Microsoft Technology Licensing, LLC

Benjamin Zorn

Karthik Pattabiraman

Darko Kirovski

Ray Bittner

Rishabh Singh

Sumit Gulwani

Benjamin Livshits

Todd A Proebsting

Vinod Grover

Andrew Donald Gordon

Vu Minh Le

Advait Sarkar

David R Hanson

Christian Seifert

Oleksandr Polozov

Carina Suzana Negreanu

Neil Toronto

Dany Rouhana

Emery D Berger

Brian Paul Slininger

Sruti Srinivasa Ragavan

Ramarathnam R Venkatesan

Galen Clyde Hunt

Eric Cohen

Thomas Moscibroda

Karin Strauss

James R Larus

Steven P Levi

Kris Kuppuswamy Ganjam

Chad B Rothschiller

George M Moore

John D DeTreville

Daniel C Battagin

Emre Mehmet Kiciman

Stuart Hillary Schaefer

Wolfgang Grieskamp

Christopher W Fraser

Manuel Alfred Fahndrich

Thomas Michael Roeder

Elnaz Nouri

Mark Marron

Pallavi Choudhury

Marc Manuel Johannes Brockschmidt

Weide Zhong

Saswat Padhi

Jeffrey Todd Cohen

Ben Livshits

Dennis B Gannon

Shraddha Govind Barke

Charles M Curtsinger

Benjamin Stock

Mikael Mayer

Maxim Grechkin

Song Liu

Clemens Kolbitsch

Rahul Nagpal

Christopher Benedict

Rodrigo Moreira-Silveira

Gustavo Araujo Soares

Dmitri O Danilov

Christian Leopold Bejamin Poelitz

The indirect querying of models to determine capabilities possessed by the model. Such indirect queries take the form of model input that potentially includes a natural language input user data. Such model input is structured such that the output of the model is either not natural language at all, or else is natural language that is not semantically responsive to the natural language input. Nevertheless, the output is evaluated to estimate or determine the capability possessed by the model. Thus, models may be more fully utilized to their better potential.

Model capability extraction

The generation of a response to a task prompt that represents a task to perform on declarative code. The response is generated with the aid of a language model that was trained on imperative code. The declarative code includes declarations about data. A task prompt represents a task to perform on the declarative code. At least a portion of the declarative code and at least a portion of the task prompt are converted into input imperative code. The input imperative code is then caused to be provided as input to the language model, resulting in the language model generating output imperative code. At least a portion of the output imperative code is then converted into a response to the task prompt.

Responding to task prompt on declarative code using language model

Systems, methods, and computer-readable storage devices are disclosed for improved table identification in a spreadsheet. One method including: receiving a spreadsheet including at least one table; identifying, using machine learning, one or more classes of a plurality of classes for each cell of the received spreadsheet, wherein the plurality of classes include corners and not-a-corner; and inducing at least one table in the received spreadsheet based on the one or more identified classes for each cell of the received spreadsheet.

Systems, methods, and computer-readable media for improved table identification using a neural network

Embodiments are described that leverage variability of a chip. Different areas of a chip vary in terms of reliability under a same operating condition. The variability may be captured by measuring errors over different areas of the chip. A physical factor that affects or controls the likelihood of an error on the chip can be varied. For example, the voltage supplied to a chip may be provided at different levels. At each level of the physical factor, the chip is tested for errors within the regions. Some indication of the error statistics for the regions is stored and then used to adjust power used by the chip, to adjust reliability behavior of the chip, to allow applications to control how the chip is used, to compute a signature uniquely identifying the chip, etc.

Leveraging chip variability

A device includes a logic machine and a data-holding machine having instructions executable by the logic machine to receive a spreadsheet including a plurality of cells, apply an abstraction to the spreadsheet that defines one or more features of a cell set including one or more cells of the plurality of cells to form an abstracted representation of the spreadsheet, form, for the cell set, an input vector for a machine-learning prediction function from the abstracted representation of the spreadsheet, the machine-learning prediction function configured to output a prediction of one or more properties of the cell set based on the input vector, wherein the machine-learning prediction function is previously trained based on a plurality of previously-created spreadsheets, provide the input vector to the machine-learning prediction function; and output the prediction from the machine-learning prediction function.

