Artificial neural network guessing method and game

• US 20060230008A1
• Filed: 04/08/2005
• Published: 10/12/2006
• Est. Priority Date: 04/08/2005
• Status: Abandoned Application

• Current Assignee: 20Q.net Inc.
• Original Assignee: 20Q.net Inc.
• Inventors: Burgener, Robin
• Application Number:
  US11/102,105
• Publication Number:
  US 20060230008A1
• Time in Patent Office: Days
• US Class Current: 706/25
• CPC Class Codes: G06N 3/02 Neural networks
• CPC Class Codes: G09B 7/02 of the type wherein the student is expected to construct an answer to the question which is presented or wherein the machine gives an answer to the question presented by a student

Abstract

A method for guessing, in an electronic game, an object that a user is thinking of, from a set of target objects, after asking the user at least one question, the method utilizing a neural network structured in a target objects-by-questions matrix format, wherein each cell of the matrix defines an input-output connection weight, and the neural network can be utilized in a first mode, whereby answers to asked questions are input nodes and the target objects are output nodes, and in a second mode, whereby the target objects are input nodes and the questions are output nodes, the method comprising the steps of ranking the target objects by utilizing the neural network in the first mode; ranking the questions by utilizing the neural network in the second mode; and providing a guess in accordance with the ranking of the target objects.

Claims

1. A method for guessing, in an electronic game, an object that a user is thinking of, from a set of target objects, the method comprising:

asking the user at least one question;

utilizing a neural network structured in a target objects-by-questions matrix format, wherein each cell of the matrix defines an input-output connection weight, and the neural network can be utilized in a first mode, whereby answers to asked questions are input nodes and the target objects are output nodes, and in a second mode, whereby the target objects are input nodes and the questions are output nodes;

ranking the target objects by utilizing the neural network in the first mode;

ranking the questions by utilizing the neural network in the second mode; and

providing a guess in accordance with the ranking of the target objects.

2. The method of claim 1 wherein ranking the target objects comprises:

mapping at least one answer to a weight;

comparing the weight of the answer to weights of cells in the neural network corresponding to that question and the target objects being ranked, and temporarily changing weights of corresponding cells in accordance with agreeability; and

rating the target objects in accordance with the changed cell weights.

3. The method of claim 1 wherein ranking the questions comprises:

mapping predictable answers to questions to positive and negative weights with respect to target objects highly ranked;

totaling all agreeable weights and all disagreeable weights for each question, and computing a margin between the agreeable weights and disagreeable weights totals for each question; and

rating the questions in accordance with the margins.

4. The method of claim 1 wherein ranking the questions comprises:

computing, for each question, margins between a weight of a cell corresponding to a most highly ranked target object and that question, and weights corresponding to other highly ranked target objects and that question; and

rating the questions by comparing the margins of each question with the margins of other questions.

5. The method of claim 1 further comprising adjusting weights of cells corresponding to a guessed object in accordance to agreeability between a mapped weight of an answer and the cell weight before adjusting.

6. The method of claim 5 further comprising:

classifying the user according to user-specific information;

storing the adjusted weights in a database associated with a class of the user; and

using the associated database with a different user that belongs to the class of the user.

7. The method of claim 6 wherein the user-specific information is acquired from the user.

8. The method of claim 6 wherein the user-specific information is inferred from at least one answer to said at least one question.

9. A neural network comprising:

an X-by-Y matrix;

a plurality of cells in the matrix, wherein each cell of the matrix defines an input-output connection weight, and the neural network can be utilized in a first mode, whereby elements of the X-axis are input nodes and elements of the Y-axis are output nodes, and in a second mode, whereby the elements of the Y-axis are input nodes and the elements of the X-axis are output nodes.

10. A game for guessing an object that a user is thinking of, from a set of target objects, after asking the user at least one question, the game comprising:

a neural network structured in a target objects-by-questions matrix format, the neural network having a plurality of cells wherein each cell of the matrix defines an input-output connection weight, and the neural network can be utilized in a first mode, whereby answers to asked questions are input nodes and the target objects are output nodes, and in a second mode, whereby the target objects are input nodes and the questions are output nodes;

means for ranking the target objects by utilizing the neural network in the first mode;

means for ranking the questions by utilizing the neural network in the second mode; and

means for providing a guess in accordance with the ranking of the target objects.

11. A method comprising:

bearing representations from a computer-readable medium of instructions and data;

causing a computer to perform a method for guessing an object that a user is thinking of, from a set of target objects, after asking the user at least one question;

utilizing a neural network structured in a target objects-by-questions matrix format, wherein each cell of the matrix defines an input-output connection weight, and the neural network can be utilized in a first mode, whereby answers to asked questions are input nodes and the target objects are output nodes, and in a second mode, whereby the target objects are input nodes and the questions are output nodes;

ranking the target objects by utilizing the neural network in the first mode;

ranking the questions by utilizing the neural network in the second mode; and

providing a guess in accordance with the ranking of the target objects.

12. The method of claim 12, wherein the computer is a hand-held device.