In many types of systems (e.g., manufacturing systems, production systems, etc.), maintenance costs may constitute a major part of the total operating costs. Generally, the aim of an optimal maintenance policy is to provide optimal machine and/or plant availability, reliability, and safety at the lowest possible cost. Different systems may use any number of different failure identification procedures. Examples of such procedures include failure mode effects analysis (FMEA) procedures; failure mode, effects and criticality analysis (FMECA) procedures; fault tree analysis (FTA) procedures; etc. In some instances, failure identification procedures and design of experiments are being used for quality control and for the detection of potential failure modes during the detail design stage and/or the post product launch stage.
In some embodiments, a non-transitory machine-readable medium stores a program. The program retrieves a set of notifications. Each notification includes a set of words describing a failure that occurred on a monitored device in a set of monitored devices. The program further determines a set of topics based on the sets of words of the set of notifications. For each topic in the set of topics, the program also determines a failure mode associated with the topic from a plurality of failure modes defined for the set of monitored devices. For each notification in the set of notifications, the program further determines a failure mode associated with the notification based on the set of topics and the failure modes associated with the set of topics. The program also receives a particular notification that includes a particular set of words describing a failure that occurred on a particular monitored device in the set of monitored devices. The program further determines a failure mode associated with the particular notification based on the particular set of words and the determined failure modes associated with the set of notifications.
In some embodiments, a topic in the set of topics may include a set of related words or phrases that occur in the set of words of the set of notifications. Determining the set of topics may be further based on a Latent Dirichlet allocation (LDA) model. Determining, for each topic in the set of topics, the failure mode associated with the topic from the plurality of failure modes defined for the set of monitored devices may be based on a Latent Dirichlet allocation (LDA) model.
In some embodiments, determining, for each notification in the set of notifications, the failure mode associated with the notification may include determining a topic in the set of topics associated with the notification and determining the failure mode associated with the notification based on the failure modes associated with the set of topics. Determining, for each notification in the set of notifications, the topic in the set of topics associated with the notification may be based on a Latent Dirichlet allocation (LDA) model. Determining, for each topic in the set of topics, the failure mode associated with the topic from the plurality of failure modes defined for the set of monitored devices may include determining a device model associated with the set of monitored devices, determining that the plurality of failure mode definitions are defined for the device model, and retrieving the plurality of failure modes from a storage configured to store failure mode definitions.
In some embodiments, a method retrieves a set of notifications. Each notification includes a set of words describing a failure that occurred on a monitored device in a set of monitored devices. The method further determines a set of topics based on the sets of words of the set of notifications. For each topic in the set of topics, the method also determines a failure mode associated with the topic from a plurality of failure modes defined for the set of monitored devices. For each notification in the set of notifications, the method further determines a failure mode associated with the notification based on the set of topics and the failure modes associated with the set of topics. The method also receives a particular notification comprising a particular set of words describing a failure that occurred on a particular monitored device in the set of monitored devices. The method further determines a failure mode associated with the particular notification based on the particular set of words and the determined failure modes associated with the set of notifications.
In some embodiments, a topic in the set of topics may include a set of related words or phrases that occur in the set of words of the set of notifications. Determining the set of topics may be further based on a Latent Dirichlet allocation (LDA) model. Determining, for each topic in the set of topics, the failure mode associated with the topic from the plurality of failure modes defined for the set of monitored devices may be based on a Latent Dirichlet allocation (LDA) model.
In some embodiments, determining, for each notification in the set of notifications, the failure mode associated with the notification may include determining a topic in the set of topics associated with the notification and determining the failure mode associated with the notification based on the failure modes associated with the set of topics. Determining, for each notification in the set of notifications, the topic in the set of topics associated with the notification may be based on a Latent Dirichlet allocation (LDA) model. Determining, for each topic in the set of topics, the failure mode associated with the topic from the plurality of failure modes defined for the set of monitored devices may include determining a device model associated with the set of monitored devices, determining that the plurality of failure mode definitions are defined for the device model, and retrieving the plurality of failure modes from a storage configured to store failure mode definitions.
In some embodiments, a system includes a set of processing units and a non-transitory machine-readable medium that stores instructions. The instructions cause at least one processing unit to retrieve a set of notifications. Each notification includes a set of words describing a failure that occurred on a monitored device in a set of monitored devices. The instructions further cause the at least one processing unit to determine a set of topics based on the sets of words of the set of notifications. For each topic in the set of topics, The instructions also cause the at least one processing unit to determine a failure mode associated with the topic from a plurality of failure modes defined for the set of monitored devices. For each notification in the set of notifications, The instructions further cause the at least one processing unit to determine a failure mode associated with the notification based on the set of topics and the failure modes associated with the set of topics. The instructions also cause the at least one processing unit to receive a particular notification comprising a particular set of words describing a failure that occurred on a particular monitored device in the set of monitored devices. The instructions further cause the at least one processing unit to determine a failure mode associated with the particular notification based on the particular set of words and the determined failure modes associated with the set of notifications.
