Browse Prior Art Database

Auditory Status of End-User Query Outcomes

IP.com Disclosure Number: IPCOM000245181D
Publication Date: 2016-Feb-17
Document File: 3 page(s) / 114K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a system for monitoring information retrieval systems using ambient auditory alerts. Specific sounds will change depending on the magnitude, general sentiment (positive or negative), and type of event.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 3

Auditory Status of End -

Disclosed is a system for monitoring information retrieval systems using ambient auditory alerts. Specific sounds will change depending on the magnitude, general sentiment (positive or negative), and type of event. Understanding operational status for a deployed information retrieval system requires administrators to monitor and interpret multiple variables simultaneously. Some examples of such variables include the availability and responsiveness of the system as well the overall user experience with search. Actively monitoring and interpreting streaming data in a live system is not realistic without automated assistance. As a result, many system administrators rely on reports for specified time intervals and alerts via email, phone, text, pager, and similar devices.

    While these monitoring systems are effective, because of the disruptive nature of alerting systems, they are generally only used for monitoring emergencies. Further, common alert systems cannot easily distinguish between events of varying degrees of interest - for example, an emergency (failure) and positive event (high water mark of unique users) will both be sent using the same mechanism. This is a problem for information retrieval systems which have a large number of interesting, non-emergency events.

    Finally, information retrieval systems require a high degree of experience to interpret data for the purpose of evaluating user experience. User experience can be measured by way of common metrics including the number of results in a query, relevancy scores for a given set of search results, and whether a search leads to a result click among other metrics. While visual dashboards and similar devices can be used to create a continuous, visual feedback loops, what can be seen is but one variable for a multivariate problem. Further, visible indicators are only useful for individuals who can see them. There are many environments and situations (such as the visually impaired) in which visual cues are not useful. Sound can help address many of these shortcomings.

The disclosed system works in the following way.

Events of interest are configured by an administrator. In the context of information retrieval, these would include query events such as user searches or result clicks, and ingestion events such as number of records ingested (e.g., every 10,000 records), conversion failures, or interesting annotations, among other events.

As interesting events occur in the information retrieval system being monitored, they are registered with the auditory status system. When an event is registered, it is given a label, a type, a sentiment (positive, negative, or neutral), and a magnitude (numeric value indicating relative strength or weakness of the event). Assignment is usually determined deterministically at code time but is not registered until the event occurs at run time.

Event registration occurs via an HTTP request made to the auditory status system...