Hello Team Canada,

I would like to introduce myself and the project that I am leading. My name is Ovidiu Anghelidi and I am working on the Artificial Intelligence System. A couple of users from Team Canada have noticed that first, this is a BOINC project and second, it is Canadian. They would like to support it, but because the project is started by a private company and aims to utilize public resources they have raised the issue of resource utilization and have asked us for clarification on that. I think that is a fair question and we would like to briefly explain what we try to ccomplish so that everyone can make an informed decision in regards to participating in the project.

First here is a bit of background on a similar project that proved that private and public resources can be joined together and bring about successful results for everyone involved. Many of you are maybe familiar with the recently started Blue Brain project (e.g. http://en.wikipedia.org/wiki/Blue_brain), a computational neuroscience collaboration effort between IBM and a laboratory from University of Lausanne, Switzerland. Computational neuroscience is the field that studies the computational aspect of the brain and artificial intelligence has its deep roots in this field.

The project was a huge success because it showed that first, large scale neuronal simulations can be performed and second that important research advances are within our reach and are not constrained by lack of funding on either side, public or private. This collaborative approach has benefited both IBM and the University. It helped IBM prepare and understand what are the requirements and issues arising from running large scale brain simulations, the bedstep of future artificial intelligence solutions. The University gained worlwide recognition and will certainly be able to attract skillful scientists and make their Brain and Mind Institute an important research center for the neuroscience community. We look forward to that.

IBM is not the only company though, that is looking to participate in large scientifical research projects; in the near future we expect to see many other companies involved in this type of collaborative efforts. The Boinc home page is also encouraging in this aspect. Over the next couple of years, our aim is to reach the same level of development and large-scale
computations and then to further push this limit, but in order to do that we need the support of the BOINC community.

Our position is that it does not matter who is going to implement Artificial Intelligence, if it is going to be a company or a University laboratory or other organization; that is because, after we reach that point many things will positively change.

In the same way that computers changed our lives and improved them, AI will do the same. Unfortunately we do not see that many projects like this one being built, but that will hopefully change. In the meantime we will continue our advancements. We have started development more than one year ago, as soon as we understood how knowledge is represented in the brain and we have made available the work in the form of the source code for the simulator which can be downloaded using the link from the project's home page (i.e. http://www.intelligencerealm.com/aisystem/system.php). We will also publish results and source code and continue to share information because that it is only fair when building something with public resources even though we are a small company.

Please note that we are also advanced in our research and that fundamental knowledge that we acquired prior to the start of this project will not be made available to the public. That is only fair, too.

I hope this provides answers to some of the questions raised and helps everyone in making informed decisions.

Thank you.

Your support is appreciated.

Best regards,
Ovidiu Anghelidi

Thanks for the info Ovidiu, it certainly answered my questions.

Thanks for taking the time to post and explain your position.

I'm not trying to start a big debate on this one (although that may be an unfortunate result), but one thing I have a bit of an issue with is your reference of Univeristy institutions as "public" as it relates to the "public" in general. The biggest factor I see is that, as in the case of the University in Lusanne, the Blue Brain project provided a lot of public attention which would then be converted into money in the way of extra grants/donations/funding/etc for futher research and development. This certainly doesn't apply to the average BOINC user and their home PC - they will not personally recieve any public fame or additional cash in the pocket as a result of assisting this project (unlike a University). (I skip by the whole private funding for University research aspect as that might get us a little off topic!).

That being said, a lot of the projects currently in BOINC are University driven and have the same underlying issues as mentioned above, and several others are even purely private company driven, and it doesn't stop people who are interested in the science (or points) from crunching happily away. As long as the science is worthwhile, you shouldn't find too much trouble getting people on board to assist in your efforts (especially if you respond quickly to issues and keep the public informed with frequent progress updates!).

I'm still not entirely sure what your stance on information sharing information, in particular the level of sharing involved (ie. what level of sharing will be involved with the results/knowledge obtained), but I'll more than likely be crunching as it's a topic I've always been somewhat fascinated with.

Thanks for taking the time to give us some info, hopefully it'll answer questions many were thinking about.
Oddly enough, this reminds me how how Seti@Home came around. Ya ya, i know, its not the same exact thing but the parrallel between the two are there... arent they?

Hmm, need more coffee apparently.

Thanks for the info. Count me in too. Just give us a buzz when your ready and will be there.

Thank you.

I don't know how Seti@Home started but David did a good job in there and now with Boinc.

Debates are interesting but unfortunately they take time.



You are right. I can only thank to the future participants.

(I skip by the whole private funding for University research aspect as that might get us a little off topic!).



That is mainly because companies that do have resources are not interested in building solutions on Open Source software like the BOINC platform.



I wasn't aware of that but it is a normal evolution of the platform. I saw that on the Boinc home page companies are also invited to use the platform.



The vast majority of AI researchers are into the field of AI and some in other fields, like math, physics, or in our case computational neuroscience. Each field has different models and theories and everyone favors what they understand best.



You are right to point that but you should also know that data sharing is a contentious issue in many fields of science. Mainly because science is very competitive and funds are scarce. Here is a brief perspective: if a researcher will publish everything he/she knows then there is a good chance that someone else might get to reuse their results and be able to obtain funds to continue to explore that result. That would not be beneficial to the person who put a lot of work into that and got results because it can take away their edge. And what would be the guarantee that other people will actually be able to explore that in depth? They may get funds based on that but funds do not equal results. On the other hand if someone is not willing to present results and keeps everything for themselves than the community is not helped at all. Some other important discoveries might take longer to be found based on the lack of bridging results. We are on the opinion that a balanced approach is needed and in our case we will post results and source code that we feel are not of vital importance to us but that are still very valuable to the community and can can help others. Take for example the source code that we posted on SourceForge. Inside there is a database structure that allows everyone to build and access quickly large numbers of neurons using an embedded database on a single machine. Hundreds of millions of neurons or more. That is new. None of the other major simulators (e.g. Neuron, Genesis) store information this way and in the Blue Brain project that was employing Neuron almost a single neuron was stored per machine; 10,000 neurons on 8,000 nodes. Highly inefficient. This information is going to be very valuable for many researchers while for us it is not vital. So we shared it so that others can reuse it.
You can search for data sharing on the Internet and you should be able to find many papers on this issue.



AI is fascinating indeed in possibilities.

Thank you.

It's not just the time, they tend to get off track on forums and spiral off into nowhere quickly..



I'm well aware of the various intricacies of data sharing and agree there definitely has to be a balance in order for both economic and scientific progress reasons (I've been closely associated with both master/phd level University research projects as well as intelectual property law in the past and present). I can't personally see too many projects surviving without withholding some level of proprietary information. You seem to be taking a fairly balanced approach in this respect and that's good to see. I'm looking forward to the project initiation and wish you guys luck in the future (I'll be crunching for you anyway)...

Ovidiu,

Thank You for taking the time and making the effort to educate us about the new project. I've read and digested what you and others have written and feel confident that Team Canada's resources will be put to good use.

When the time comes, we will "crunch" and spread the word for your project.

Thank You again,
Paul

PS - Seeing as Dr. David Anderson joined up with Team BOINC Synergy, Maybe you could join Canada??? 8^D That would be a nice way of "Paying" us back! HAHAHA! http://www.intelligencerealm.com/aisyst ... teamid=116

Thanks again.

Thanks Pawly.



I joined the team but because we maintain a separate copy of the project where we test new applications we use the machines mostly in there. Nonetheless it feels good to be part of the team.

Excellent! Welcome to the team!

Welcome aboard ovidiu.

Great to have you aboard Ovidiu!!

Hope this wont sount "too" repetitive but, Welcome aboard!

FYI - there have been a few wus come down the pipeline recently. They only have a Linux client for now (Windows coming soon apparently), and the wus were somewhat buggy (I only got credit for 1 of 5 downloaded), but at least it's some progress. Combine really large downloads (40MB - they're reducing this to 20MB for the next round) and short run times (800-850s on my PentiumD) they still have some issues to iron out, but I encourage all our Linux crew to attach and help out with their troubleshooting when the next round of wus hit.

We are 13th by RAC:
http://www.intelligencerealm.com/aisystem/top_teams.php


There is work for Linux.

We have moved up to 11th by RAC!

Good work team.

C'moooooooooooonnnnnnnnnnnnnnn Windows App!!

Hey Pawly, You can always install Linux as a second operating system on your computer. You can even install it to a flash drive.

You could even vritualize it, but you'd take a bit of a performance hit...

I'm hoping for a new computer soon, when that happens the Linux experiments will start here in my basement!

PS3?

Naw, my girlfriend would KILL me! lol Says I spend too much time "playing" on the computer as it is. I have my eye on a few computers that people will be needing to vacate their houses when they upgrade. I'll keep y'all informed.