Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

COLLABORATION WORKSPACE RECOMMENDATION SYSTEM

IP.com Disclosure Number: IPCOM000247668D
Publication Date: 2016-Sep-26

Publishing Venue

The IP.com Prior Art Database

Related People

Aseem Asthana: AUTHOR [+4]

Abstract

Machine learning algorithms are used to provide a more rich and accurate online collaboration workspace recommendations. Additionally, a mechanism is provided for administrators to enforce organization specific policies and to provision users or organization with recommended workspaces. This solution integrates of social graphs/interest graphs with online collaboration workspaces.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 17% of the total text.

Page 01 of 14

COLLABORATION WORKSPACE RECOMMENDATION SYSTEM

AUTHORS:

   Aseem Asthana
Amarkanth Ranganamayna Ashish Chotai
Shamim Pirzada

CISCO SYSTEMS, INC.

ABSTRACT

    
Machine learning algorithms are used to provide a more rich and accurate online collaboration workspace recommendations. Additionally, a mechanism is provided for administrators to enforce organization specific policies and to provision users or organization with recommended workspaces. This solution integrates of social graphs/interest graphs with online collaboration workspaces.

DETAILED DESCRIPTION

     Team-based online collaboration systems have been in existence. However, none of the existing systems have been able to solve the problem of automatic workspace discovery. A new user that joins any online collaboration system/service is currently confronted with the problem of finding the right workspaces to join.

    Presented herein is a dynamic way of discovering collaboration workspaces. A machine learning based algorithm provides a recommendation system for users/admins to join appropriate collaboration workspaces.

    The machine learning algorithm can focus on multiple aspects. When a user is newly joined he/she will be offered the workspaces that are popular and mandatory in the existing organization. Over time the activities of the on-boarded users will be monitored and the machine learning algorithm computes: a. Correlation of the user with similar users. For example, this can be obtained by roles in organization, expertise in an area, common teams, etc.

Copyright 2016 Cisco Systems, Inc.

1


Page 02 of 14

b. Correlation of user activities with similar activities. For example, activities will be mined using attributes like "workspace name/subject", keywords in content share, topics of discussion in conversation etc.

Using these aspects described above, the machine learning algorithm provides recommendations.

    These recommendations may be notified to the administrator who can then choose to configure the appropriate workspaces for the users or even for the organization. For example, in the Sparkā„¢ service of Cisco Systems, Inc., this may be achieved by a mechanism that provides for an administrator to control and comply with organization policies for workspaces. This will typically be true for non-moderated workspaces. In another embodiment, these recommendations will be notified to the users who can then choose to join the suggested workspace. For example, in the Spark service, this may be achieved via the various Spark clients.

Further details are now provided on the algorithms used to provide a

recommendation for users/admins to join appropriate collaboration room/workspaces.

The algorithm involves two unique aspects:


A)

Automatic "user profile" generation and its use in content-based filtering to

recommend appropriate rooms for a user.


B) Automatic determination of "room ratings" for a given user and its use in

collaborative filtering to recommend appropriate rooms for a user.

This a...