专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

61. 发明专利

CA3043825A1 SYSTEMS AND METHODS FOR CLASSIFICATION OF MULTI-DIMENSIONAL TIME SERIES OF PARAMETERS 未知
公开(公告)号：CA3043825A1
公开(公告)日：2020-01-09
申请号：CA3043825
申请日：2019-05-21
申请人： TATA CONSULTANCY SERVICES LTD
发明人： MALHOTRA PANKAJ , GUPTA PRIYANKA , VIG LOVEKESH , SHROFF GAUTAM
IPC分类号： G06N3/02 , G06V30/144 , G06N3/08 , G16H50/70
摘要： Traditional systems and methods have implemented hand-crafted feature extraction from varying length time series that results in complexity and requires domain knowledge. Building classification models requires large labeled data and is computationally expensive. Embodiments of the present disclosure implement learning models for classification tasks in multi-dimensional time series by performing feature extraction from entity's parameters via unsupervised encoder and build a non-temporal linear classifier model. A fixed-dimensional feature vector is outputted using a pre-trained unsupervised encoder, which acts as off the shelf feature extractor. Extracted features are concatenated to learn a non-temporal linear classification model and weight is assigned to each extracted feature during learning which helps to determine relevant parameters for each class. Mapping from parameters to target class is considered while constraining the linear model to use only subset of large number of features.

62. 发明专利

CA3037024A1 FAILED AND CENSORED INSTANCES BASED REMAINING USEFUL LIFE (RUL) ESTIMATION OF ENTITIES 未知
公开(公告)号：CA3037024A1
公开(公告)日：2020-01-09
申请号：CA3037024
申请日：2019-03-18
申请人： TATA CONSULTANCY SERVICES LTD
发明人： MALHOTRA PANKAJ , TV VISHNU , VIG LOVEKESH , SHROFF GAUTAM
IPC分类号： G06N3/02 , G06N3/08
摘要： Estimating Remaining Useful Life (RUL) from multi-sensor time series data is difficult through manual inspection. Current machine learning and data analytics methods, for RUL estimation require large number of failed instances for training, which are rarely available in practice, and these methods cannot use information from currently operational censored instances since their failure time is unknown. Embodiments of the present disclosure provide systems and methods for estimating RUL using time series data by implementing an LSTM-RNN based ordinal regression technique, wherein during training RUL value of failed instance(s) is encoded into a vector which is given as a target to the model. Unlike a failed instance, the exact RUL for a censored instance is unknown. For using the censored instances, target vectors are generated and the objective function is modified for training wherein the trained LSTM-RNN based ordinal regression is applied on an input test time series for RUL estimation.

63. 发明专利

CA2997797C BILSTM-SIAMESE NETWORK BASED CLASSIFIER FOR IDENTIFYING TARGET CLASS OF QUERIES AND PROVIDING RESPONSES THEREOF 未知
公开(公告)号：CA2997797C
公开(公告)日：2019-12-03
申请号：CA2997797
申请日：2018-03-07
申请人： TATA CONSULTANCY SERVICES LTD
发明人： AGARWAL PUNEET , KHURANA PRERNA , SHROFF GAUTAM , VIG LOVEKESH , SRINIVASAN ASHWIN
IPC分类号： G06K9/62 , G06F40/00 , G06F40/20
摘要： Organizations are constantly flooded with questions, ranging from mundane to the unanswerable. It is therefore respective department that actively looks for automated assistance, especially to alleviate the burden of routine, but time-consuming tasks. The embodiments of the present disclosure provide BiLSTM-Siamese Network based Classifier for identifying target class of queries and providing responses to queries pertaining to the identified target class, which acts as an automated assistant that alleviates burden of answering queries in well- defined domains. Siamese Model (SM) is trained for a epochs, and then the same Base-Network is used to train Classification Model (CM) for b epochs iteratively until best accuracy is observed on validation test, wherein SM ensures it learns which sentences are similar/dissimilar semantically while CM learns to predict target class of every user query. Here a and b are assumed to be hyper parameters and are tuned for best performance on the validation set.

64. 发明专利

AU2018201670A1 BiLSTM-SIAMESE NETWORK BASED CLASSIFIER FOR IDENTIFYING TARGET CLASS OF QUERIES AND PROVIDING RESPONSES THEREOF 未知
公开(公告)号：AU2018201670A1
公开(公告)日：2019-03-28
申请号：AU2018201670
申请日：2018-03-07
申请人： TATA CONSULTANCY SERVICES LTD
发明人： AGARWAL PUNEET , KHURANA PRERNA , SHROFF GAUTAM , VIG LOVEKESH , SRINIVASAN ASHWIN
IPC分类号： G06N5/00 , G06F40/00 , G06F40/20 , G06Q10/10
摘要： Organizations are constantly flooded with questions, ranging from mundane to the unanswerable. It is therefore respective department that actively looks for automated assistance, especially to alleviate the burden of routine, but time-consuming tasks. The embodiments of the present disclosure provide BiLSTM-Siamese Network based Classifier for identifying target class of queries and providing responses to queries pertaining to the identified target class, which acts as an automated assistant that alleviates burden of answering queries in well defined domains. Siamese Model (SM) is trained for a epochs, and then the same Base-Network is used to train Classification Model (CM) for b epochs iteratively until best accuracy is observed on validation test, wherein SM ensures it learns which sentences are similar/dissimilar semantically while CM learns to predict target class of every user query. Here a and b are assumed to be hyper parameters and are tuned for best performance on the validation set.

65. 发明专利

AU2018201487A1 METHOD AND SYSTEM FOR HEALTH MONITORING AND FAULT SIGNATURE IDENTIFICATION 未知
公开(公告)号：AU2018201487A1
公开(公告)日：2019-03-07
申请号：AU2018201487
申请日：2018-03-01
申请人： TATA CONSULTANCY SERVICES LTD
发明人： MALHOTRA PANKAJ , TV VISHNU , GUGULOTHU NARENDHAR , VIG LOVEKESH , AGARWAL PUNEET , SHROFF GAUTAM
IPC分类号： G06F11/32 , G06Q99/00
摘要： METHOD AND SYSTEM FOR HEALTH MONITORING AND FAULT SIGNATURE IDENTIFICATION System and method for health monitoring and fault signature identification of a system are disclosed. In an embodiment, the system, estimates Health Index (HI) of the system as time series data. By analyzing data corresponding to the estimated HI, the system identifies one or more time windows in which majority of the estimated HI values are low as a low HI window, and one or more time windows in which majority of the estimated HI values are high as a high HI window. Upon identifying a low HI window, which indicates an abnormal behavior of the system being monitored, the system, based on a local Bayesian Network generated for the system being monitored, generates an Explainability Index (EI) for each sensor, wherein the El quantifies contribution of the sensor to the low HI. Further, associated component(s) is identified as contributing to the abnormal/faulty behavior of the system. Estimating Health Index (HI) of the system as a time series data Identifying a low HI as an abnormal behavior of the system Abnormal behavior No detected? Yes Detecting system component(s) responsible for the abnormal behavior, based on the HI values and sensor readings Fig. 2

66. 发明专利

CA2997489A1 METHOD AND SYSTEM FOR HEALTH MONITORING AND FAULT SIGNATURE IDENTIFICATION 未知
公开(公告)号：CA2997489A1
公开(公告)日：2019-02-18
申请号：CA2997489
申请日：2018-03-06
申请人： TATA CONSULTANCY SERVICES LTD
发明人： MALHOTRA PANKAJ , TV VISHNU , GUGULOTHU NARENDHAR , VIG LOVEKESH , AGARWAL PUNEET , SHROFF GAUTAM
IPC分类号： G01D21/00 , G06N3/02
摘要： System and method for health monitoring and fault signature identification of a system are disclosed. In an embodiment, the system, estimates Health Index (HI) of the system as time series data. By analyzing data corresponding to the estimated HI, the system identifies one or more time windows in which majority of the estimated HI values are low as a low HI window, and one or more time windows in which majority of the estimated HI values are high as a high HI window. Upon identifying a low HI window, which indicates an abnormal behavior of the system being monitored, the system, based on a local Bayesian Network generated for the system being monitored, generates an Explainability Index (EI) for each sensor, wherein the EI quantifies contribution of the sensor to the low HI. Further, associated component(s) is identified as contributing to the abnormal/faulty behavior of the system.

67. 发明专利

CA2967641A1 TRAINING INDUCTIVE LOGIC PROGRAMMING ENHANCED DEEP BELIEF NETWORK MODELS FOR DISCRETE OPTIMIZATION 未知
公开(公告)号：CA2967641A1
公开(公告)日：2018-06-05
申请号：CA2967641
申请日：2017-05-16
申请人： TATA CONSULTANCY SERVICES LTD
发明人： SAIKIA SARMIMALA , VIG LOVEKESH , SHROFF GAUTAM , AGARWAL PUNEET , RAWAT RICHA , SRINIVASAN ASHWIN
IPC分类号： G06F15/18 , G06N3/02 , G06N3/08
摘要： System and method for training inductive logic programming enhanced deep belief network models for discrete optimization are disclosed. The system initializes (i) a dataset comprising values and (ii) a pre-defined threshold, partitions the values into a first set and a second set based on the pre-defined threshold. Using Inductive Logic Programming (ILP) engine and a domain knowledge associated with the dataset, a machine learning model is constructed on the first set and the second set to obtain Boolean features, and using the Boolean features that are being appended to the dataset, a deep belief network (DBN) model is trained to identify an optimal set of values between the first set and the second set. Using the trained DBN model, the optimal set of values are sampled to generate samples. The pre-defined threshold is adjusted based on the generated samples, and the steps are repeated to obtain optimal samples.

