International Business Machines Corporation

Devin A Conley

Lakshminarayanan Krishnamurthy

Oren Sar-Shalom

Priscilla S Moraes

James Patrick Hoff

Sridhar Sudarsan

Scott Christopher Moonen

Priscilla Santos Moraes

Simon J Kofkin-Hansen

Chris Moss

Jordan Shamir

Iain Mccown

Bryan M Buckland

An approach is provided that receives a message and applies a deep analytic analysis to the message. The deep analytic analysis results in a set of enriched message embedding (EME) data that is passed to a trained neural network. Based on a set of scores received from the trained neural network, a conversation is identified from a number of available conversations to which the received message belongs. The received first message is then associated with the identified conversation.

Enriched message embedding for conversation deinterleaving

A method comprises obtaining a set of log files for a software system. The set of log files applies to an extended window. A periodic pattern in a first set of error-event surges in the set of log files is identified. The error-event surges in the first set is identified as event noise. A second set of log files for the software system is obtained. The second set of log files applies to a shortened window. Timeseries analysis on the second set of log files is performed. A particular error-event surge in a detection period in the second set of log files that is abnormal as compared to the shortened window is detected based on the timeseries analysis. That the particular error-event surge does not fit into the periodic pattern is determined, the particular error-event surge is characterized as an anomaly, based on the determining.

Anomaly recognition in information technology environments

An engagement classifier for a group chatbot is trained by leveraging the implicit dataset generated by humans engaging in both direct messages as well as group conversations. Human-to-human direct messages are used as an approximate representation of the domain knowledge and expertise of each user. The decision to engage in a group conversation is assumed to be based on that domain knowledge. The knowledge representations and instances of engagements in group conversations yields an effective set of features and labels which can be used to model the engagement decision. The same transfer learning technique is used to generate a knowledge representation for the group chatbot. Given this representation of the domain knowledge of the chatbot, the classifier can predict whether it should engage in any particular group conversation.

Unsupervised technique for training an engagement classifier in chat-based group conversation

Growing community of inventors

Austin, TX, United States of America

Devin A Conley