In some embodiments, a topic in the set of topics may include a set of related words or phrases that occur in the set of words of the set of notifications. Determining the set of topics may be further based on a Latent Dirichlet allocation (LDA) model. Determining, for each topic in the set of topics, the failure mode associated with the topic from the plurality of failure modes defined for the set of monitored devices may be based on a Latent Dirichlet allocation (LDA) model.
In some embodiments, determining, for each notification in the set of notifications, the failure mode associated with the notification may include determining a topic in the set of topics associated with the notification and determining the failure mode associated with the notification based on the failure modes associated with the set of topics. Determining, for each notification in the set of notifications, the topic in the set of topics associated with the notification may be based on a Latent Dirichlet allocation (LDA) model.
The following detailed description and accompanying drawings provide a better understanding of the nature and advantages of the present invention.
In the following description, for purposes of explanation, numerous examples and specific details are set forth in order to provide a thorough understanding of the present invention. It will be evident, however, to one skilled in the art that the present invention as defined by the claims may include some or all of the features in these examples alone or in combination with other features described below, and may further include modifications and equivalents of the features and concepts described herein.
Described herein are techniques for determining failure modes of devices based on text analysis. In some embodiments, a system utilizes a two-stage data training approach to determining failure modes of devices. In the first data training stage (also referred to as an unsupervised training stage), the system may retrieve and analyze text or documents describing failures that have occurred on devices. Such text may be in the form of notifications, messages, logs, etc., that a user provides to the system when the user is determining the cause of the failures of the devices. Based on the analysis of the text, the system may determine one or more topics using a topic model. In some embodiments, a topic specifies a group of related terms (e.g., words, phrases, etc.) or clusters of similar terms that occur in the text. In some embodiments, a topic model captures this intuition in a mathematical framework, which allows examining a set of documents and discovering based on statistics of the terms in each document, what the topics might be and what each document's balance of such a topic might be. The system may then retrieve failure mode definitions and determine, for with each topic, a failure mode defined in the failure mode definitions that is associated with the topic. The system can infer underlying topics that explain a specific instance of text based on the generative process which was used to train the model. In some cases, this is achieved by inferring the conditional distribution (posterior) of the hidden variables given the observed variables. In other words, the underlying topic distribution that coherently explains the specific failure mode is determined based on the generative process that was used to train the model.
In the first data training stage, the system may determine, for each instance of text (e.g., the text of a notification, a text of a message, a text of a log, etc.) used to determine the one or more topics described above, a topic that is associated with the instance of text. Next, the system may determine, by inference, a failure mode, for each instance of text, that is associated with the instance of text based on the analysis of the text mentioned above. Once the first data training stage is complete, the system may use the data generated from the first data training stage to curate a labeled dataset of notification texts and failure modes as labels, which is then used for a second supervised training stage. The second training stage includes the usage of text classification techniques in order learn a model that determines categories or labels for documents. By means of scoring on the learnt classification model, the system determines failure modes that are associated with any new notifications, messages, logs, etc. received by the system that have text describing failures that have occurred on devices.
In some embodiments, the system uses a Latent Dirichlet allocation (LDA) model for various aspects of data training. Latent Dirichlet Allocation (LDA) model may be a Bayesian mixture model for discrete data where topics are assumed to be uncorrelated. LDA assumes that each document can be represented as a probabilistic distribution over latent topics, and that topic distribution in all documents share a common Dirichlet prior. Each latent topic in the LDA model can be represented as a probabilistic distribution over words and the word distributions of topics share a common Dirichlet prior as well.
The techniques described in the present application provide a number of benefits and advantages over conventional data processing systems. For example, a system that uses a two stage data training approach can improve the accuracy of failure modes determined to be associated with device failure data received by the system that have text describing failures that have occurred on devices. Additionally, such a system provides an automated approach to failure mode identification based on information embedded in texts, thereby enabling faster processing of notifications and aiding in root cause analysis.
Client device 110 is configured to communicate and interact with computing system 115. For example, a user of client device 110 may send computing system 115 data associated with monitored devices 105a-n. The data associated with a monitored device 105 may include text describing failures that have occurred on the monitored device 105. In some cases, a user of client device 110 can provide validation for data generated and/or determined by computing system 115 during various stages of data training. A user of client device 110 may send computing system 115 requests for data associated with monitored devices 105a-n (e.g., device failure data, device failure determinations, device failure modes, etc.).
