Take the 2-minute tour ×
Cognitive Sciences Stack Exchange is a question and answer site for practitioners, researchers, and students in cognitive science, psychology, neuroscience, and psychiatry. It's 100% free, no registration required.

I want to build a system that has the ability to gather data from the internet in order to build a cognitive model of the human brain. The model should be able to answer the questions required by a robot. Also, I would like to apply an AI inference engine to get more important information to the robot.

If I want to do build such a model, what "skill list" do I need? What skills would be most important? What "opensource project" might be helpful?


added:

Thank everyone who encourage me. You let me have more brave to build a human brain mind model(not brain structure).

Let me introduce myself. I have some CS and EE background and now research on robotics area. I love cognitive science so much so I want to figure out how to let robot really have the mind like human to achieve goal more clever. Thank the hint of @Jeromy Anglim, I give a video youtube example:James and Rosie Preparing Popcorn and Sandwiches

That video shows two robot cooperate together to make a sandwich with cognitive architecture(CRAM or Cogito) based on ROS(Robot OS,ROS map,ROS Navigation Video) system. After reading their papers I found that they just using database or downloading web page instructions or using very well structured pages which are for human (ex.CYC,OpenMind Indoor Common Sense database,Semantic web,ORO,SUMO,Wikipedia,google images,germandeli).

Also I found that many free resources,any article on WWW should be most important database. The most important thing is how to build a system that can read natural language and have a model (maybe based on CRAM cognitive architecture) to understand it,because CRAM has a very good knowledge processor(KNOWROB,M. Tenorth and M. Beetz, “KnowRob — Knowledge Processing for Autonomous Personal Robots,” in IEEE/RSJ International Conference on Intelligent RObots and Systems., 2009.). If perhaps, I want to collect all linux opencourse libraries to make it more fast and robust. I also begin to take online machine learning course on stanford.

My goal is to build a human mind model can learn,have some innovation,and find intelligent solutions to dedicate itself to human society.But my first step is to find a way that robot can try to understand all kinds of data on the WWW (any data that it can't understand just warn the first 10,and ignore others), and structure them as an intelligent interaction system which can answer knowledge or skill questions which can be search on google. The most important of that system is try to use all information on the web clever with robust libraries( I don't want to reinvent wheels).

I know it is a long term project, but I loving and willing to achieve in my whole life. Because this is an integration project and any success in that will make a big change in the world just as ROS in robotics also an integration project and very active on most top robotics research centers and in the beginning on industry now. I know My own energy is not enough for this purpose, but I can gather all the power of opensource and volunteers in the world to do it together~^^ (Now is the day which techs are really enough for this project like computer speed,WWW,advanced research papers,opensourse libraries,free opencourseware lessons,..., what a lack is a group of excellent people volunteer to make it become true to help anyone who use this project.)

share|improve this question
1  
Hi Sam. Welcome to the site. Could you clarify more precisely what your goal is, perhaps with some references to examples on the web of something similar? –  Jeromy Anglim Feb 6 '12 at 3:00
6  
Suggested requirements: a) a post-graduate degree in cogsci or related field and b) several years experience programming AI –  Jeff Feb 6 '12 at 3:16
    
Do you want to build a model of the brain or the model of the mind? These might be quite distinctive goals ... –  Felix S Feb 6 '12 at 8:04
    
Wow that is one big question. –  Oriesok Vlassky Oct 30 '12 at 11:19
add comment

closed as not a real question by Artem Kaznatcheev, Ben Brocka, Steven Jeuris Apr 3 '12 at 23:43

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.If this question can be reworded to fit the rules in the help center, please edit the question.

2 Answers

"how to build a system that can read natural language and ... understand it"

I am not sure you appreciate what you are wanting here. Just this one piece requires an understanding of the processes of the mind that I don't believe we are at yet. There is a lot of material about how we MIGHT achieve this, but you would need to appreciate the whole cognative processing model ( or rather, find one of the models that you like best ), and be able to mirror it.

There is also, I think a requirement in here to automatically search the web and identify good quality answers. Well, if you do this, I am sure that Google will be very interested in you. Most people are poor at doing this reliably, so trying to enable a computer system to achieve this is probably not practical at this moment.

