Processing apparatus, processing method, estimating apparatus, estimating method, and program

Patent metadata

Field	Value
Publication number	US-11227228-B2
Application number	US-202016792472-A
Country	US
Kind code	B2
Filing date	Feb 17, 2020
Priority date	Dec 18, 2014
Publication date	Jan 18, 2022
Grant date	Jan 18, 2022

Abstract

A processing apparatus is disclosed for representing cognitively biased selection behavior of a consumer as a learnable model with high prediction accuracy taking into account even feature values of a product and the consumer. The processing apparatus generates a selection model obtained by modeling selection behavior of a selection entity that selects at least one choice out of presented input choices. The processing apparatus includes an acquiring unit to acquire training data including a plurality of input feature vectors that indicate features of a plurality of the choices presented to the selection entity and an output feature vector that indicates a feature of an output choice. The processing apparatus further includes an input combining unit to combine the plurality of input vectors to generate an input combined vector, and a learning processing unit to learn a selection model on the basis of the input combined vector and the output vector.

First claim

What is claimed is:

1. An information processing system comprising:
memory;
a processor communicatively coupled to the memory; and
a processing apparatus communicatively coupled to the memory and the processor, the processing apparatus comprising:
an acquiring unit configured to acquire training data including
a plurality of input feature vectors that each indicate an input choice of a first plurality of input choices and one or more features of the input choice, wherein the first plurality of input choices is presented to a selection entity in a choice set from a plurality of choice sets, and
a plurality of output feature vectors each indicating at least one output choice and one or more features of the at least one output choice, wherein each of the at least one output choice is an input choice selected by the selection entity from one choice set of the plurality of choice sets;

an input combining unit configured to combine the plurality of input feature vectors to generate an input combined vector;
a learning processing unit configured to learn a selection model based on the input combined vector and the plurality of output feature vectors, wherein the selection model models how a fraction of selection rates of choices included in input choices change based on a combination of other choices included in the input choices and one or more feature values of the input choices; and
an estimating unit programmed with the selection model and configured to determine a combination of input choices from a second plurality of input choices that increases a probability that the selection entity will select a desired output choice included within the second plurality of input choices.

2. The information processing system according to claim 1 , wherein the processing apparatus further comprises:
an output combining unit configured to combine the plurality of the output feature vectors to generate an output combined vector, wherein
the learning processing unit learns the selection model on the basis of the input combined vector and the output combined vector.

3. The information processing system according to claim 2 , wherein the selection model comprises a plurality of input nodes corresponding to a plurality of input feature values included in the input combined vector, a plurality of output nodes corresponding to a plurality of output feature values included in the output combined vector, and a plurality of intermediate nodes, and
the learning processing unit learns a plurality of weight parameters provided between the plurality of input nodes and the plurality of intermediate nodes and between the plurality of output nodes and the plurality of intermediate nodes.

4. The information processing system according to claim 2 , wherein each of the plurality of input feature vectors and the plurality of output feature vectors is a binary vector.

5. The information processing system according to claim 4 , wherein the acquiring unit comprises a binarizing unit configured to binarize feature parameters of
choices of each of the first plurality of input choices, and
each of the output choices, wherein the binarizing unit is further configured to generate a feature vector of each of the plurality of input feature vectors and each of the plurality of output feature vectors.

6. The information processing system according to claim 5 , wherein the binarizing unit converts feature parameters closer to each other into feature vectors closer to each other.

7. The information processing system according to claim 2 , wherein the input combining unit generates, for the first plurality of input choices, the input combined vector that takes the same value without depending on order of the input choices, and
the output combining unit generates, for the output choices, the output combined vector that takes the same value without depending on order of the output choices.

8. The information processing system according to claim 7 , wherein the input combining unit and the output combining unit generate the input combined vector and the output combined vector based on an average of the plurality of input feature vectors and an average of the plurality of output feature vectors.

9. The information processing system according to claim 7 , wherein the acquiring unit acquires the plurality of input feature vectors and the plurality of output feature vectors, which are binary vectors, and
the input combining unit and the output combining unit generate the input combined vector and the output combined vector based on a logical operation of the plurality of input feature vectors and a logical operation of the plurality of output feature vectors.

10. The information processing system according to claim 9 , wherein the input combining unit and the output combining unit generate the input combined vector and the output combined vector based on logical OR of each of elements of the plurality of input feature vectors and logical OR of each of elements of the plurality of output feature vectors.

11. The information processing system according to claim 1 , the processing apparatus further comprising:
a selecting unit configured to select an output choice from the second plurality of input choices as the desire output choice.

12. A method comprising:
acquiring training data comprising
a plurality of input feature vectors that each indicate an input choice of a first plurality of input choices and one or more features of the input choice, wherein the first plurality of input choices is presented to a selection entity in a choice set from a plurality of choice sets, and
a plurality of output feature vectors each indicating at least one output choice and one or more features of the at least one output choice, wherein each of the at least one output choice is an input choice selected by the selection entity from one choice set of the plurality of choice sets;

combining the plurality of input feature vectors to generate an input combined vector;
learning a selection model based on the input combined vector and the plurality of output feature vectors, wherein the selection model models how a fraction of selection rates of choices included in input choices change based on a combination of other choices included in the input choices and one or more feature values of the input choices; and
determining, utilizing the selection model, a combination of input choices from a second plurality of input choices that increases a probability that the selection entity will select a desired output choice included within the second plurality of input choices.

13. The method according to claim 12 , further comprising:
combining the plurality of the output feature vectors to generate an output combined vector, wherein the selection model is learned on the basis of the input combined vector and the output combined vector.

14. The method according to claim 13 , wherein the selection model comprises a plurality of input nodes corresponding to a plurality of input feature values included in the input combined vector, a plurality of output nodes corresponding to a plurality of output feature values included in the output combined vector, and a plurality of intermediate nodes, and
wherein the learning further comprises learning a plurality of weight parameters provided between the plurality of input nodes and the plurality of intermediate nodes and between the plurality of output nodes and the plurality of intermediate nodes.

15. The method according to claim 13 , wherein each of the plurality of input feature vectors and the plurality of output feature vectors is a

Assignees

• IBM

Inventors

• Morimura Tetsuro
• Osogami Takayuki
• Otsuka Makoto

Classifications

• G06N7/01 Primary

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N7/005 Primary

  Physics · mapped topic