# US 2019/0171726 A1 · Suggesting follow up questions from user behavior

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0171726 A1
- Document ID: US20190171726A1
- Status: Published application
- Kind: A1
- Application number: 15/829,601
- Filed: 2017-12-01
- Published: 2019-06-06
- Earliest claimed: 2017-12-01
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Mohamed N. Ahmed; Charles E. Beller; WILLIAM G. DUBYAK; Palani Sakthi; Kristen M. Summers; Andeep S. Toor
- Web page: https://patentgrep.com/patent/US20190171726A1

## CPC

- G06F 16/90324: {using system suggestions}
- G06F 16/90332: {Natural language query formulation or dialogue systems}
- G06F 16/9038: Presentation of query results
- G06F 16/9535: Search customisation based on user profiles and personalisation
- G06N 20/00: Machine learning
- G06N 5/02: Knowledge representation; Symbolic representation
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An embodiment of the invention may include a method, computer program product and system for generating follow-up questions based on machine learning utilizing a computing device. The embodiment may include receiving an input question from a user. The embodiment may include parsing the received input question to extract input question components. Parsing utilizes natural language processing techniques. The embodiment may include executing trained question component models to predict follow-up question components. The extracted input question components are utilized as inputs to the trained question component models. The embodiment may include combining the predicted follow-up question components to generate one or more follow-up questions. The embodiment may include returning the one or more follow-up questions to the user.

## Claims

### Claim 1

A method for generating follow-up questions based on machine learning utilizing a computing device, the method comprising:
receiving an input question from a user;
parsing the received input question to extract input question components, wherein parsing utilizes natural language processing techniques;
executing trained question component models to predict follow-up question components, wherein the extracted input question components are utilized as inputs to the trained question component models;
combining the predicted follow-up question components to generate one or more follow-up questions; and
returning the one or more follow-up questions to the user.

### Claim 2 (depends on 1)

The method of claim 1, wherein the trained question component models are selected from the group consisting of a model for lexical answer type prediction, a model for entity type prediction, a model for question type prediction, and a model for question relationship type prediction.

### Claim 3 (depends on 1)

The method of claim 1, wherein the trained question component models are created utilizing training data generated by an expert user, and wherein the training data comprises a sequence of questions asked by the expert user, and wherein the sequence of questions is relevant to a single line of inquiry, and wherein the sequence of questions is relevant to a single line of inquiry if it satisfies a criterion selected from the group consisting of a topic similarity criterion and a graph connection criterion.

### Claim 4 (depends on 3)

The method of claim 3, wherein the training data generated by the expert user is parsed, utilizing natural language processing techniques, to extract training data question components, and wherein machine learning algorithms are applied to the training data question components to determine training data question component patterns, and wherein the training data question component patterns are utilized to train the trained question component models.

### Claim 5 (depends on 4)

The method of claim 4, wherein, the input question components, the follow-up question components, and the training data question components comprise an element from the group consisting of an entity type, a relationship type, a question type, and a lexical answer type.

### Claim 6 (depends on 1)

The method of claim 1, wherein the one or more follow-up questions are in a same grammatical form as the input question.

### Claim 7 (depends on 1)

The method of claim 1, further comprising:
ranking the one or more follow-up questions from most useful to least useful.

### Claim 8

A computer program product for generating follow-up questions based on machine learning utilizing a computing device, the computer program product comprising:
one or more computer-readable tangible storage devices and program instructions stored on at least one of the one or more computer-readable tangible storage devices, the program instructions comprising:
program instructions to receive an input question from a user;
program instructions to parse the received input question to extract input question components, wherein parsing utilizes natural language processing techniques;
program instructions to execute trained question component models to predict follow-up question components, wherein the extracted input question components are utilized as inputs to the trained question component models;
program instructions to combine the predicted follow-up question components to generate one or more follow-up questions; and
program instructions to return the one or more follow-up questions to the user.

### Claim 9 (depends on 8)

The computer program product of claim 8, wherein the trained question component models are selected from the group consisting of a model for lexical answer type prediction, a model for entity type prediction, a model for question type prediction, and a model for question relationship type prediction.

### Claim 10 (depends on 8)

The computer program product of claim 8, wherein the trained question component models are created utilizing training data generated by an expert user, and wherein the training data comprises a sequence of questions asked by the expert user, and wherein the sequence of questions is relevant to a single line of inquiry, and wherein the sequence of questions is relevant to a single line of inquiry if it satisfies a criterion selected from the group consisting of a topic similarity criterion and a graph connection criterion.

### Claim 11 (depends on 10)

The computer program product of claim 10, wherein the training data generated by the expert user is parsed, utilizing natural language processing techniques, to extract training data question components, and wherein machine learning algorithms are applied to the training data question components to determine training data question component patterns, and wherein the training data question component patterns are utilized to train the trained question component models.

### Claim 12 (depends on 11)

The computer program product of claim 11, wherein, the input question components, the follow-up question components, and the training data question components comprise an element from the group consisting of an entity type, a relationship type, a question type, and a lexical answer type.

### Claim 13 (depends on 8)

The computer program product of claim 8, wherein the one or more follow-up questions are in a same grammatical form as the input question.

### Claim 14 (depends on 8)

The computer program product of claim 8, further comprising:
ranking the one or more follow-up questions from most useful to least useful.

### Claim 15

A computer system for generating follow-up questions based on machine learning utilizing a computing device, the computer system comprising:
one or more processors, one or more computer-readable memories, one or more computer-readable tangible storage devices, and program instructions stored on at least one of the one or more computer-readable tangible storage devices for execution by at least one of the one or more processors via at least one of the one or more computer-readable memories, the program instructions comprising:
program instructions to receive an input question from a user;
program instructions to parse the received input question to extract input question components, wherein parsing utilizes natural language processing techniques;
program instructions to execute trained question component models to predict follow-up question components, wherein the extracted input question components are utilized as inputs to the trained question component models;
program instructions to combine the predicted follow-up question components to generate one or more follow-up questions; and
program instructions to return the one or more follow-up questions to the user.

### Claim 16 (depends on 15)

The computer system of claim 15, wherein the trained question component models are selected from the group consisting of a model for lexical answer type prediction, a model for entity type prediction, a model for question type prediction, and a model for question relationship type prediction.

### Claim 17 (depends on 15)

The computer system of claim 15, wherein the trained question component models are created utilizing training data generated by an expert user, and wherein the training data comprises a sequence of questions asked by the expert user, and wherein the sequence of questions is relevant to a single line of inquiry, and wherein the sequence of questions is relevant to a single line of inquiry if it satisfies a criterion selected from the group consisting of a topic similarity criterion and a graph connection criterion.

### Claim 18 (depends on 17)

The computer system of claim 17, wherein the training data generated by the expert user is parsed, utilizing natural language processing techniques, to extract training data question components, and wherein machine learning algorithms are applied to the training data question components to determine training data question component patterns, and wherein the training data question component patterns are utilized to train the trained question component models.

### Claim 19 (depends on 18)

The computer system of claim 18, wherein, the input question components, the follow-up question components, and the training data question components comprise an element from the group consisting of an entity type, a relationship type, a question type, and a lexical answer type.

### Claim 20 (depends on 15)

The computer system of claim 15, wherein the one or more follow-up questions are in a same grammatical form as the input question.