I think the most important thing you need is an entire research department. The requirement you have is so diverse that it needs many different people working on many different areas. You would need, I woudl say, PhDs in cogsci or similar, and AI, as a minimum.

This all sounds very negative, but actually, I think it is a good question, because it is important to understand the scope of the answer, for everyone who works in machine intellegence.

share|improve this answer
    
Your answer make a great thing that localized of my project's range. I appreciated it, and if you will, I still need more and more actual help to figure out how to make that project very robust and useful in the world to help everyone in now and the future. If I have ability to achieve it, I will upgrade this project in my whole life. –  sam Feb 6 '12 at 13:31
    
I have also added great video link which on robotics areas on my origin post. –  sam Feb 6 '12 at 13:42
    
Another brand new idea is: I want to integrate all the research results means any research center in the world will become my best friend and no enemy or competition. –  sam Feb 6 '12 at 13:57
    
That was sort of my point. Pretty much each piece of this is a lifetimes work. I picked on a piece that I knew something about, but the same applies to every other piece. Each one needs a genius and a few lifetimes. –  Schroedingers Cat Feb 6 '12 at 14:41
add comment

Schroedingers Cat nailed a couple of important points and I need to expand it further.

I believe the important roadblock in the creation of an absolutely intelligent machine is the fact that human brain is still not decrypted. There isn't a holistic cognitive model that describes how the brain functions, reacts or take decisions. Even in case such a model is inferred, it still cannot be put into real use since human decision is based on knowledge and the environment (popularly called culture). So a machine being rational is hard to prove since human themselves cannot prove their rationality.

Technically, there are three problems

  1. Language - Almost all programs use one or the other form of Context Free Grammar. Humans always use context while speaking.
  2. Computationally it isn't feasible to do a simulation model similar to brain. Just assume a choice of 2 items among a set of 10 items. You have a possible 2^10 combinations. Think of 100 conditions and your computer would go into limbo.
  3. Most of the AI engines built today are a far cry from those proposed by their forefathers. You must have read the AI: A modern approach. They are mostly built for a particular domain and aren't absolutely intelligent

And regarding your robot searching Google for the answers, its already possible by using the Google API's to search for answers but it won't fulfill your requirement. Google uses the PageRank technology for its results, which is more of a machine learning technique and is still far from true semantic search.

There's hope though. I am just in the same situation (coincidentally I have subscribed for both machine learning and natural language processing, though all courses are indefinitely postponed). I could suggest learning the following

  1. Artificial Neural Networks
  2. Natural Language Processing
  3. Probabilistic Models
  4. Unsupervised machine learning techniques
  5. Advanced pattern discovery

I also suggest a change in your original model. Instead of searching Google to find the answers, it would be better you download the stack exchange data dump and start analyzing the patterns since those answers are given by true humans.

Feel free to ask more questions. There are more things to be questioned than to be answered in this field.

share|improve this answer
    
I think "reading natural language" is enough of a challenge. Listening to it and interpreting it is a whole new level of problem. As you say, humans use context to an incredible degree. We also use multi-media - that is, more than just sounds, but faces and actions. They all have to be "read". –  Schroedingers Cat Feb 6 '12 at 16:05
    
Actually, this gives me an idea of a good starting point. Download the SE data dump, and write an interface to answer any question that SE already provides the answer to. That in itself woudl be a stunning achievement. And I mean interpreting questions, so that if I ask something that has been answered in a different way from a different question, it will work it out. –  Schroedingers Cat Feb 6 '12 at 17:12
    
@sam I am also thinking about creating a research framework with a mass simulation engine that could resemble a lot of your characteristics. Feel free to ping me or send a mail for further information –  Ubermensch Feb 7 '12 at 15:19
    
@SchroedingersCat Spot on. On top of it, to build a truly intelligent machine, it must perceive things in the right sense. So what is right brings into the equation a philosopher –  Ubermensch Feb 7 '12 at 15:22
    
@SchroedingersCat And regarding reading natural language, we can cut down a few (infact a lot) of the original language comprehensions to make it simpler. NLTK is a real good starting point –  Ubermensch Feb 7 '12 at 15:24
show 3 more comments

Not the answer you're looking for? Browse other questions tagged or ask your own question.