Browse Prior Art Database

Cognitive help for software applications and system Disclosure Number: IPCOM000244409D
Publication Date: 2015-Dec-09
Document File: 3 page(s) / 46K

Publishing Venue

The Prior Art Database


Disclosed is a cognitive help model that is a generative model based on the representative dataset processing which is based on identified patterns of help usage for various categories of users: advanced, beginners, experts. In addition, the patterns of user communication with the computing device are established and processed for feature generation at the model learning or training stage.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 42% of the total text.

Page 01 of 3

Cognitive help for software applications and system

Current help systems present information in the same way to all users. This can make it very difficult and time consuming to locate the help a user needs, in a presentation that they would understand. For example, people with cognitive disabilities may get information that is not presented at a level that they can understand, and advanced users of a product can often be forced to navigate through a bunch of introductory material to get to the more detailed, deeper information that they are after. This invention solves this problem by using machine learning to tailor what help is provided to a user by examining the way a user uses the system and making

judgements about the level of experience that person has, and thus what level of information the

user could understand.

    By examining the user's navigation of the system, this invention can make judgements about how advanced the user is in using the system. For example, Lotus Notes*: if the user turns on advanced menus or has spent time in Designer adding formula language or LotusScript*, when that user seeks help, the system will show them details about advanced menus or LotusScript at the outset rather than requiring they navigate through a bunch of introductory material to get there.

    Help invocation identifies the context and provides the solution based on the context and the user profile. Example : wizard fails exception. The help system: learns the user profile based on previously submitted questions, it identifies the UI of the applications, performs data mining on the computer configuration, formulates the search criteria, performs search based on the user's preferences.

    Cognitive help system is implemented based on Machine Learning methodology. Once trained on the particular user's system and application usage, the cognitive help system "learns" the patterns in the user's system usage and evaluates the user's practical knowledge level. The system then can provide the guidelines to that user in the format and with the content that is expressed to the user level of comprehension and expertise. The cognitive help system would continue to learn new patterns in user behavior. Help could provide the recommendations to the user that are relevant to the context of the current user's activities. The help system originally is installed as a base-trained application module. It provides help interface that is the most typical for the users of that profile (what most of the time users would ask for help in the similar situation, relevant application, etc.). It is implemented as a set of rules or action vectors which is mapped to that typical user profile: what applications are mostly used, what is a typical day-to-day software usage, frequency of new installations, searches of the tools feature, etc. During the "learning" stage, system recognizes certain patterns in user - computer communications. For example, repetitiveness of the same...