual ball being balanced on a virtual plate which was controlled at four points by the simulation. The purpose was to show that the simulation was able to learn what effects its actions caused (with the explicit directive to keep the ball in the centre of the plate).
Whilst the colloquium was awash with images reminiscent of the Millennium simulation, the sheer scale of the project was just so immense, and so incredibly interesting.
One of the other side projects to emerge from the simulation of the brain is in the field of computer chip technology. If one thinks about what the brain achieves on 20 watts of power, the efficiency and the real-time processing speed is enormous. To achieve the same with supercomputers requires gigawatts of power, so if the way the brain uses ion channels and action potentials to drive processing whilst also using that energy to replenish other energy potential stores in the brain in a near perpetual cycle could be applied to computer chips, the results would be extraordinary.
It should probably be noted as well that they hope by the end of 2014 to have simulated an entire rodent brain.
Now I am hoping that the video of this colloquium will be made publicly available before long, and if I find it so I will edit this post because I think it beneficial for people to hear this guy speak with the passion of his subject.