Topics data storage 140 is configured to store topics determined and generated by topic modeler 126. In some embodiments, a topic specifies a set of related terms (e.g., words, phrases, etc.). Device failure data storage 145 stores device failure data associated with monitored devices 105a-n. Device failure data associated with a monitored device 105 may include text describing failures that have occurred on the monitored device 105. In some embodiments, the text of device failure data associated with a monitored device 105 may be in the form of notifications, messages, logs, etc. Failure mode definitions storage 150 is configured to store failure mode definitions. In some embodiments, a failure mode definition specifies a failure mode for a particular device type or device model. In some such embodiments, a failure mode describes the manner in which a particular device or device model may fail functionally. In some cases (e.g., where monitored devices 105a-n are the same device model), failure mode definitions stored in failure mode definitions storage 150 can be provided by the manufacturer of monitored devices 105a-n. In other cases, failure mode definitions stored in failure mode definitions storage 150 are provided from an industry standard. For example, in petroleum, petrochemical and/or natural gas industries, failure mode definitions can be provided by an International Organization for Standardization (ISO) 14224 standard. Device failure determinations storage 155 stores failure modes determined by notifications processor 135 to be the cause of failures of monitored devices 105a-n. In some embodiments, storages 140-155 are implemented in a single physical storage while, in other embodiments, storages 140-155 may be implemented across several physical storages. While
As illustrated in
As shown, training engine 125 includes topic modeler 126, inference manager 127, and text classifier 128. Topic modeler 126 is responsible for initiating the first data training stage. In some cases, topic modeler 126 initiates the first stage at defined intervals (e.g., once a day, once a week, once a month, etc.). In other cases, topic modeler 126 initiates the first stage in response to a request to start data training (e.g., from a user of client device 105). In the first stage of data training, topic modeler 126 may determine a set of topics based on device failure data. In some embodiments, topic modeler 126 determines the set of topics by retrieving from device failure data storage 145 device failure data associated with monitored devices 105a-n that include text describing failures that have occurred on monitored devices 105a-n. In some such embodiments, topic modeler 126 retrieves device failure data associated with monitored devices 105a-n from device failure data storage 145 that is from a defined period of time (e.g., device failure data associated with monitored devices 105a-n describing failures that occurred within the last six months, one year, two years, etc.).
Next, topic modeler 126 employs a Latent Dirichlet allocation (LDA) model to determine the set of topics based on all the text from the retrieved device failure data associated with monitored devices 105a-n. In some embodiments, topic modeler 126 first determines the number of topics to in the set of topics. To determine the optimum number of topics, topic modeler 126 may determine the mean probabilistic semantic coherence of different numbers of inferred topics based on all the text by running multiple LDA models against the retrieved device failure data associated with monitored devices 105a-n. Topic modeler 126 can then select the LDA model having the highest mean probabilistic coherence. The number of topics used to generate this model is then chosen as the optimum number of topics. The coherence score is the mean pairwise Cosine similarity between the vectors corresponding to the top terms describing the topic.
In some embodiments, the system utilizes the following equation (1) to determine the coherence of a topic:
where D(v) is the document frequency of word v (i.e., the number of documents with least one token of type v) and V(t)=(v1(t), . . . , vm(t)) is the list of the M most probable words in topic t. This metric, which relies upon word co-occurrence statistics gathered from the corpus being modeled rather than an external reference corpus, is both domain-specific, and does not require additional reference data.
Once topic modeler 126 determines the number of topics in the set of topics, topic modeler 126 uses the LDA model to determine the set of topics based on all the text from the retrieved device failure data associated with monitored devices 105a-n. In some embodiments, the LDA model utilizes the following equation (2) to determine the set of topics:
where α and β are corpus-level parameters (e.g., topic-level parameters) sampled once in the process of generating a topic. A corpus is a collection of M documents denoted by D={w1, w2, . . . , wdn}. θd are document-level variables (e.g., notification-level, message-level, log-level, etc. parameters) that are sampled once per document. Nd are N topics in single documents d. Variables zdn and wdn are word-level variables that are sampled once for each word in each document.
As mentioned above, a topic specifies a group of related terms (e.g., words, phrases, etc.) in some embodiments. Here, each term in the group of related terms of a topic determined by topic modeler 126 occurs in the text from the device failure data associated with monitored devices 105a-n. That is, each term in the group of related terms of a topic occurs in at least one notification, message, log, etc. (but not necessarily every notification, message, log, etc.) in the retrieved device failure data associated with monitored devices 105a-n. In some embodiments, each topic determined by topic modeler 126 is a distribution of words in the text from the device failure data associated with monitored devices 105a-n. Once topic modeler 126 determines the set of topics, topic modeler 126 stores them in topics data 140.
