Microsoft Technology Licensing, LLC

Microsoft Corporation

Michael Gamon

Patrick Pantel

Ahmed Hassan Awadallah

Sujay Kumar Jauhar

Mark James Encarnacion

Eric K Ringger

Simon H Corston-Oliver

William B Dolan

Ye-Yi Wang

Saliha Azzam

Ariel Fuxman

Nicholas Caldwell

Ashequl Qadir

Anthony Aue

Nirupama Chandrasekaran

Madian Khabsa

Yizheng Cai

Jianfeng Gao

Robert C Moore

Anitha Kannan

Christopher John Brockett

Lucretia Henrica Vanderwende

Gary Kacmarcik

Elnaz Nouri

Yu Su

Thomas Lin

Eric J Horvitz

Qiang Wu

Li Deng

Ming Zhou

Ryen William White

Robert George Atkinson

Johnson T Apacible

Malcolm Erik Pearson

Xiaodong He

Paul Nathan Bennett

Danyel Aharon Fisher

Daniel Morris

Sumit Basu

Bernhard Kohlmeier

Oriana Riva

David Richard Reed

Pradeep Chilakamarri

Xinying Song

Nikrouz Ghotbi

Richard Lloyd Hughes

Simon H Corston

Matthew Francis Hurst

Prabhdeep Singh

Richard G Campbell

Timothy C Franklin

Rajatish Mukherjee

Lucy H Vanderwende

Bahareh Sarrafzadeh

Arjun Mukherjee

Mark James Encarnación

Christian König

Diego Perez Del Carpio

Mei Yang

Chris Brockett

Naoki Otani

Nalin Singal

Shawon Sarkar

Nouran Soliman

Liye Fu

Sri Raghu Malireddi

Xuchao Zhang

Roger P Menezes

Dongyeop Kang

Martine Smets

Zhu Zhang

Hassan Sajjad

Tae Yano

Julie A Medero

Alexandre Klementiev

Hsiao-Wen Hon

Dmitriy A Belenko

Monwhea Jeng

Chu-Cheng Lin

Simon Corston-Oliver

Eric Ringger

Eric Joel Horvitz

Richard L Hughes

Systems and methods are described that are generally directed to generating a general task embedding representing task information. In examples, the generated task embedding may include predicted task information such that, rather being underspecified, the task embedding representative of the task may include additional specified information, where the task embedding can then be utilized in many different models and applications. Thus, task data may be received and at least a portion of the task data may be encoded using an encoder. Based on one or more outputs generated by the encoder and a type embedding associated with the task data, a task intent may be extracted or otherwise predicted based on the task data and one or more type encodings associated with the task data. The intent extractor may be trained on multiple auxiliary tasks with weak supervision that provide semantic augmentation to under-specified task texts.

Intent-based task representation learning using weak supervision

Generally discussed herein are devices, systems, and methods for determining a relationship between an edit and a comment. A system can include a memory to store parameters defining a machine learning (ML) model, the ML model to determine a relationship between an edit, by an author or reviewer, of content of a document and a comment, by a same or different author or reviewer, regarding the content of the document, and processing circuitry to provide the comment and the edit as input to the ML model, and receive, from the ML model, data indicating a relationship between the comment and the edit, the relationship including whether the edit addresses the comment or a location of the content that is a target of the comment.

Architectures for modeling comment and edit relations

Aspects of the present disclosure relate to task template generation and social task discovery. In examples, a task template catalog comprises task templates, which may be automatically generated and/or user-submitted, among other examples. Task templates can be reviewed, shared, and curated within the task template catalog. A user may browse the task catalog or search the task catalog for task templates. Once the user selects a task template, a task is generated based on the task template and added to the user's task list. In some examples, aspects of a task template may be customized. For example, a task may comprise parametric or conditional subtasks, thereby enabling a user to further tailor the task template to his or her needs. Thus, the task catalog provides a starting point from which the user can author a task in a task management application.