68. 发明专利

CA2873834C SYSTEM AND METHOD FOR CREATING STRUCTURED EVENT OBJECTS 未知
公开(公告)号：CA2873834C
公开(公告)日：2017-07-18
申请号：CA2873834
申请日：2013-03-22
申请人： TATA CONSULTANCY SERVICES LTD
发明人： AGARWAL PUNEET , VAITHIYANATHAN RAJGOPAL , SHROFF GAUTAM
IPC分类号： G06F17/27 , G06F19/00
摘要： The present invention envisages a system and method for converting a voluminous stream of unstructured short text messages into event-objects of specific event type that may be of potential interest to users at far away locations. The method of structuring involves detecting long tail of events in spite of their sparsity. This is followed by extracting and correlating detected short text messages that describe same event type to create structured event objects.

69. 发明专利

AU2016201689A1 Methods and systems for searching logical patterns 未知
公开(公告)号：AU2016201689A1
公开(公告)日：2017-01-12
申请号：AU2016201689
申请日：2016-03-17
申请人： TATA CONSULTANCY SERVICES LTD
发明人： HASSAN EHTESHAM , YADAV MOHIT , AGARWAL PUNEET , SHROFF GAUTAM , SRINIVASAN ASHWIN
IPC分类号： G06K11/00 , G07C5/00
摘要： METHODS AND SYSTEMS FOR SEARCHING LOGICAL PATTERNS Methods and systems for searching logical patterns in voluminous multi sensor data from the industrial internet is provided. The method retrieves instances of patterns in time-series data where patterns are specified logically, using a sequence of symbols. The logical symbols used are a subset of the qualitative abstractions specifically, the concepts of steady, increasing, decreasing. Patterns can include symbol-sequences for multiple sensors, approximate duration as well as slope values for each symbol. To facilitate efficient querying, each sensor time-series is pre-processed into a sequence of logical symbols. Each position in the resulting compressed sequence is registered across a TRIE-based index structure corresponding to the multiple logical patterns it may belong to. Logical multi-sensor patterns are efficiently retrieved and ranked using such a structure. This method of indexing and searching provides an efficient mechanism for exploratory analysis of voluminous multi-sensor data. Use FIG.1 Pattern Retriever 30 Pattern Generator 28 Ranking Engine 26 L- Search Engine 24 Disk Memory Classifier Extractor Pre-processor Sensors Time- series data Input Module FIG.tI

70. 发明专利

AU2015268759A1 Interpretation of a dataset 未知
公开(公告)号：AU2015268759A1
公开(公告)日：2016-07-07
申请号：AU2015268759
申请日：2015-12-16
申请人： TATA CONSULTANCY SERVICES LTD
发明人： AGARWAL PUNEET , SHROFF GAUTAM , SAIKIA SARMIMALA , SRINIVASAN ASHWIN
IPC分类号： G06F17/00
摘要： INTERPRETATION OF A DATASET A method and a system for interpreting a dataset comprising a plurality of items is described herein. The method may include computing a rule set pertaining to the dataset, generating a rule cover, calculating a plurality of distances between the plurality of rule pairs in the rule cover and generating a distance matrix based on the calculated plurality of distances between the plurality of rule pairs, storing the calculated plurality of distances between the plurality of rule pairs, clustering the overlapping rules within the rule cover using the distance matrix; selecting a representative rule from each cluster, determining at least one exception for each representative rule in the rule cover selected from each cluster and interpreting the dataset using the representative rules and the at least one exception determined for each representative rule in the rule set. To be published with Fig. 1 104-1 DATA INTERPRETATIONSYSTEM 104-2 PROCESSOR(S)1 TFA S MODULE(S) 11M RULE GENERATION MODULE12f 102 NETWORK INTERPRETATIONIMODULE122 OTHER MODULE(S) 1210 DATA l= RULE SET DATA 12& INTERPRETATION DATAl2f OTHER DATA lQ

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式