53 results about "Classification procedure" patented technology

An illustrative storage system operating within an environment bounded by a managed network may perform a sensitivity classification procedure with respect to a data instance stored within the storage system, the sensitivity classification procedure comprising determining a sensitivity classification of the data instance, and generating metadata representative of the sensitivity classification; detecting a request to send the data instance to a remote system by way of the managed network; and sending, based on the request, the data instance and the metadata to a network monitoring system that performs a network traffic monitoring service with respect to outgoing network traffic from the managed network.

The invention designs a software security vulnerability mining method for natural language processing, which can be used for detecting backdoor vulnerabilities in a text classification program trained by a small number of poisoning samples. By means of the method, the backdoor vulnerability in the NLP model can be effectively detected under the condition that a small number of backdoor samples are mixed, the effectiveness and concealment of the generated backdoor samples are improved, and a new challenge and defense thought is provided for the safety of the current NLP model. Experimental results prove that backdoor vulnerabilities can be effectively detected on different pre-training models, and compared with other prior art, the method has remarkable advantages in the aspects of quality and effectiveness of backdoor samples.

A system may include an article stream. Variable-appearance articles within the stream may have various orientation states. In some cases, to support later processing, articles may be transitioned from anomalous orientations into one or more nominal orientations which are later processing ready. In some cases, determining the current orientation of the variable-appearance articles may be performed using an adaptive classification routine.

A first learning unit performs learning of a first neural network that extracts a feature vector in each pixel of a target image including a plurality of objects and that outputs a feature map in which feature vectors of pixels belonging to individual objects included in the target image are clustered and distributed as a plurality of the feature vector groups in a feature space which is a space of the feature vector. A second learning unit performs learning of a second neural network that outputs a class classification result of a plurality of objects belonging to the same category included in the target image in response to input of the feature vector of the target image.

A feature extraction unit (510) outputs first and second feature vectors of an input image. An averaged first / second feature calculation unit (520a, 520b) calculates an averaged first / second feature vector by averaging first / second feature vectors of a given class and obtains an averaged first / second feature matrix by aggregating averaged first / second feature vectors of all classes. A first / second feature similarity calculation unit (532a, 532b) calculates a first / second similarity from the first / second feature vector of the input image and a first / second weight matrix. The averaged first / second feature calculation unit (520a, 520b) replaces the first / second weight matrix of the first / second feature similarity calculation unit (532a, 532b) with the averaged first / second feature matrix.

To detect waste plastic.SOLUTION: An information processing device (1) includes: a similarity calculation part (107) for calculating a similarity between a histogram of a brightness value in a target area of a difference image generated from images obtained by capturing the inside a waste pit at a predetermined time interval and a histogram of a brightness value generated from images in which a bagged waste is captured; and a classification part (108) for classifying the waste in the image in the target area based on the calculated similarity.SELECTED DRAWING: Figure 1

To provide a data classification device capable of preventing deterioration of classification accuracy of an object.SOLUTION: An object classification device includes a data acquisition unit (101) that acquires sensor data from a sensor (20), a data processing unit (104) that generates classification data from the sensor data acquired by the data acquisition unit (101) using representative shape data that is sensor data indicating a feature of a representative shape of an object belonging to a class to be classified, and a classification unit (105) that classifies the object on the basis of the classification data generated by the data processing unit (104).SELECTED DRAWING: Figure 3

Turning assistance device (2) as a retrofit solution for a vehicle (1) for retrofitting the vehicle (1) with the turning assistance device (2), comprising a radar sensor device (3) and a control device (4), wherein the radar sensor device (3) is configured to - to detect an object (10) and its relative position (11) with respect to the vehicle (1) in a predetermined spatial detection area (8) in an external environment of the vehicle (1) according to a predetermined detection method, characterized in that the control device (4) is configured to - to classify the object (10) as a road user (6) using a predetermined classification procedure, and - to issue a warning signal (12) when the object (10) classified as a road user (6) is located in a predetermined spatial danger zone (7) related to the vehicle (1), wherein a warning is only issued if the object (10) in the danger zone (7) is the road user (6), and that the control device (4) is configured to - to determine a relative motion (16) of the object (10) with respect to the vehicle (1) according to a predetermined motion determination procedure, - to classify the object (10) as either a stationary object (17) or a dynamic object (18) using a predetermined determination procedure, and - to determine a movement (19) of the vehicle (1) from the relative movement (16) of at least one object (10) classified as a stationary object (17), so that the turning assistance device (2) can independently determine the movement (19) of the vehicle (1).

The present application provides a method for road classification, a device, an apparatus, a system, a medium, and a product thereof, all of which can be used in the field of vehicle control. The method captures a set of chassis signals from a vehicle and then inputs them into a road detection model for classification to determine a road type. A corresponding driving parameter of the vehicle is then determined, according to which the vehicle operates. The present embodiment effectively improves the adaptability of the road type to the degree of influence of the road on the vehicle by performing the classification using the chassis signals related to vehicle operation and the road detection model resulting from a subsequent reclassification of the road.

The present disclosure includes a method of operating an exoskeleton system. The method includes: determining a first state estimate of a current classification procedure implemented by the exoskeleton system; determining a second state estimate of a reference classification procedure; determining that a difference between the first state estimate and the second state estimate is greater than a classification procedure replacement threshold; generating an updated classification procedure; and replacing the current classification procedure with the updated classification procedure based at least in part on determining that the difference between the first state estimate and the second state estimate is greater than the classification procedure replacement threshold.

