公开(公告)号：US20220392429A1

公开(公告)日：2022-12-08

申请号：US17337518

申请日：2021-06-03

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Kaizhi Qian ,  Yang Zhang ,  Shiyu Chang ,  Jinjun Xiong ,  Chuang Gan ,  David Cox

IPC: G10L13/10 ,  G06N20/00 ,  G10L21/013 ,  G10L17/04 ,  G10L25/63

Abstract: A computer-implemented method is provided of using a machine learning model for disentanglement of prosody in spoken natural language. The method includes encoding, by a computing device, the spoken natural language to produce content code. The method further includes resampling, by the computing device without text transcriptions, the content code to obscure the prosody by applying an unsupervised technique to the machine learning model to generate prosody-obscured content code. The method additionally includes decoding, by the computing device, the prosody-obscured content code to synthesize speech indirectly based upon the content code.

公开(公告)号：US20220253426A1

公开(公告)日：2022-08-11

申请号：US17170164

申请日：2021-02-08

Applicant: International Business Machines Corporation ,  The Board of Trustees of the University of Illinois

Inventor： Yada Zhu ,  Jinjun Xiong ,  Jingrui He ,  Lecheng Zheng ,  Xiaodong Cui

IPC: G06F16/23 ,  G06N3/04 ,  G06N3/08

Abstract: Time series data can be received. A machine learning model can be trained using the time series data. A contaminating process can be estimated based on the time series data, the contaminating process including outliers associated with the time series data. A parameter associated with the contaminating process can be determined. Based on the trained machine learning model and the parameter associated with the contaminating process, a single-valued metric can be determined, which represents an impact of the contaminating process on the machine learning model's future prediction. A plurality of different outlier detecting machine learning models can be used to estimate the contaminating process and the single-valued metric can be determined for each of the plurality of different outlier detecting machine learning models. The plurality of different outlier detecting machine learning models can be ranked according to the associated single-valued metric.

公开(公告)号：US11327973B2

公开(公告)日：2022-05-10

申请号：US16751317

申请日：2020-01-24

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： I-Hsin Chung ,  Jinjun Xiong ,  Carl Pearson

IPC: G06F11/00 ,  G06F16/2455 ,  G06F11/34 ,  G06F11/30

Abstract: Embodiments for providing critical path analysis of active trace files in a cloud computing environment. A critical path may be identified using a trace of time spans and activities of a plurality of applications, wherein the critical path is a set of activities having time spans free of overlap with other activities.

公开(公告)号：US20220121921A1

公开(公告)日：2022-04-21

申请号：US17074054

申请日：2020-10-19

Applicant: International Business Machines Corporation ,  University of Illinois at Urbana-Champaign

Inventor： Pin-Yu Chen ,  Yada Zhu ,  Jinjun Xiong ,  Kumar Bhaskaran ,  Yunan Ye ,  Bo Li

IPC: G06N3/08 ,  G06Q40/06 ,  G06Q10/10 ,  G06F40/279

Abstract: A processor training a reinforcement learning model can include receiving a first dataset representing an observable state in reinforcement learning to train a machine to perform an action. The processor receives a second dataset. Using the second dataset, the processor trains a machine learning classifier to make a prediction about an entity related to the action. The processor extracts an embedding from the trained machine learning classifier, and augments the observable state with the embedding to create an augmented state. Based on the augmented state, the processor trains a reinforcement learning model to learn a policy for performing the action, the policy including a mapping from state space to action space.

公开(公告)号：US20210303803A1

公开(公告)日：2021-09-30

申请号：US16830106

申请日：2020-03-25

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  The Board of Trustees of the University of Illinois

Inventor： Lingfei Wu ,  Jinjun Xiong ,  Hongyu Gong ,  Suma Bhat ,  Wen-Mei Hwu

IPC: G06F40/56 ,  G06N3/04 ,  G06F40/35 ,  G06F40/253

Abstract: A computer-implemented method is provided for transferring a target text style using Reinforcement Learning (RL). The method includes pre-determining, by a Long Short-Term Memory (LSTM) Neural Network (NN), the target text style of a target-style natural language sentence. The method further includes transforming, by a hardware processor using the LSTM NN, a source-style natural language sentence into the target-style natural language sentence that maintains the target text style of the target-style natural language sentence. The method also includes calculating an accuracy rating of a transformation of the source-style natural language sentence into the target-style natural language sentence based upon rewards relating to at least the target text style of the source-style natural language sentence.

