| Home | Table of Contents |

General Intelligence :

Collaborative web platform

version 0.4

Summary:

Basically the platform allows members to make contributions to a collaborative project. Contents are open to all members and membership is granted automatically. A member claims credits for each piece of contribution s/he makes. If the self-claimed amount is "outrageous" other members can complain. The income generated by the product is re-distributed back to the contributors' accounts instantaneously as it is earned. A special agreement specifies how contributed materials may be used outside of the web site. Also, the web site supports branching of the "project tree" which users can browse with a web-based GUI.


  1. Registration
  2. Nondisclosure and noncompetition
  3. Structure of the project tree
  4. Making a contribution
  5. Open-free contributions
  6. Peer rating
  7. Generating income
  8. Virtual credits
  9. "External-use" agreement
  10. Software license

Registration

  1. Each new member signs up with his/her real name.
  2. Membership is granted automatically
  3. [ There will be an online interview session where questions are posed to and answered by the candidate. ]
  4. [ Membership will be granted by majority voting by all existing members. The votes are weighted by each member's amount of contributions. ]
  5. An external method is used to verify the candidate's real-life identity (eg credit card, reference checking, or some form of real-life digital identity).
  6. Every new member is required to sign the following:
    1. nondisclosure agreement
    2. noncompetitive agreement
    3. "external-use" agreement

Nondisclosure and noncompetition

  1. Everyone is allowed to use his/her own contributions in projects outside the platform.
  2. Members are NOT allowed to disclose or use other members' contributions in projects outside the platform.

But members can select a set of items and obtain the rights to use them in external projects by applying to the Managerial Board. See "external-use" agreement.


Structure of the Project Tree

  1. The Project Tree is a tree-like data structure.
  2. The project's goal is "GI" (general intelligence) which is at the top node.
  3. A node may split into alternatives.
    For example: "GI" may split into "logic-based approach" and "neural-based approach".
  4. A node may decompose into subtasks.
    For example: "task A" may be decomposed into "task X" + "task Y" + "task Z" + ...

The difference between alternatives and subtasks is that alternatives represent choices whereas all subtasks must be sequentially completed.

For example:

The "tree view" of this example is like this:

The data structure should be a tree and linked-list hybrid.

There should be a web-based GUI for browsing and modifying this tree structure.

Modifications include:

  1. creating a branching
  2. creating a subtask breakdown
  3. modifying a branching (eg, add branches)
  4. modifying a subtask breakdown (eg, add subtasks)
  5. posting new content to an empty node
  6. modifying the content on a non-empty node

Histories of modifications must be stored and browsable.


Making a contribution

  1. A member makes a contribution (an idea, algorithm, or piece of code) by posting it on a wiki. Source code may be posted as attachments or links.
  2. S/he must provide an estimation of how much the contribution is worth (eg 100 Virtual Credits).
  3. S/he must state whether the idea is original or is taken from somewhere (indicate source).
  4. If the idea/code is derived from some existing contributions, s/he must specify the dependencies.

Open-free contributions

Members can specify certain items to be free-opensource. They will automatically be under an opensource license, except that they are only visible to members of the platform. As long as the disclosure of the opensource items do not violate our protected contents (for example, with references that will reveal protected ideas), they will be published on our publicly-accessible web pages.


Knowledge collection

This allows to collect knowledge about:

  1. common sense
  2. knowledge of programming in various languages (C/C++/C#, Lisp, Prolog, etc)
  3. knowledge of library functions in the above languages

Client-side intelligent agent

An intelligent agent, called G0, will reside on the client side.

It can help:

Cross validation

There is no need for voting on the collected facts, but there is a need for cross-validation by other users.

Cross verification means whether the fact is incorrect or partially correct (correct but has exceptions), or estimate P/Z.

This is done on the site.

Classification

The facts need to be classified into various categories, such as: C/C++/C#.

Maybe this is done on the GUI.

Check for duplications

Done on the site.

Convert English to logic

Can be done on site or on G0.

Storage

For facts, simply throw into a central database, and give the contributor the fact id.


Peer rating

  1. Any member can estimate the values of other members' contributions.
  2. The average values and individual values are visible on the corresponding nodes of the Project Tree
  3. A member can file a complaint if s/he thinks a certain rating is outrageous, by clicking on a "complain" button.

Comments / discussions

Comments are a way for members to discuss the contributions. Comments do not carry any Credits. They appear below the wiki entries (similar to the comments on YouTube).


Generating income

The branches of the project can be downloaded by the public with a fee, thus generating income for the web site.

At the initial stage, customers can pay a fixed fee and download all the branches of the project. Later, when the branches have become sufficiently differentiated, we will allow selective payment and downloading of branches.


Virtual Credits

Re-distribution of income

Income is re-distributed to the corresponding contributors instantly. This makes sense because contributions yet to be made in the future have nothing to do with the income generated at present. The income generated at present should be distributed to members who contributed thus far.

Can Virtual Credits be traded like stocks?

It seems that Virtual Credits are fundamentally different from real stocks in the sense that every member's set of contributions is unique, and they generate income differently according to which branch(es) they belong to. Therefore the value of a contributor's portfolio is dependent on its contents. Real stocks, on the other hand, are essentially interchangeable.

In principle, Virtual Credits can be traded in a market, but we'll explore this possibility later.


"External use" agreement

This agreement states that, if a member decides to use other members' contributions in an external project, s/he must pay our company a percentage of his/her new project's revenue. The percentage will depend on the selected items, and will be decided by the managerial board.

A member can always use his/her own contributions in external projects, without payment.


Cash funding from Investors

The collaborative platform can accept cash funding from investors.

First, the investor needs to register as a member.

The investment mechanism is as follows:

  1. An investor invests $W on the project;
  2. In return, the investor gets a percentage, %P, of future revenues of the project.
  3. $W will be re-distributed to Members over a period of time, following a scheme where $W is partly paid to already-contributed Members and partly reserved for future contributors.

Note: Currently, all investments will be made on the entire project (all branches), but later we may allow investors to make branch-specific investments. When we allow that, re-distribution of funds will be handled accordingly.


Software license

I'm still preparing it...


| Home | Table of Contents |

6/Jan/2008, 24/Sept, 17/Jun/2007 (C) GIRG

Hosted by www.Geocities.ws

1