After topic modeler 126 determines the set of topics, inference manager 127 retrieves from failure mode definitions storage 150 failure mode definitions defined for monitored devices 105a-n. Inference manager 127 can then determine, for each topic in the set of topics, a failure mode from the retrieved failure mode definitions that is associated with the topic. In some embodiments, inference manager 127 determines a failure mode from the retrieved failure mode definitions that is associated with a topic by determining the similarity between the topic and each failure mode and then selecting the failure mode having the highest similarity with the topic to be associated with the topic. Determining the similarity between a topic and a failure mode may be referred to as computing the distances between the topic and the failure modes. In some instances, inference manager 127 determines the similarity between a topic and a failure mode by computing the posterior prior probability between the topic and the failure mode. In other instances, inference manager 127 determines the similarity between a topic and a failure mode by computing the cosine similarities between the topic and the failure mode. In some embodiments, inference manager 127 uses an LDA model to determine a failure mode for a topic. For example, inference manager 127 may infer the conditional distribution (posterior) of the hidden variables given the observed variables from the learned LDA Model. One approach is to use the following equation (3) to determine a topic distribution for failure mode using collapsed Gibbs Sampler:
where p(w) represents the posterior distribution of the hidden variables, given a document, given α and β, the corpus-level parameter. θi are document-level variables representing distribution of topics in document. k is the number of topics, V is the number of words in the vocabulary. wnt is the identity of jth word in nth document. Once inference manager 127 determines a failure mode for each topic in the set of topics, inference manager 127 stores the determinations in topics storage 140 and sends device failure manager 130 a notification that training engine 125 has completed the first stage of data training. In some embodiments, inference manager 127 sends application 120 a request to present the determinations to client device 110 in order to allow a user of client device 110 to validate the determinations made by topics training engine 125.
Device failure manager 130 handles the second data training stage. Upon receiving a notification from training engine 125 that training engine 125 has completed the first stage of data training, device failure manager 130 starts the second data training stage by retrieving device failure data from device failure data storage 145 that is associated with monitored devices 105a-n. In some embodiments, device failure manager 130 retrieves the same device failure data from device failure data storage 145 that was used by training engine 125 in the first stage of the data training. Next, device failure manager 130 retrieves the set of topics determined by topics engine 125 from topics data storage 140. The retrieved device failure data and the retrieved set of topics can constitute an input training dataset of labeled failure data.
Device failure manager 130 is now able to join the topics per device failure in topics storage data 140 determined by topic modeler 126 during the unsupervised topic modeling, with the topics per failure mode in topics data storage 140 determined by inference manager 127 during the inference operations. Device failure manager 130 now can prepare the input training dataset for the second stage of training. Device failure manager 130 thus determines a failure mode that is associated with each notification, message, log, etc. in the retrieved device failure data based on the set of topics and failure modes associated with the set of topics determined in the first stage of the data training. In some embodiments, device failure manager 130 determines a failure mode that is associated with a notification by determining a topic in the set of topics that is associated with the notification and then determining the failure mode that is associated with the notification based on the failure modes associated with the set of topics that topic modeler 126 determined in the first stage of the data training. In some such embodiments, device failure manager 130 uses an LDA model to determine a topic associated with a notification. After device failure manager 130 determines a failure mode that is associated with each notification, message, log, etc. in the retrieved device failure data, device failure manager 130 stores the determinations in device failure data storage 145.
Text classifier 127 uses failure data that have text describing failures that have occurred on monitored devices 105a-n and known failure modes in order to perform the second stage of training. Classification technique training is a type of machine learning technique where a training set of correctly identified observations is available. In some embodiments, this dataset is also termed as labeled dataset, implying category assignment of each record of the data is known. The objective of the classification algorithm is then to learn classifiers from the known examples, so that category assignment can be automatically done for a new example, for which category assignment is unknown. In some systems, the simplest form of text classification, uses the concept of building features using a bag of words model, where the classifier would learn the presence of words in a text. Content-based classification is classification in which the weight given to particular subjects in a document determines the class to which the document is assigned. A common methodology used to do this is TF-IDF (term frequency—inverse document frequency). The TF-IDF weighting for a word increases with the number of times the word appears in the document but decreases based on how frequently the word appears in the entire document set. This can be mathematically represented via the following equation (4):
where TF(i,j) means how many times does the ith word occurs in the jth document, N is the number of documents, and DF(i) counts the documents containing the ith word at least once. The transformed documents now form the document term matrix. After pre-processing and transformations, a machine learning algorithm is used for learning how to classify documents, i.e. creating a model for input-output mappings. Several classification models can be used for the classification. However, linear models such as Support Vector Machines, Logistic Regression are expressive enough to achieve good results.
Support Vector Machine (SVM) is primarily a classier method that performs classification tasks by constructing hyperplanes in a multidimensional space that separates cases of different class labels. SVM searches for a separating hyperplane, which separates positive and negative examples from each other with maximal margin. Such a hyperplane can be expressed via the following equation (5):
w
T
x+b=0
where wT is the decision hyperplane normal vector, x is the datapoint in the training example, y is the class of the data point, b is a scalar value representing the intercept. Text classifier 128 learns the optimal hyperplane which separates one class from the rest of the classes. The one-versus-rest (1VR) approach constructs k separate binary classifiers for k-class classification. For example, for monitored device 105, K classes here are represented by the total number of failure mode definitions storage 150 failure mode definitions defined for the monitored device 105.
Scoring is also called prediction, and is the process of generating values based on a trained machine learning model, given some new input data. The values or scores that are created can represent predictions of future values, but they might also represent a likely category or outcome. In some embodiments, scoring refers to predicting the most likely failure mode assignment for incoming device failure data based on the model that is trained by text classifier engine 128.
Scoring engine 135 applies the trained models learnt via text classifier 128 in order to determine failure modes that are associated with incoming device failure data (e.g., notifications, message, logs, etc.) that have text describing failures that have occurred on monitored devices 105a-n. For example, scoring engine 135 may receive from application 120 an instance (e.g., a notification, a message, a log, etc.) of device failure data that has text describing a failure that has occurred on a monitored device 105. In response, scoring engine 135 applies the model that has been trained by training engine 125c, which has learnt a classifier that has determined the failure mode from failure mode definitions defined for the monitored device 105 stored in failure mode definitions storage 150. Next, scoring engine 135 stores the instance of device failure data and the determination in device failure determinations storage 155. In some instances, scoring engine 135 may receive from application 120 a request for determined failure modes for failures described in new notifications, messages, logs, etc. In response to such a request, scoring engine 135 retrieves the requested data from device failure determinations storage 155 and sends it to application 120.
Classification via support vector machines (svm) of an unseen test example x is based on the sign of wTx+b. The separator property can be formalized via the following equation (6):
w
T
x
t
+b≥1tffyi=+1
w
T
x
t
+b≤1tffyi=−1
where wT is the decision hyperplane normal vector, xi is the ith datapoint, yi is the class of the ith data point, b is a scalar value representing the intercept.
An example operation of system 100 will now be described by reference to
Continuing with the example operation, topic modeler 126 uses an LDA model to determine a set of topics based on all the text from the retrieved device failure data associated with monitored devices 105a-n. In some embodiments, the LDA model may determine the number of topics in the set of topics by determining the mean probabilistic coherence of different numbers of topics based on all the text from the retrieved device failure data associated with monitored devices 105a-n using equation (1) and then selecting the number of topics having the highest mean probabilistic coherence as the number of topics. Then topic modeler 126 uses the equation (2) describe above to determine the set of topics and then stores the set of topics in topics data storage 145.
After determining the set of topics, inference manager 127 then retrieves from failure mode definitions storage 150 failure mode definitions defined for monitored devices 105a-n.
Next, inference manager 127 determines, for each topic in the set of topics, a failure mode from the retrieved failure mode definitions that is associated with the topic. As described above, inference manager 127 can determine a failure mode from the retrieved failure mode definitions that is associated with a topic by either determining the similarity between the topic and each failure mode and then selecting the failure mode having the highest similarity with the topic to be associated with the topic or determining the similarity between a topic and a failure mode by computing the posterior prior probability between the topic and the failure mode. In this example, inference manager 128 uses inference based on a trained LDA model learnt by topic modeler 126 to determine a failure mode for a topic according to equation (2) described above. After determining a failure mode for each topic in the set of topics, inference manager 127 stores the determinations in topics storage 140 and sends device failure manager 130 a notification that training engine 125 has completed the first stage of data training.
When device failure manager 130 receives the notification from training engine 125 that training engine 125 has completed the first stage of data training, device failure manager 130 starts the second data training stage. To start the second data training stage, device failure manager 130 retrieves from device failure data storage 145 the same device failure data from device failure data storage 145 that was used by training engine 125 in the first stage of the data training. In this example, device failure manager 130 table 200. Device failure manager 130 then retrieves the set of topics determined by topics engine 125 from topics data storage 140. For this example, the set of topics that device failure manager 130 retrieves is the data illustrated in table 300 of
Next, device failure manager 130 determines a failure mode that is associated with each notification, message, log, etc. in the retrieved device failure data based on the set of topics and failure modes associated with the set of topics determined by training engine 125. For this example, device failure manager 130 determines a failure mode that is associated with a notification by using an LDA model to determine a topic in the set of topics that is associated with the notification.
After determining a topic in the set of topics that is associated with each notification, device failure manager 130 determines a failure mode that is associated with each notification based on the failure modes associated with the set of topics that inference manager 127 determined in the first stage of the data training. For this example, device failure manager 130 determines a failure mode that is associated with each notification based on the data illustrated in table 300 of
Once the two stages of data training are complete, computing system 115 can process new device failure data (e.g., notifications, messages, logs, etc.) that have text describing failures that have occurred on monitored devices 105a-n. For instance, a user of client device 110 may send application 110 an instance (e.g., a notification, a message, a log, etc.) of device failure data that has text describing a failure that has occurred on a monitored device 105. Application 120 can store the instance of device failure data in device failure mode data storage 145 and then forward the instance of device failure data to scoring engine 135 for processing. Upon receiving the instance of device failure data describing a failure that has occurred on a monitored device 105, scoring engine 135 retrieves the trained model from topics data storage 140. Then, scoring engine 135 can determine, based on the retrieved failure mode definitions and trained data, a failure mode that is associated with the instance of device failure data in the same manner described above. After determining the failure mode, scoring engine 135 stores the instance of device failure data along with the determination in device failure determinations storage 155.