公开(公告)号：US20190354835A1

公开(公告)日：2019-11-21

申请号：US15982181

申请日：2018-05-17

Applicant: International Business Machines Corporation ,  The Board of Trustees of the University of Illinois

Inventor： Khoi-Nguyen C. Mac ,  Raymond Alexander Yeh ,  Dhiraj Joshi ,  Minh N. Do ,  Rogerio Feris ,  Jinjun Xiong

IPC: G06N3/04 ,  G06N3/08 ,  G06K9/00

Abstract: Techniques regarding action detection based on motion in receptive fields of a neural network model are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in the memory. The computer executable components can comprise a motion component that can extract a motion vector from a plurality of adaptive receptive fields in a deformable convolution layer of a neural network model. The computer executable components can also comprise an action detection component that can generate a spatio-temporal feature by concatenating the motion vector with a spatial feature extracted from the deformable convolution layer.

公开(公告)号：US09836434B2

公开(公告)日：2017-12-05

申请号：US14823889

申请日：2015-08-11

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Alexey Y. Lvov ,  Jinjun Xiong ,  Vladimir Zolotov

IPC: G06F17/16 ,  G06T1/20

CPC classification number: G06F17/16 ,  G06T1/20 ,  G06T2200/28

Abstract: Methods for improving matrix multiplication runtimes are provided. A method includes determining, by a GPU, optimal partitions for matrix-by-matrix multiplication of two factor matrices having sizes known a priori. The determining step includes performing offline a plurality of matrix-by-matrix multiplication executions, each for a respective different combination of two-way partitions across a plurality of partition sizes. The determining step further includes determining offline a respective performance number for each of the executions based on runtime. The determining step also includes recursively repeating offline said performing and determining steps until the respective performance number ceases to improve for best-performing combinations of the two-way partitions and saving the best performing combinations of the two-way partitions as the optimal partitions. The method further includes performing online, by the GPU, the matrix-by-matrix multiplication of the two factor matrices using calls for a given one of the best performing combinations of the two-way partitions.

公开(公告)号：US09665538B2

公开(公告)日：2017-05-30

申请号：US14145300

申请日：2013-12-31

Applicant: International Business Machines Corporation

Inventor： Gradus Janssen ,  Jinjun Xiong

IPC: G06F17/30 ,  G06F17/11 ,  G06F17/50

CPC classification number: G06F17/11 ,  G06F17/5027

Abstract: One embodiment of a method for solving an input satisfiability instance includes searching a database for a stored satisfiability instance that matches the input satisfiability instance and outputting a solution to the input satisfiability instance. One embodiment of method for converting an input satisfiability instance into a standardized representation includes applying a plurality of syntactical simplification rules to the input satisfiability instance until no conditions of any of the plurality of syntactical simplification rules can be met, thereby producing a simplified instance, uniformly replacing each variable in the simplified instance with a unique, consecutively chosen even number, annotating each literal in the simplified instance to indicate whether the each literal is positive or negative, ordering all literals in the simplified instance, and ordering all clauses in the simplified instance to produce the standardized representation.

公开(公告)号：US09599973B2

公开(公告)日：2017-03-21

申请号：US13826419

申请日：2013-03-14

Applicant: International Business Machines Corporation

Inventor： Gary S. Ditlow ,  Brian P. Gaucher ,  Clifford A. Pickover ,  Jinjun Xiong

IPC: G05B15/02

CPC classification number: G05B15/02 ,  G05B2219/2642

Abstract: A method, system and apparatus for operating an energy-using device are disclosed. Current data related to operation of the energy-using device is received at a remote device. An operating specification for the energy-using device is received at the remote device from a database. A recommended setting of the energy-using device is determined from the current data and the operating specification. The remote device communicates the recommended setting to the energy-using device. A control unit at the energy-using device receives the recommended setting and implements the recommended setting at the energy-using device.

公开(公告)号：US20160148171A1

公开(公告)日：2016-05-26

申请号：US14551522

申请日：2014-11-24

Applicant: International Business Machines Corporation

Inventor： Dung Phan ,  Jinjun Xiong

IPC: G06Q10/00 ,  G01M99/00

CPC classification number: G06Q10/06312 ,  G01M99/008 ,  G06Q10/20 ,  Y02P90/86

Abstract: There are provided a system, a method and a computer program product for generating an optimal preventive maintenance/replacement schedule for a set of assets. The method includes receiving data regarding an asset, said data including a failure rate function of said asset, a cost of preventative maintenance (PM) of said asset, a cost of an asset failure, and a cost of replacing an asset. An optimal number K of preventative maintenance time intervals tk and an indication of a possible replacement is computed and stored for each asset by minimizing a mean cost-rate value function with respect to an electrical age of the asset. A first PM schedule is formed without consideration of labor and budget resource constraints. The method further generates a second maintenance schedule for a system of assets by minimizing a deviation from the optimal PM time intervals subject to the labor and budget resource constraints.