Task templates and social task discovery

A computing system for identifying tasks at risk in a collaborative project includes one or more processors configured to execute, during an inference-time phase, a collaborative project management program and a machine learning model. The collaborative project management program is configured to receive telemetry data associated with a task, process the telemetry data based at least in part on one or more task attributes, and output at least one feature associated with the task. The machine learning model is configured to receive, as inference-time input, the at least one feature associated with the task, and, responsive to receiving the at least one feature, output a risk prediction for the task. The system is configured to output an alert when the task is predicted to be at risk of not being completed by a predetermined due date.

Identification of tasks at risk in a collaborative project

Systems, storage media and methods for providing information for user prioritization of tasks associated with collaboratively developed content are described. Some examples may include: receiving a conversation thread associated with collaboratively developed content, the conversation thread including a plurality of comments authored by multiple different authors, generating a predicted measure of completion for the received conversation thread, the predicted measure of completion being at least one of a predicted number of remaining actions until the received conversation thread is resolved or a predicted number of total actions for the conversation thread to be resolved and providing, for display at a user interface, the predicted measure of completion for the received conversation thread, the predicted measure of completion being associated with the conversation thread at the user interface.

Collaborative communication triage assistance

A task agnostic framework for neural model transfer from a first language to a second language, that can minimize computational and monetary costs by accurately forming predictions in a model of the second language by relying on only a labeled data set in the first language, a parallel data set between both languages, a labeled loss function, and an unlabeled loss function. The models may be trained jointly or in a two-stage process.

Minimization of computational demands in model agnostic cross-lingual transfer with neural task representations as weak supervision

Automatically detected and identified tasks and calendar items from electronic communications may be populated into one or more tasks applications and calendaring applications. Text content retrieved from one or more electronic communications may be extracted and parsed for determining whether keywords or terms contained in the parsed text may lead to a classification of the text content or part of the text content as a task. Identified tasks may be automatically populated into a tasks application. Similarly, text content from such sources may be parsed for keywords and terms that may be identified as indicating calendar items, for example, meeting requests. Identified calendar items may be automatically populated into a calendar application as a calendar entry.

Automatic task extraction and calendar entry

This disclosure describes techniques and architectures that involve a latent activity model for workplace emails. Such a model is based, at least in part, on a concept that communications, such as email at a workplace, are purposeful and organized by activities. An activity is a set of interrelated actions and events around a common goal, involving a particular group of people, set of resources, and time framework, for example. The latent activity model involves a probabilistic inference in graphical models that jointly captures the interplay between latent activities and the email contexts governed by the emails. Such contexts may be email recipients, subject and body of the email, and so on.

Activity modeling in email or other forms of communication

A method includes receiving an email addressed to a recipient user, processing the received email using a reparametrized recurrent neural network model to identify an action based on the received email, and wherein the reparametrized recurrent neural network model has been trained on an email dataset annotated with recipient corresponding actions and reparametrized on unannotated conversation data having structures similar to email data.

AI system to determine actionable intent

Subject matter involves using natural language to Web application program interfaces (API), which map natural language commands into API calls, or API commands. This mapping enables an average user with little or no programming expertise to access Web services that use API calls using natural language. An API schema is accessed and using a specialized grammar, with the help of application programmers, canonical commands associated with the API calls are generated. A hierarchical probabilistic distribution may be applied to a semantic mesh associated with the canonical commands to identify elements of the commands that require labeling. The identified elements may be sent to annotators, for labeling with NL phrases. Labeled elements may be applied to the semantic mesh and probabilities, or weights updated. Labeled elements may be mapped to the canonical commands with machine learning to generate a natural language to API interface. Other embodiments are described and claimed.

Natural language interface to web API

In one embodiment, a web service engine servermay predict a successive action by a user based on an entity reference. The web service engine servermay identify an entity referencein a data transmission caused by a user. The web service engine servermay determine from the data transmission a user intention towards the entity referenceusing an intention model based on a transmission log. The web service engine servermay predict a related successive web action optionfor the entity referencebased on the user intention.

Modeling actions for entity-centric search

Growing community of inventors

Seattle, WA, United States of America

Michael Gamon