If that simulation would be conscious, wouldn't it be just as bad (if not worse) as performing experiments on animals ? and if it wouldn't be conscious, wouldn't it conflict with the beliefs of most scientists on the origin of consciousness ?
If the HBP is really a success and the simulation is conscious, then the simulation should probably get similar rights and protections as a human. ( Whatever freedom of movement means for a supercomputer is a rather interesting additional question. ) And this would likely mean, that the possible experiments on such a simulation need to be severely restrained. On the other hand, if a sufficiently advanced simulation of the human brain would not be conscious, then we run into trouble regarding the purely physical nature of human brains.
However, I tend to think that a more rigid definition of consciousness is needed ( and may well be one of the major results of the HBP). I think this, because I am thinking about similar complex programs as a full brain simulation, which demonstrate complex behavior, but no consciousness. ( For example very detailed simulations of super nova explosions.)
A digital computer simulation would be fundamentally different from a brain. It could be replicated, backed up, and restored with perfect fidelity. In my view that completely changes the ethics of experimentation.
> Modern computing technology has brought these goals within sight.
Haven't we been saying this since we had vacuum tubes and early chess programs?
We thought machine vision would be easy and that turned out pretty tricky.
> ICT is ready to give us a completely new understanding of the brain and its diseases; understanding the brain will lead inevitably to radical innovation in computing.
I can't help feeling this is backwards. As we learn more about the brain our simulations become better. But I'd be really interested in any computer simulations that have improved our understanding of brain stuff.
What they are trying to say is that we can put a lot more data into our simulations and run them at different abstraction levels, which will help us improve our understanding of the brain. This improved understanding can be used for innovation in computing, e.g. through neuromorphic processors, see http://www.kip.uni-heidelberg.de/cms/groups/vision/projects/... which is one of the project partners.
This sounds really interesting - I would definitely apply to work on such a thing. But I am scared away by the science career path. So optimistically you get to develop some neuro simulation code, maybe get a couple of papers out it. What do you do then? Who in industry hires people who worked on this kind of stuff? I'm not trying to be pessimistic - if I figure out a good idea then I'll go for it and apply.
I know several people who have done PhDs in neural simulation and then gone on to work in industry (e.g. banking, IT consultancy).
I've also come across a couple of startups building on brain-inspired computing. The only one that comes to mind at the moment is Brain Corporation in San Diego http://www.braincorporation.com
At what resolution would you have to simulate the brain to get "human like" properties like consciousness? (quantum/subatomic? molecular? cellular? biologically inspired mathematical abstractions of neurons? simplest computationally useful mathematical abstractions of neurons - like current artificial nns?) - this is the question I find most interesting ..instead of unrealistic goals like predicting drug effects in-silico.
It would be helpful to run simulation algorithms with "tunable resolution" on supercomputers to see at what level the interesting properties appear. Though I have better hopes of seeing an answer to this question from the guys doing AI research than from a medical mega-research project or a collaboration... This is very different from the human genome project (where people alredy knew what information they needed and they put together resources to obtain it faster) that they try to imitate in the PR vids...
I agree. To simulate the effects of a drug, you would need to simulate not only the neurons, but the circulatory system, the blood-brain barrier, perhaps even the immune system, and all of their respective interactions. Give the state of our knowledge, a massive publicly-funded project seems premature, like setting out to do the moon landing in 1870.
I don't know about that part of the project, but you could for example simulate the effect of the local interaction between a molecule and a neuron. The mechanisms to put the molecule there could be investigated separately.
> Give the state of our knowledge, a massive publicly-funded project seems premature
One billion might sound like bucket loads of money, but it is not. There are close to 90 institutions involved and a huge percentage of the funding will need to go to the platforms that need to built. This is more about building the necessary infrastructure to do higher-impact research.
> but you could for example simulate the effect of the local interaction between a molecule and a neuron
Of course you can... but I sincerely hope the PR video was inaccurate, because lumping together goals like: (1) small scale cell simulation for basic pharmacology research and, (2) better medical imaging for clinical purposes and (3) large scale simulation for really understanding how things work seems a recipe for confusion and badly allocating resources... yeah, it's a common goal, but the people doing this have very different expertise and are close to "not speaking the same language" in terms of the way they approach problems... an you'd have them competing among each other for funding instead of competing for funding in the overall "research market" and then, let's say, the guys doing low level simulation and more clinical oriented research win the big bucks over the guys interested in large-scale simulations and understanding.
I'm speaking out of my ass a little bit, but what I've learned from my contact with medical research is that the whole system is extremely ("criminally" I'd say...) "good" at mismanaging resources (hardware, smart people, money... everything), and the only hope to have "decent" resource utilization would be by letting small teams self-manage and compete for funding in the broadest "market" not just heir own area of study...
If this is really just an umbrella term for a bunch of independent research, that's different. The problem with a big project organization is that the administrative overhead involved in coordinating all of those 90 institutions will eat up valuable time and research dollars. And with no clear roadmap to guide priorities, that overhead won't provide any benefit. There's too much basic research left to be done before people start trying to coordinate on building ambitious 'platforms'. Better to simply fund a broad range of independent research and see what emerges.
The project is organized into different divisions and measures have been taken to reduce the administrative overhead where possible. Having 50 small projects does not reduce administrative costs, you would still have an enormous amount of reporting, both financial and scientific to deal with (on the EU and on the projects side).
