Question Answering with Time-Based Weighting

Patent metadata

Field	Value
Publication number	US-2016239740-A1
Application number	US-201514624030-A
Country	US
Kind code	A1
Filing date	Feb 17, 2015
Priority date	Feb 17, 2015
Publication date	Aug 18, 2016
Grant date	—

Abstract

A mechanism is provided in a data processing system for question answering using time weighted evidence. The mechanism receives an input question. The mechanism determines a time focus for the input question and defines a weighting function. The weighting function is a bell curve having a peak at the time focus on a time axis. The mechanism decomposes the input question into one or more queries and applies the one or more queries to a corpus of information to obtain a set of hypothesis evidence. Each item of information within the hypothesis evidence has an associated time value. The mechanism weights the set of hypothesis evidence based on the associated time values according to the weighting function to form time weighted evidence and generates hypotheses for answering the input question based on the time weighted evidence.

First claim

What is claimed is:

1 . A method, in a data processing system, for question answering using time weighted evidence, the method comprising:
receiving, by the data processing system, an input question; determining, by the data processing system, a time focus for the input question; defining, by the data processing system, a weighting function, wherein the weighting function is a bell curve having a peak at the time focus on a time axis; decomposing, by the data processing system, the input question into one or more queries; applying, by the data processing system, the one or more queries to a corpus of information to obtain a set of hypothesis evidence, wherein each item of information within the hypothesis evidence has an associated time value; weighting, by the data processing system, the set of hypothesis evidence based on the associated time values according to the weighting function to form time weighted evidence; and generating, by the data processing system, hypotheses for answering the input question based on the time weighted evidence.

2 . The method of claim 1 , wherein determining the time focus for the input question comprises determining that a majority of the set of hypothesis evidence have associated time values in a given time period.

3 . The method of claim 2 , wherein determining the time focus for the input question further comprises identifying the given time period as the time focus.

4 . The method of claim 2 , wherein determining the time focus for the input question further comprises:
prompting a user to select the given time period as the time focus; and responsive to the user selecting the given time period as the time focus, identifying the given time period as the time focus.

5 . The method of claim 1 , wherein the weighting function is a Gaussian function.

6 . The method of claim 1 , further comprising determining a standard deviation of the bell curve based on user input.

7 . The method of claim 1 , wherein the time focus is a first time focus for a first dimension of evidence, the method further comprising:
determining a second time focus for a second dimension of evidence; and defining the weighting function, wherein the weighting function is a two-dimensional Gaussian function.

8 . The method of claim 1 , wherein determining the time focus for the input question comprises:
determining a plurality of time foci for the input question; defining a weighting function for each of the plurality of time foci, wherein each weighting function is a bell curve having a peak at a given time focus within the plurality of foci; weighting the set of hypothesis evidence according to each weighting function; aggregating results for the plurality of time foci; and establishing a final weighting function along a time axis according to the aggregated results.

9 . A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device, causes the computing device to:
receive an input question; determine a time focus for the input question; define a weighting function, wherein the weighting function is a bell curve having a peak at the time focus on a time axis; decompose the input question into one or more queries; apply the one or more queries to a corpus of information to obtain a set of hypothesis evidence, wherein each item of information within the hypothesis evidence has an associated time value; weight the set of hypothesis evidence based on the associated time values according to the weighting function to form time weighted evidence; and generate hypotheses for answering the input question based on the time weighted evidence.

10 . The computer program product of claim 9 , wherein determining the time focus for the input question comprises determining that a majority of the set of hypothesis evidence have associated time values in a given time period.

11 . The computer program product of claim 10 , wherein determining the time focus for the input question further comprises identifying the given time period as the time focus.

12 . The computer program product of claim 10 , wherein determining the time focus for the input question further comprises:
prompting a user to select the given time period as the time focus; and responsive to the user selecting the given time period as the time focus, identifying the given time period as the time focus.

13 . The computer program product of claim 9 , wherein the weighting function is a Gaussian function.

14 . The computer program product of claim 9 , wherein the computer readable program further causes the computing device to determine a standard deviation of the bell curve based on user input.

15 . The computer program product of claim 9 , wherein the time focus is a first time focus for a first dimension of evidence, wherein the computer readable program further causes the computing device to:
determine a second time focus for a second dimension of evidence; and define the weighting function, wherein the weighting function is a two-dimensional Gaussian function.

16 . The computer program product of claim 9 , wherein determining the time focus for the input question comprises:
determining a plurality of time foci for the input question; defining a weighting function for each of the plurality of time foci, wherein each weighting function is a bell curve having a peak at a given time focus within the plurality of foci; weighting the set of hypothesis evidence according to each weighting function; aggregating results for the plurality of time foci; and establishing a final weighting function along a time axis according to the aggregated results.

17 . An apparatus comprising:
a processor; and a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor, cause the processor to:
receive an input question;
determine a time focus for the input question;
define a weighting function, wherein the weighting function is a bell curve having a peak at the time focus on a time axis;
decompose the input question into one or more queries;
apply the one or more queries to a corpus of information to obtain a set of hypothesis evidence, wherein each item of information within the hypothesis evidence has an associated time value;
weight the set of hypothesis evidence based on the associated time values according to the weighting function to form time weighted evidence; and
generate hypotheses for answering the input question based on the time weighted evidence.

18 . The apparatus of claim 17 , wherein the weighting function is a Gaussian function.

19 . The apparatus of claim 17 , wherein the time focus is a first time focus for a first dimension of evidence, wherein the instructions further cause the processor to:
determine a second time focus for a second dimension of evidence; and define the weighting function, wherein the weighting function is a two-dimensional Gaussian function.

20 . The apparatus of claim 17 , wherein determining the time focus for the input question comprises:
determining a plurality of time foci for the input question; defining a weighting function for each of the plurality of time foci, wherein each weighting function is a bell curve having a peak at a given time focus within the plurality of foci; weighting the set of hypothesis evidence according to each weighting funct

Assignees

• IBM

Inventors

• Baughman Aaron K
• Diamanti Gary F
• Marzorati Mauro
• Valletti Elizabeth M

Classifications

• G06N20/00

  Machine learning · CPC title

• G06N5/04 Primary

  Inference or reasoning models · CPC title

• G06N99/005

  Physics · mapped topic