公开(公告)号：US11475303B2

公开(公告)日：2022-10-18

申请号：US16848707

申请日：2020-04-14

Applicant: Microsoft Technology Licensing, LLC

Inventor： Andrew Wagner ,  Tiyasa Mitra ,  Sujeeth Subramanya Bharadwaj ,  Saurabh Mohan Kulkarni ,  Marc Tremblay

IPC: G06N3/08 ,  G06F17/15 ,  G06F17/16 ,  G06N3/04

Abstract: Techniques for training neural networks are provided. According to one set of embodiments, a first array is processed in a spreading component to produce a second array, where a first dimension of the first array corresponds to at least one sequence of approximately orthogonal numeric vectors representing tokens, and where the spreading component combines values along the first dimension. The second array is processed in a transformer neural network to determine correlations between the sequence, which produces a third array. One or more batches of the third array are processed in a de-spreading component to produce a fourth array.

公开(公告)号：US11921911B2

公开(公告)日：2024-03-05

申请号：US17374942

申请日：2021-07-13

Applicant: Microsoft Technology Licensing, LLC

Inventor： Stavros Volos ,  David Thomas Chisnall ,  Saurabh Mohan Kulkarni ,  Kapil Vaswani ,  Manuel Costa ,  Samuel Alexander Webster ,  Cédric Alain Marie Fournet ,  Richard Osborne ,  Daniel John Pelham Wilkinson ,  Graham Bernard Cunningham

IPC: H04L29/00 ,  G06F21/60 ,  G06F21/85 ,  H04L9/30 ,  H04L9/32

CPC classification number: G06F21/85 ,  G06F21/602 ,  H04L9/30 ,  H04L9/3265

Abstract: A peripheral device, for use with a host, comprises one or more compute elements a security module and at least one encryption unit. The security module is configured to form a trusted execution environment on the peripheral device for processing sensitive data using sensitive code. The sensitive data and sensitive code are provided by a trusted computing entity which is in communication with the host computing device. The at least one encryption unit is configured to encrypt and decrypt data transferred between the trusted execution environment and the trusted computing entity via the host computing device. The security module is configured to compute and send an attestation to the trusted computing entity to attest that the sensitive code is in the trusted execution environment.

公开(公告)号：US11126757B2

公开(公告)日：2021-09-21

申请号：US16166047

申请日：2018-10-19

Applicant: Microsoft Technology Licensing, LLC

Inventor： Stavros Volos ,  David Thomas Chisnall ,  Saurabh Mohan Kulkarni ,  Kapil Vaswani ,  Manuel Costa ,  Samuel Alexander Webster ,  Cédric Alain Marie Fournet

IPC: H04L29/00 ,  G06F21/85 ,  G06F21/60 ,  H04L9/30 ,  H04L9/32

Abstract: A peripheral device, for use with a host, comprises one or more compute elements a security module and at least one encryption unit. The security module is configured to form a trusted execution environment on the peripheral device for processing sensitive data using sensitive code. The sensitive data and sensitive code are provided by a trusted computing entity which is in communication with the host computing device. The at least one encryption unit is configured to encrypt and decrypt data transferred between the trusted execution environment and the trusted computing entity via the host computing device. The security module is configured to compute and send an attestation to the trusted computing entity to attest that the sensitive code is in the trusted execution environment.

公开(公告)号：US11544537B2

公开(公告)日：2023-01-03

申请号：US16848748

申请日：2020-04-14

Applicant: Microsoft Technology Licensing, LLC

Inventor： Andrew Wagner ,  Tiyasa Mitra ,  Sujeeth Subramanya Bharadwaj ,  Marc Tremblay ,  Saurabh Mohan Kulkarni

IPC: G06F40/289 ,  G06N3/08 ,  G06N3/04 ,  G06K9/62

Abstract: Embodiments of the present disclosure include a method for token-position handling comprising: processing a first sequence of tokens to produce a second sequence of tokens, wherein the second sequence of tokens has a smaller number of tokens than the first sequence of tokens; masking at least some tokens in the second sequence to produce masked tokens; moving the masked tokens to the beginning of the second sequence to produce a third sequence; encoding tokens in the third sequence into a set of numeric vectors in a first array; and processing the first array in a transformer neural network to determine correlations among the third sequence, the processing the first array producing a second array.

公开(公告)号：US10564930B2

公开(公告)日：2020-02-18

申请号：US15949048

申请日：2018-04-09

Applicant: Microsoft Technology Licensing, LLC

Inventor： Saurabh Mohan Kulkarni ,  Marc Tremblay

IPC: G06F7/38 ,  G06F1/03

Abstract: Reduced precision computer number formats inherently limit the quantity of discrete numeric values that can be represented. Therefore, the solution values of an arithmetic function, for each numeric value that is individually and uniquely expressible utilizing such a reduced precision computer number format, can be precomputed since the quantity of unique solution values can be limited to a quantity that can be conveniently stored, such as in an array. Subsequently, rather than computing the solution value of such an arithmetic function, for a given input value, the precomputed array can be referenced and a solution value corresponding to the given input value can be read from the array. Reading numeric values from an array can be substantially faster than computing solution values of a computationally-expensive arithmetic function.