> And with no clear roadmap to guide priorities, that overhead won't provide any benefit
But that is one of the purported advantages of a flagship project. You do have a roadmap and a unifying goal that is provided by the project. Thus you can avoid different institutions performing the same research over and over without any concern to how this relates to previous results and other relevant areas. The difference is that the roadmap is provided by the project instead of the funding agency and as such there is more flexibility.
> building ambitious 'platforms'
To take one of the platforms as an example, the idea behind the brain simulation platform is to be able to aggregate scientific data collected one way or another (even outside the project) to build a simulation model. The more data collected the better the model. Then scientists can come and test their hypotheses or run scenarios. Based on the results the brain model may be adjusted. The way I see it, this will be an evolving tool that will facilitate basic research. I am not a neuroscientist so I can't really comment on whether this tool makes sense or not.
From the FAQ -
Why not begin with simple organisms like C.elegans?
There are two problems here. The first is feasibility; the second is the relevance of our results.
Feasibility. Neuroscientists have mapped all of C.elegans’ 300 or so neurons. However, enormous amounts of key data needed are still missing. For instance we do not have enough data on the physiology and pharmacology of C. Elegans neurons and synapses. And we still have limited data on the distribution of ion channels, receptors and other proteins on neurons, synapses and glia. Without this data we cannot build unifying models. A second problem is how easy it is to obtain the data. The crucial requirement for unifying models is the ability to access the data needed. Obtaining a deep understanding of the molecular machinery of a single neuron or a single synapse is just as difficult in C. Elegans as in human beings. And many datasets – particularly data on cognition - are actually easier to acquire in rodents, or even in humans. So we can’t just say: “let’s do this quickly in worms and do complex brains later”: we have to solve the same basic challenges, whatever brain we model. What we are actually doing is building a generic strategy we can use to reconstruct any brain.
Relevance: Studying the “simple” nervous systems of organisms like C.elegans or drosophila, is obviously very important, particularly for molecular and genetic studies. However the organization, electrophysiology and function of the mammalian brain are quite different. One of the HBP’s most important goals is to contribute to the development of new treatments for brain disease. But pharmaceutical companies already have great difficulties in translating results from mouse to human beings; with simpler organisms these problems become much worse. If we want to make a real contribution to clinical research, it is probably unwise to invest heavily in simple systems, so distant from the human brain.
I'm not a neuroscientist and may be missing something, but on the surface this seems like a pretty poor argument: citing feasibility as a justification for not going with the (overwhelmingly) simpler organisms first, and going after humans with several orders of magnitude more complex brains. If I were to guess, I'd say that building the unified model of the simplest organisms would be tremendously relevant in helping to model higher-level organisms including humans.
So sure there are low-level details that must be understood for any brain, but come on, 300 vs. 100B?
Once the big projects have cracked the low-level details, someone will figure out C. elegans. Until we have some smaller organisms emulated, don't expect to have a conversation with their emulated human brain, I just can't see that happening out of order.
According to the work plan (not public as far as I know), all the platforms will be open to researchers not just from Europe but from anywhere in the world.
The "Human Brain Project" was also an NIH funding initiative in the 1990s, which was hugely influential in kickstarting the field of neuroinformatics (see http://en.wikipedia.org/wiki/Neuroinformatics)
Okay, so to correct my earlier comment to (hopefully) exclude the earlier project, Google Scholar shows about 105 results for "human brain project" in 2012-2013.
Yes, but one day they will have - hopefully - results. And a commitment to publish in open access journals should have been part and parcel of a funding round like this. A billion is not exactly pocket change.
Henry Markram, the coordinator of the Human Brain Project, is also one of the founders of Frontiers (http://frontiersin.org), a scientist-led, open-access publisher, so I think it's fair to say there is already a strong commitment to open access publishing within the project.
For some projects it is already mandated in the grant agreement. I've heard that this will be extended to all project for Horizon 2020. I am also under the impression, that the EU protects researchers who post the full version of their papers on their websites even when the publisher does not permit it. I could be wrong on the latter.
It'd be interesting if that page listed jobs at partner organisations.. I'd love to be involved with a project like that, but it's difficult to get involved.
Honestly, have you tried simply contacting the project? That's how I got an internship on the Blue Brain Project a few years ago (and in fact have a cameo in the video).
I would suggest checking back after the summer, which is when the project should actually start.
Also note, "The HBP has allocated a large proportion of its funding to what we call Competitive Calls. The HBP Competitive Calls Programme will allow researchers from outside the HBP Consortium to propose research and applications development projects using the HBP platforms and to receive funding from the HBP. Proposals will be evaluated by peer reviewers from outside the Consortium. We expect to see many proposals for research and approaches that the current Consortium has never considered"
and
"From year 3 onwards the HBP will fund a large number of HBP Ph.D studentships and post-doctoral fellowships."
(my understanding is that these are in addition to any postdocs recruited by partner organizations using their individual funding for the project).
I don't necessarily doubt the scientists' competence here, but they're just playing the fundraising game when they talk about "disease" in this context.
I agree that the word "disease" is certainly overused these days, there are still some things that should rightfully use the word.
Besides, whether you call them "diseases" or "disorders" or another term, there are many ways that our neurology can fail us. The "Classification" section of this Wikipedia article lists several major ones, but there's countless more:
