Browse Prior Art Database

Automatic Deployment of Cognitive Computing Component With Triggered Meeting Cancellations

IP.com Disclosure Number: IPCOM000238576D
Publication Date: 2014-Sep-04
Document File: 2 page(s) / 41K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method for predicting the ability of machine learning technology or cognitive computing technology to solve a problem that a scheduled meeting aims to solve. Using historical data of meeting inputs and associated outcomes, the method assigns a meeting replacement cost and benefit to the cognitive computing element in terms of time to compute, expected error, and savings in human cognitive investment, in order to determine if a meeting may be cancelled.

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

Page 01 of 2

Automatic Deployment of Cognitive Computing Component With Triggered Meeting Cancellations

In an enterprise, meetings are often scheduled to solve problems in the space of the cognitive processes of individual meeting participants, rather than in the

lower dimensional spaces of current machine learning methods . As these machine learning methods improve, and simulation of the cognitive processes of individuals become possible, certain meetings can be automated. Evaluating when a meeting may be replaced with a machine learning artifact or a cognitive computing artifact , and when an actual meeting is required, is therefore a problem that enterprises can expect to confront more often with the rise of cognitive computing .

The novel solution is a method to evaluate cognitive computing technology with respect to historical meeting data in order to automatically cancel a scheduled meeting . The method includes an estimate of cost and risk associated with using the cognitive computing element over using a traditional meeting. If a meeting is cancelled, then the meeting cancellation notice includes a description of meeting outcomes computed by the cognitive computing element, and the meeting is automatically removed from a calendar.

Taking inputs from historical data of meetings of a particular type and the meetings' associated inputs and outcomes, the system evaluates a set of cognitive computing elements for the ability to generate the same outcomes given the inputs . The cost of computing these outcomes is then estimated and compared to the human cognitive costs. The expected error of the cognitive computing outcomes is also estimated , and compared against acceptable limits set by the enterprise , as well as the expected error of the meeting. If the cost of using cognitive computing to achieve the outcomes is

within some tolerance, then the system automatically cancels the meeting and initiates the computation.

To implement the method for automatic deployment of a cognitive computing component with triggered meeting cancellations in a preferred embodiment :


1. Construct a database of historical meeting categories , inputs, and outcomes


2. Given a set of cognitive computing elements (M1) n), evaluate each for the

associated error predicting outcomes of historical meetings given the inputs (Em) of each

3. Calculate the cost (Tm) of applying each cognitive computing elements as a

function of E and the associated computational costs (Cm) of each element; T=f(Mm, Em, Cm)

4. Compare the cost Tm to the similarly expected errors and cognitive costs of a scheduled set of meeting participants (G)

5. If G> Tm, then automatically cancel the scheduled meeting and send notification to the meeting organizer that a cognitive computing element can provide the meeting outcomes

1


Page 02 of 2

No other automatic meeting cancellation system makes use of an evaluation of estimates of cognitive computing elements errors and...