Next, process 800 determines, at 820, a set of topics based on the sets of words of the set of notifications. In some embodiments, process 800 performs operation 820 as part of the first data training stage. Referring to
For each notification in the set of notifications, process 800 then determines, at 840, a failure mode associated with the notification based on the set of topics and the failure modes associated with the set of topics. In some embodiments, process 800 performs operation 840 as part of the second data training stage. Referring to
Process 800 then receives, at 860, a particular notification comprising a particular set of words describing a failure that occurred on a particular monitored device in the set of monitored devices. In some embodiments, process 800 performs operation 850 after data training is finished. Referring to
Finally, process 800 determines, at 870, a failure mode associated with the particular notification based on the particular set of words and the determined failure modes associated with the set of notifications. Referring to
Bus subsystem 926 is configured to facilitate communication among the various components and subsystems of computer system 900. While bus subsystem 926 is illustrated in
Processing subsystem 902, which can be implemented as one or more integrated circuits (e.g., a conventional microprocessor or microcontroller), controls the operation of computer system 900. Processing subsystem 902 may include one or more processors 904. Each processor 904 may include one processing unit 906 (e.g., a single core processor such as processor 904-1) or several processing units 906 (e.g., a multicore processor such as processor 904-2). In some embodiments, processors 904 of processing subsystem 902 may be implemented as independent processors while, in other embodiments, processors 904 of processing subsystem 902 may be implemented as multiple processors integrate into a single chip or multiple chips. Still, in some embodiments, processors 904 of processing subsystem 902 may be implemented as a combination of independent processors and multiple processors integrated into a single chip or multiple chips.
In some embodiments, processing subsystem 902 can execute a variety of programs or processes in response to program code and can maintain multiple concurrently executing programs or processes. At any given time, some or all of the program code to be executed can reside in processing subsystem 902 and/or in storage subsystem 910. Through suitable programming, processing subsystem 902 can provide various functionalities, such as the functionalities described above by reference to process 800, etc.
I/O subsystem 908 may include any number of user interface input devices and/or user interface output devices. User interface input devices may include a keyboard, pointing devices (e.g., a mouse, a trackball, etc.), a touchpad, a touch screen incorporated into a display, a scroll wheel, a click wheel, a dial, a button, a switch, a keypad, audio input devices with voice recognition systems, microphones, image/video capture devices (e.g., webcams, image scanners, barcode readers, etc.), motion sensing devices, gesture recognition devices, eye gesture (e.g., blinking) recognition devices, biometric input devices, and/or any other types of input devices.
User interface output devices may include visual output devices (e.g., a display subsystem, indicator lights, etc.), audio output devices (e.g., speakers, headphones, etc.), etc. Examples of a display subsystem may include a cathode ray tube (CRT), a flat-panel device (e.g., a liquid crystal display (LCD), a plasma display, etc.), a projection device, a touch screen, and/or any other types of devices and mechanisms for outputting information from computer system 900 to a user or another device (e.g., a printer).
As illustrated in
As shown in
Computer-readable storage medium 920 may be a non-transitory computer-readable medium configured to store software (e.g., programs, code modules, data constructs, instructions, etc.). Many of the components (e.g., application 120, training engine 125, device failure manager 130, and scoring engine 135) and/or processes (e.g., process 800) described above may be implemented as software that when executed by a processor or processing unit (e.g., a processor or processing unit of processing subsystem 902) performs the operations of such components and/or processes. Storage subsystem 910 may also store data used for, or generated during, the execution of the software.
Storage subsystem 910 may also include computer-readable storage medium reader 922 that is configured to communicate with computer-readable storage medium 920. Together and, optionally, in combination with system memory 912, computer-readable storage medium 920 may comprehensively represent remote, local, fixed, and/or removable storage devices plus storage media for temporarily and/or more permanently containing, storing, transmitting, and retrieving computer-readable information.
Computer-readable storage medium 920 may be any appropriate media known or used in the art, including storage media such as volatile, non-volatile, removable, non-removable media implemented in any method or technology for storage and/or transmission of information. Examples of such storage media includes RAM, ROM, EEPROM, flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disk (DVD), Blu-ray Disc (BD), magnetic cassettes, magnetic tape, magnetic disk storage (e.g., hard disk drives), Zip drives, solid-state drives (SSD), flash memory card (e.g., secure digital (SD) cards, CompactFlash cards, etc.), USB flash drives, or any other type of computer-readable storage media or device.
Communication subsystem 924 serves as an interface for receiving data from, and transmitting data to, other devices, computer systems, and networks. For example, communication subsystem 924 may allow computer system 900 to connect to one or more devices via a network (e.g., a personal area network (PAN), a local area network (LAN), a storage area network (SAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), a global area network (GAN), an intranet, the Internet, a network of any number of different types of networks, etc.). Communication subsystem 924 can include any number of different communication components. Examples of such components may include radio frequency (RF) transceiver components for accessing wireless voice and/or data networks (e.g., using cellular technologies such as 2G, 3G, 4G, 5G, etc., wireless data technologies such as Wi-Fi, Bluetooth, ZigBee, etc., or any combination thereof), global positioning system (GPS) receiver components, and/or other components. In some embodiments, communication subsystem 924 may provide components configured for wired communication (e.g., Ethernet) in addition to or instead of components configured for wireless communication.
One of ordinary skill in the art will realize that the architecture shown in
Processing system 1002, which can be implemented as one or more integrated circuits (e.g., a conventional microprocessor or microcontroller), controls the operation of computing device 1000. As shown, processing system 1002 includes one or more processors 1004 and memory 1006. Processors 1004 are configured to run or execute various software and/or sets of instructions stored in memory 1006 to perform various functions for computing device 1000 and to process data.
Each processor of processors 1004 may include one processing unit (e.g., a single core processor) or several processing units (e.g., a multicore processor). In some embodiments, processors 1004 of processing system 1002 may be implemented as independent processors while, in other embodiments, processors 1004 of processing system 1002 may be implemented as multiple processors integrate into a single chip. Still, in some embodiments, processors 1004 of processing system 1002 may be implemented as a combination of independent processors and multiple processors integrated into a single chip.
Memory 1006 may be configured to receive and store software (e.g., operating system 1022, applications 1024, I/O module 1026, communication module 1028, etc. from storage system 1020) in the form of program instructions that are loadable and executable by processors 1004 as well as data generated during the execution of program instructions. In some embodiments, memory 1006 may include volatile memory (e.g., random access memory (RAM)), non-volatile memory (e.g., read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory, etc.), or a combination thereof.
I/O system 1008 is responsible for receiving input through various components and providing output through various components. As shown for this example, I/O system 1008 includes display 1010, one or more sensors 1012, speaker 1014, and microphone 1016. Display 1010 is configured to output visual information (e.g., a graphical user interface (GUI) generated and/or rendered by processors 1004). In some embodiments, display 1010 is a touch screen that is configured to also receive touch-based input. Display 1010 may be implemented using liquid crystal display (LCD) technology, light-emitting diode (LED) technology, organic LED (OLED) technology, organic electro luminescence (OEL) technology, or any other type of display technologies. Sensors 1012 may include any number of different types of sensors for measuring a physical quantity (e.g., temperature, force, pressure, acceleration, orientation, light, radiation, etc.). Speaker 1014 is configured to output audio information and microphone 1016 is configured to receive audio input. One of ordinary skill in the art will appreciate that I/O system 1008 may include any number of additional, fewer, and/or different components. For instance, I/O system 1008 may include a keypad or keyboard for receiving input, a port for transmitting data, receiving data and/or power, and/or communicating with another device or component, an image capture component for capturing photos and/or videos, etc.
Communication system 1018 serves as an interface for receiving data from, and transmitting data to, other devices, computer systems, and networks. For example, communication system 1018 may allow computing device 1000 to connect to one or more devices via a network (e.g., a personal area network (PAN), a local area network (LAN), a storage area network (SAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), a global area network (GAN), an intranet, the Internet, a network of any number of different types of networks, etc.). Communication system 1018 can include any number of different communication components. Examples of such components may include radio frequency (RF) transceiver components for accessing wireless voice and/or data networks (e.g., using cellular technologies such as 2G, 3G, 4G, 5G, etc., wireless data technologies such as Wi-Fi, Bluetooth, ZigBee, etc., or any combination thereof), global positioning system (GPS) receiver components, and/or other components. In some embodiments, communication system 1018 may provide components configured for wired communication (e.g., Ethernet) in addition to or instead of components configured for wireless communication.
Storage system 1020 handles the storage and management of data for computing device 1000. Storage system 1020 may be implemented by one or more non-transitory machine-readable mediums that are configured to store software (e.g., programs, code modules, data constructs, instructions, etc.) and store data used for, or generated during, the execution of the software.
In this example, storage system 1020 includes operating system 1022, one or more applications 1024, I/O module 1026, and communication module 1028. Operating system 1022 includes various procedures, sets of instructions, software components and/or drivers for controlling and managing general system tasks (e.g., memory management, storage device control, power management, etc.) and facilitates communication between various hardware and software components. Operating system 1022 may be one of various versions of Microsoft Windows, Apple Mac OS, Apple OS X, Apple macOS, and/or Linux operating systems, a variety of commercially-available UNIX or UNIX-like operating systems (including without limitation the variety of GNU/Linux operating systems, the Google Chrome® OS, and the like) and/or mobile operating systems such as Apple iOS, Windows Phone, Windows Mobile, Android, BlackBerry OS, Blackberry 10, and Palm OS, WebOS operating systems.
