公开(公告)号：WO2021055007A1

公开(公告)日：2021-03-25

申请号：PCT/US2020/022541

申请日：2020-03-13

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： KIM, Chung Hwan ,  RHEE, Junghwan ,  YU, Xiao ,  TANG, LuAn ,  CHEN, Haifeng ,  KIM, Kyungtae

IPC: G06N20/00 ,  G06N3/08 ,  G06F21/12

Abstract: A computer-implemented method for efficient and scalable enclave protection for machine learning (ML) programs includes tailoring at least one ML program to generate at least one tailored ML program for execution within at least one enclave, and executing the at least one tailored ML program within the at least one enclave.

公开(公告)号：WO2015057617A1

公开(公告)日：2015-04-23

申请号：PCT/US2014/060358

申请日：2014-10-14

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： RHEE, Junqhwan ,  ZHANG, Hui ,  ARORA, Nipun ,  JIANG, Guofei ,  KIM, Chung Hwan

IPC: G06F11/36

CPC classification number: G06F11/3636 ,  G06F11/3419

Abstract: Methods and systems for performance inference include inferring an internal application status based on a unified call stack trace that includes both user and kernel information by inferring user function instances. A calling context encoding is generated that includes information regarding function calling paths. The analysis includes performing a top-down latency breakdown and ranking calling contexts according to how costly each function calling path is.

Abstract translation: 用于性能推理的方法和系统包括通过推断用户功能实例来推断基于包括用户和内核信息的统一调用堆栈跟踪的内部应用程序状态。 生成包含有关函数调用路径的信息的调用上下文编码。 分析包括根据每个功能调用路径的代价昂贵地执行自上而下的延迟故障和排序呼叫上下文。

公开(公告)号：EP3058465A1

公开(公告)日：2016-08-24

申请号：EP14854428.1

申请日：2014-10-14

Applicant: NEC Laboratories America, Inc.

Inventor： RHEE, Junqhwan ,  ZHANG, Hui ,  ARORA, Nipun ,  JIANG, Guofei ,  KIM, Chung Hwan

IPC: G06F11/36

CPC classification number: G06F11/3636 ,  G06F11/3419

Abstract: Methods and systems for performance inference include inferring an internal application status based on a unified call stack trace that includes both user and kernel information by inferring user function instances. A calling context encoding is generated that includes information regarding function calling paths. The analysis includes performing a top-down latency breakdown and ranking calling contexts according to how costly each function calling path is.