International Business Machines Corporation

Michael Desmond

Harold Leon Ossher

Jacquelyn Annette Martino

John Thomas Richards

Paul Michael Matchen

Rachel K E Bellamy

Calvin Bruce Swart

Evelyn Duesterwald

Matthew R Arnold

Sophia Krasikov

Mark E Wallace

Edith Gail Schonberg

Jeffrey Scott Boston

Matthew J Callery

James Johnson

Darrell Christopher Reimer

Robert Leslie Yates

Pietro Mazzoleni

Casey Dugan

Yunfeng Zhang

Harini Srinivasan

Qingzi Vera Liao

Ian D Simmonds

David John Piorkowski

Rachel Katherine Emma Bellamy

Qian Pan

Kristina Marie Brimijoin

Stephen Renwick

Zahra Ashktorab

Michelle Brachman

Maire Kehoe

Darragh Curran

Yiyun Chen

Martin Harrigan

Trevor Johnston

Cenk Kuzucu

Provided are techniques for moderating Artificial Intelligence (AI) agent interlocution in group dialog environments. Under control of an interlocution module that has been trained with dialog content and dialog turns, an indication that interlocution is to be determined for a group dialog is received. Under control of the interlocution module, it is determined whether an AI agent is to participate in the group dialog based on a current dialog context and a dialog response. Under control of the interlocution module, in response to determining that the AI agent is to participate in the group dialog, the AI agent is triggered to post the dialog response to the group dialog.

Moderating artificial intelligence (AI) agent interlocution in group dialog environments

Techniques are provided for decision making tasks using a hybrid approach where cooperation between an AI assessor and a human labeler controls automation of the process. In one aspect, a method for hybrid decision making automation includes: monitoring interactions between an AI assistant and a human decision maker; tracking, from the interactions, agreement of the human decision maker with decision predictions made by the AI assistant; determining a predicted performance of data tasks by the AI assistant on unseen data based on the agreement of the human decision maker with the decision predictions over time; and assessing delegation of remaining data tasks on the unseen data to the AI assistant using the predicted performance.

Hybrid decision making automation

An approach is provided for training classifiers used in machine learning. A corpus of training data is received. One or more clusters of the training data is generated according to features of the training data. The one or more clusters are refined using user-specified rules. One or more classifiers are trained for use in machine learning based upon the refined one or more clusters.

Training classifiers in machine learning

A method optimizes a training of a machine learning system. A conflict detection system discovers a conflict between a first training data and a second training data for a machine learning system, where the first training data and the second training data are ground truths that describe a same type of entity, and where the first training data and the second training data have different labels. In response to discovering the conflict between the first training data and the second training data for the machine learning system, an oracle adjusts the different labels of the first training data and the second training data. The machine learning system is then trained using the first training data and the second training data with the adjusted labels.

Discovering and resolving training conflicts in machine learning systems

Methods and systems for multi-dimensional data labeling. A structured data set having a plurality of rows is obtained, the structured data set comprising a set of data attributes, each data attribute having a data value for each of the plurality of rows of the structured data set. The structured data set is decomposed into a plurality of dimensions, each dimension defining a proper subset of the data attributes based on coherence criterion. A dimension label is obtained for each dimension of at least a portion of the plurality of rows of the structured data set and the dimension labels for a given one of the rows of the structured data set are consolidated into at least one row label for the given one of the rows.

Multi-dimensional data labeling

The generating of actionable recommendations for tuning model metrics of an Artificial Intelligence (AI) system includes partitioning a key performance indicator (KPI) range associated with a target system into a plurality of buckets. Log data including at least one KPI of the target system and one or more AI model metrics is partitioned and distributed across the plurality of buckets. For each bucket, an aggregate value of the one or more AI model metrics across the log data is computed and weighted according to the volume of log data in that bucket. A correlation factor between the aggregate value and a representative KPI value for each bucket is determined. A model tuning recommendation to increase ranking of the AI model metrics according to the determined correlation factor is provided to an output device and/or to the AI system for updating the one or more AI model.

Generating model insights by progressive partitioning of log data across a set of performance indicators

Methods, systems and computer program products for improving ground truth quality for modeling are provided. Aspects include receiving a plurality of data inputs, wherein each of the plurality of data inputs has an associated label. Aspects also include training a model based on the plurality of data inputs. Aspects also include generating a plurality of vector representations corresponding to the plurality of data inputs based on the model. Aspects also include clustering the plurality of vector representations into one or more clusters. Aspects also include identifying at least one anomalous data input based on the one or more clusters. The at least one anomalous data input can be a data input of the plurality of data inputs that is mislabeled, contributes to an ambiguous class structure or is an outlier. Aspects also include outputting a notification that provides an indication of the at least one anomalous data input.

Ground truth quality for machine learning models

A computing method receives a labeled sample from an annotator. The method may determine a plurality of reference model risk scores for the first labeled sample, where each reference model risk score corresponds to an amount of risk associated with adding the first labeled sample to a respective reference model of a plurality of reference models. The method may determine an overall risk score for the first labeled sample based on the plurality of reference model risk scores. The method may further determine a probe for confirmation of the first labeled sample and a trust score for the annotator by sending the probe to one or more annotators. In response to determining a trust score for the annotator the method may add the labeled sample to a ground truth or reject the labeled sample.

Filter for harmful training samples in active learning systems

An aspect of the disclosure includes a method, a system and a computer program product for matching a participant with an interest group at a venue. The system including computer inputs that receive signals identifying locations of participants in a venue having sessions, the computer inputs identifying a session location for each participant. An aggregator is coupled the computer inputs and determines interests of each participant by aggregating common session characteristics with participant behavioral characteristics and ranking a degree of interests for each participant. A matcher is coupled to the aggregator receives interest group ranked lists for each interest group, the matcher further matching the interest group to the participant by matching the degree of interests for each participant to the interest group interests. A signal output is coupled to the matcher and transmits an output signal to a representative based on a relevancy parameter being above a threshold.

Real-time discovery of interests of individuals and organizations participating in a physical event

Mechanisms are provided for generating a shape tag cloud display. A user input is received that specifies an arbitrarily hand-drawn shape tag. A set of shape tag equivalence classes is updated based on the received user input to assign the arbitrarily hand-drawn shape tag to a shape tag equivalence class having similar previously entered arbitrarily hand-drawn shape tags. Rankings of the shape tags relative to one another are generated. The shape tag cloud display is generated based on the rankings. A representation of each shape tag within the shape tag cloud display has display characteristics based on the rankings. The shape tag cloud display is then output.

Creating tag clouds based on user specified arbitrary shape tags

Disclosed is a computer-implemented method to create a presentation having at least one sheet, where each sheet contains one or more visual elements, as well as a system and computer program embodied on a computer-readable storage medium. The method includes, in response to input from a user, representing visual elements as depictions in a model, where each depiction depicts a single item in the model, where each item is a unit of information with a unique identifier, where each item can be depicted by at least one depiction and each depiction defines the visual representation of the item's information via the visual element. The method further includes, in response to a change to an item made by altering the visual element associated with the depiction of the item, automatically propagating the change to the item to all other depictions of the item across all sheets of the presentation. The method further includes updating the associated visual elements to reflect the change to the item.

Smart slide creation and presentation

Techniques are presented for providing a software fitting assessment. The techniques may be performed by methods, apparatus, and/or computer program products. The techniques include automatically matching on a computer system one or more specified requirements for a project with one or more software functions stored in a repository. The automatically matching includes mining the repository in order to match requirements. The repository includes software functions, requirements accumulated from previous projects, and results of stored matches between the software functions and the requirements accumulated from previous projects. The techniques include outputting by the computer system one or more results of the matching.

Techniques providing a software fitting assessment

Automatically creating a series of intermediate states may include receiving a start state and an end state of a reactive system, identifying one or more components of the start state and the end state and determining one or more events associated with the one or more components. One or more intermediate states between the start state and the end state, and one or more transitions from and to the one or more intermediate states are created using the one or more components of the start state and the end state and the one or more events associated with the one or more components. The one or more intermediate states and the one or more transitions form one or more time-based paths from the start state to the end state occurring in response to applying the one or more events to the associated one or more components.

In-betweening interactive states

A mark-based containment system, in one aspect, may include a graphical user interface operable to receive an indication of a mark as a container marking, attribute one or more container attributes to the container marking, and store the container mark and the container attributes associated with the container mark.

Growing community of inventors

White Plains, NY, United States of America

Michael Desmond