Applications 1024 can include any number of different applications installed on computing device 1000. Examples of such applications may include a browser application, an address book application, a contact list application, an email application, an instant messaging application, a word processing application, JAVA-enabled applications, an encryption application, a digital rights management application, a voice recognition application, location determination application, a mapping application, a music player application, etc.
I/O module 1026 manages information received via input components (e.g., display 1010, sensors 1012, and microphone 1016) and information to be outputted via output components (e.g., display 1010 and speaker 1014). Communication module 1028 facilitates communication with other devices via communication system 1018 and includes various software components for handling data received from communication system 1018.
One of ordinary skill in the art will realize that the architecture shown in
As shown, cloud computing system 1112 includes one or more applications 1114, one or more services 1116, and one or more databases 1118. Cloud computing system 1100 may provide applications 1114, services 1116, and databases 1118 to any number of different customers in a self-service, subscription-based, elastically scalable, reliable, highly available, and secure manner.
In some embodiments, cloud computing system 1100 may be adapted to automatically provision, manage, and track a customer's subscriptions to services offered by cloud computing system 1100. Cloud computing system 1100 may provide cloud services via different deployment models. For example, cloud services may be provided under a public cloud model in which cloud computing system 1100 is owned by an organization selling cloud services and the cloud services are made available to the general public or different industry enterprises. As another example, cloud services may be provided under a private cloud model in which cloud computing system 1100 is operated solely for a single organization and may provide cloud services for one or more entities within the organization. The cloud services may also be provided under a community cloud model in which cloud computing system 1100 and the cloud services provided by cloud computing system 1100 are shared by several organizations in a related community. The cloud services may also be provided under a hybrid cloud model, which is a combination of two or more of the aforementioned different models.
In some instances, any one of applications 1114, services 1116, and databases 1118 made available to client devices 1102-1108 via networks 1110 from cloud computing system 1100 is referred to as a “cloud service.” Typically, servers and systems that make up cloud computing system 1100 are different from the on-premises servers and systems of a customer. For example, cloud computing system 1100 may host an application and a user of one of client devices 1102-1108 may order and use the application via networks 1110.
Applications 1114 may include software applications that are configured to execute on cloud computing system 1112 (e.g., a computer system or a virtual machine operating on a computer system) and be accessed, controlled, managed, etc. via client devices 1102-1108. In some embodiments, applications 1114 may include server applications and/or mid-tier applications (e.g., HTTP (hypertext transport protocol) server applications, FTP (file transfer protocol) server applications, CGI (common gateway interface) server applications, JAVA server applications, etc.). Services 1116 are software components, modules, application, etc. that are configured to execute on cloud computing system 1112 and provide functionalities to client devices 1102-1108 via networks 1110. Services 1116 may be web-based services or on-demand cloud services.
Databases 1118 are configured to store and/or manage data that is accessed by applications 1114, services 1116, and/or client devices 1102-1108. For instance, storages 140-155 may be stored in databases 1118. Databases 1118 may reside on a non-transitory storage medium local to (and/or resident in) cloud computing system 1112, in a storage-area network (SAN), on a non-transitory storage medium local located remotely from cloud computing system 1112. In some embodiments, databases 1118 may include relational databases that are managed by a relational database management system (RDBMS). Databases 1118 may be a column-oriented databases, row-oriented databases, or a combination thereof. In some embodiments, some or all of databases 1118 are in-memory databases. That is, in some such embodiments, data for databases 1118 are stored and managed in memory (e.g., random access memory (RAM)).
Client devices 1102-1108 are configured to execute and operate a client application (e.g., a web browser, a proprietary client application, etc.) that communicates with applications 1114, services 1116, and/or databases 1118 via networks 1110. This way, client devices 1102-1108 may access the various functionalities provided by applications 1114, services 1116, and databases 1118 while applications 1114, services 1116, and databases 1118 are operating (e.g., hosted) on cloud computing system 1100. Client devices 1102-1108 may be computer system 900 or computing device 1000, as described above by reference to
Networks 1110 may be any type of network configured to facilitate data communications among client devices 1102-1108 and cloud computing system 1112 using any of a variety of network protocols. Networks 1110 may be a personal area network (PAN), a local area network (LAN), a storage area network (SAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), a global area network (GAN), an intranet, the Internet, a network of any number of different types of networks, etc.
The above description illustrates various embodiments of the present invention along with examples of how aspects of the present invention may be implemented. The above examples and embodiments should not be deemed to be the only embodiments, and are presented to illustrate the flexibility and advantages of the present invention as defined by the following claims. Based on the above disclosure and the following claims, other arrangements, embodiments, implementations and equivalents will be evident to those skilled in the art and may be employed without departing from the spirit and scope of the invention as defined by the claims.
Number | Date | Country | |
---|---|---|---|
62576555 | Oct 2017 | US |