Browse Prior Art Database

Adaptive Intelligent Personal Assistant Dialoguing

IP.com Disclosure Number: IPCOM000247276D
Publication Date: 2016-Aug-18
Document File: 3 page(s) / 22K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a system that adapts and customizes intelligent personal assistant dialogues for a user. In addition to adapting the level of granularity in the response, the system detects a user’s vocabulary awareness and then modifies/select s words that are appropriate for the user’s knowledge base and language style.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 3

Adaptive Intelligent Personal Assistant Dialoguing

Intelligent Personal Assistants (IPAs)/agents are revolutionizing human-computer-interaction and human-to-human interaction. Users speak to the agent to manage schedules, read/write communications, reserve time at restaurants, and connect with/search vast knowledge bases. However, users are not always satisfied with the responses that the intelligent personal assistant provides (e.g., unfamiliar vocabulary, etc.).

For example, a user with limited experience using a computer needs a basic response to a technical, computer related question; whereas, a user who is an experienced computer user and programmer can receive more complex information. A method and system is needed to adapt intelligent personal assistants dialogues to the conditions of the user.

The novel contribution is a system that adapts and customizes intelligent personal assistant dialogues for a user. In addition to adapting the level of granularity in the response, the system detects a user's vocabulary awareness and then modifies/select s words that are appropriate for the user's knowledge base and language style.

At a high level, the system performs the following:

1. Detects the activation of an IPA (e.g., long thumb press, specific action, voice command, etc.)

2. Retrieves the user's prior language usage Analyzes:
A. Prior language usage
B. Voice timbre
C. Language complexity


3. Determines, based on the analysis, an appropriate response to the user

The system may also capture contextual data (e.g., time of day, location, background voices, etc.) to determine whether others are present for the response (i.e., can hear it). If third parties are likely to hear the response, then the system customizes it for a broader audience.

Following are the steps for implementing the system for adaptive intelligent personal assistant dialoguing. The system:

1. Detects the activation of an IPA via a specific action, such as
A. Menu action
B. Device action (e.g., long thumb press, feedback to the system, etc.)
C. Voice activation (e.g., "Hey IPA" or other command)

2. Records the voice print of the user for further analysis 3. Retrieves the prior language usage for the user from any online social network or messaging system

A. Detects that the user is creating a new message through: i. push event: publishes content (each event) to a subscriber and the subscriber acts on the message (e.g., via social media, enterprise,

1


Page 02 of 3

or compliance Application Programming Interfaces (APIs))

ii. pull event: continuously monit...