Predicting spreadsheet properties

Data formatting rules to convert data from one form to another form are automatically determined based on a user's edits. A machine learning heuristic is applied to a user's edits to determine a data formatting rule that may be applied to data. For example, a user may make edits that add/remove characters from data, concatenate data, extract data, rename data, and the like. The machine learning heuristic may be automatically triggered in response to an event (e.g. after a predetermined number of edits are made to a same type of data) or manually triggered (e.g. selecting a user interface option). The data formatting rule may be applied to other data and the results of the formatting reviewable by the user. Based on further edits/reviews, the data formatting rule may be updated. The data formatting rules may be stored for later use.

Formatting data by example

The techniques described herein use statistical reasoning to determine whether a spreadsheet (e.g., cells) includes potential errors. The techniques determine a partition within a spreadsheet where the partition includes cells that share characteristics (e.g., same row or column, same type of content, same formatting, etc.). Once determined, the partition is evaluated based on defined properties. A property is applied to generate property values so that an anomaly can be identified. An anomaly can occur when a cell in the partition has a property value that is inconsistent with other property values of other cells in the same partition (e.g., an intra-partition anomaly). An anomaly can also occur when a cell in the partition has a property value that is inconsistent with property values of cells in a different partition (e.g., an inter-partition anomaly). The techniques analyze the anomalies to determine a priority value indicative of a likelihood of a potential error.

Detecting errors in spreadsheets

Disclosed herein are systems and methods for detecting script code malware and generating signatures. A plurality of script code samples are received and transformed into a plurality of tokenized samples. The tokenized samples are based on syntactical elements of the plurality of script code samples. One or more clusters of samples are determined based on similarities in different ones of the plurality of tokenized samples, and known malicious code having a threshold similarity to a representative sample of the cluster of samples is identified. Based on the identifying, the cluster of samples is identified as malicious. Based at least on respective ones of the plurality of tokenized samples associated with the cluster of samples, a generalized code signature usable to identify the script code samples in the cluster of samples is generated.

Code labeling based on tokenized code samples

A database implemented by storing information encoded in DNA molecules provides high information density but the information is more difficult to access than in conventional electronic storage media. A relational database is a way of organizing information by using multiple related tables. Relational algebra operations are performed on relational databases to locate and manipulate information. This disclosure provides techniques for implementing relational algebra operations on a relational database that uses DNA molecules to store information. The techniques of this disclosure relate to the structure of DNA molecules used to store the information and to correlations between relational algebra operations and manipulations of DNA molecules.

Relational DNA operations

Systems and methods for increasing user confidence in results that are produced by one or more programs that are generated by an underlying Programming-By-Example (PBE) system based on user input examples. A plurality of generated programs that have been generated using one or more user input examples that are indicative of an output that should be achieved to comply with a user determined result are received. The generated programs are narrowed based on one or more sub-expressions of the programs that are likely to cause the resultant program to comply with the user determined result. The one or more sub-expressions are exposed. Input that selects at least one of the one or more exposed sub-expressions to thereby identify the one of the generated programs that will result in the user determined result is received.

User interaction models for disambiguation in programming-by-example

A program creation system is described which generates sets of subprograms for respective input-output examples. The program creation system then groups the sets into partitions by performing an intersection operation. According to one aspect, the program creation system generates subprograms so as to exclude tokens that are not represented by the input strings of the input-output examples. According to another aspect, the program creation system first generates the subprograms without attempting to generate loop-type expressions. If this operation produces unsatisfactory results, the program creation system repeats its processing, this time including loop-type expressions. According to another aspect, the program creation system performs the grouping operation using an expedited graph-intersection operation. According to another aspect, the program creation system ranks programs (which are created based on the results of the grouping operation) based on the presence of preferred features found in the programs.

Expedited techniques for generating string manipulation programs

Power management functionality is described for implementing an application in an energy-efficient manner, without substantially degrading overall performance of the application. The functionality operates by identifying at least first data and second data associated with the application. The first data is considered to have a greater potential impact on performance of the application compared to the second data. The functionality then instructs a first set of hardware-level resources to handle the first data and a second set of hardware-level resources to handle the second data. The first set of hardware-level resources has a higher reliability compared to the second set of hardware-level resources. In one case, the first and second hardware-level resources comprise DRAM memory units. Here, the first set of hardware-level resources achieves greater reliability than the second set of hardware-level resources by being refreshed at a higher rate than the second set of hardware-level resources.

Providing hardware resources having different reliabilities for use by an application

Growing community of inventors

Woodinville, WA, United States of America

Benjamin Zorn