The described technology includes a method and a system for audio-based measurement of the condition and subsequent classification of arteries, as well as a means of displaying the arterial condition for monitoring progress. The technology uses external sensors, specifically MEMS-based microphones, which are placed on the skin surface near the stent. These sensors detect the sounds of blood flow, including those caused by constrictions and turbulence. The captured sounds are filtered and pre-processed by signal processing algorithms. Subsequently, classification programs based on artificial intelligence are used to characterize the condition of the artery. The innovation of this technology lies in its ability to monitor the condition of arteries by analyzing flow sounds. This provides a simple, cost-effective, and patient-friendly method for detecting arterial narrowing and changes in blood flow, particularly in the vicinity of stents. The technology also offers the capability to generate alerts in critical situations and, if necessary, transmit them to clinical partners.

Improve the accuracy of subject classification by using information highly relevant to the subject in the image. [Solution] The associative label extraction unit 10 extracts multiple associative labels associated with the true label of the main subject of the image from the associative knowledge graph. The embedding representation vector calculation unit 20 calculates the embedding representation vectors of the multiple associative labels and calculates the embedding representation vectors of multiple surrounding labels obtained from the image for the true label of the main subject and averages them. The similarity calculation unit 30 calculates the cosine similarity between the embedding representation vector of each associative label and the averaged embedding representation vector of the surrounding labels. The associative label selection unit 40 selects the embedding representation vector of the associative label that has a high similarity to the averaged embedding representation vector of the surrounding labels. The surrounding label expansion unit 50 uses the selected associative labels as surrounding labels to generate a cell graph with the main subject as the center node and the surrounding labels as side nodes.

In order to accurately and reliably classify a document to be classified, a document classification device (1) is provided with: a hypothetical sentence generation unit (11) that generates a plurality of hypothetical sentences, which are sentences related to a candidate classification into which a document is to be classified; an implication evaluation unit (12) that performs a process for evaluating, for each of the plurality of hypothetical sentences, whether the document implies the hypothetical sentence; and an evaluation result aggregation unit (13) that determines the classification into which the document is to be classified, on the basis of the result of the evaluation of each of the plurality of hypothetical sentences by the implication evaluation unit (12).

METHOD FOR CLASSIFYING FORECASTED MATERIAL WASTE IN AN INDUSTRIAL MANUFACTURING PROCESS, ASSOCIATED SYSTEM Method for handling material waste in an industrial manufacturing process comprising: Acquisition (ACQ1) of a first plan (PL1); Acquisition (ACQ2) of a first nesting plan (PIMA); Characterization (CAR1) of at least one forecasted waste item (CHk) defining at least one area of ​​the nesting plan (PIMA); Querying a first item database (BD1) containing a set of parts (Pn) to be produced; First comparison (COMP1) between the dimensions (Kc) of at least one forecasted waste item (CHk) and the dimensions of the parts to be produced (Pn) extracted from the first item database (BD1); Determination of a list (LIST1) of compatible parts to be produced; Classification (CLASS1) of each forecasted waste item (CHk);Updating a database of falls (BD2). Figure for the abbreviation: Fig.1

To enable a reader of a comment added to a code to efficiently grasp knowledge related to the code according to the purpose of the reader of the comment from the comment.SOLUTION: A comment classification system 100 detects at least one piece of information out of comment common information 121, additional comment difference information 122 and generated comment difference information 123 based on additional comment information 112 added to code information 111 and generated comment information 113 generated from code information 111. The comment classification system 100 controls at least one piece of the information detected in the comment common information 121, the additional comment difference information 122, and the generated comment difference information 123 to display or output a comment for the code information 111 in a manner distinguished from other pieces of the information.SELECTED DRAWING: Figure 1

Fault classification device (1) for detecting a location that is the cause of a fault in a mechanism unit of a machine with multiple components, comprising: a fault unit detection unit (12) that captures a data set in which a fault unit of a machine is correlated with one or more components, wherein the fault unit is a group of one or more components whose causal fault locations cannot be separated from each other when designing the machine; a fault history recording unit (11) that records a fault history containing an event of faults that have occurred in the past and countermeasures for eliminating the faults; and a correlation unit (13) which stores information about one or more events in correlation with the fault unit in a fault history database (21) by comparing the component with the countermeasure component; a recovery receiving unit (15) that receives a recovery request containing a search keyword; and a recovery execution unit (16) that recovers information about the event based on the search keyword and outputs the fault unit correlated with the information about the event in order to specify the location in the mechanism unit of the machine that caused the fault.

Systems, computer program products, and methods are described herein for process optimization using advanced computational models for data analysis and automated processing. The present disclosure is configured to collect metadata from a node associated with data, wherein the node is configured to process the data, and wherein the metadata comprises real-time parameters of the data; train an instantaneous process identifier (IPI) using the metadata, wherein the IPI comprises a deep learning neural network; analyze the metadata using a classification procedure, wherein the classification procedure determines resources needed to process the data; determine a processing node to process the data, wherein determining the processing node is based on the processing node's availability and the processing node's processing capabilities; and allocate resources to process the data, wherein allocating the resources comprises an instance-based allocation determined by the IPI.

Methods and devices for a condition classification of a power network asset of a power network asset are disclosed. The methods and devices may combine an automatic classification procedure with a missing data replacement procedure.

At least one or more cameras configured to capture images of at least one capture state and one assembly state of a component and a component feed position, a defect detection section configured to control the camera to obtain an image of at least one capture state and one assembly state during the manufacture of a circuit board, and performing at least one defect detection processing based on an image in the capture state and one defect detection processing based on an image in the assembly state, an inspection section and a classification section are provided.The inspection section checks for the presence or absence of the component at the component feed position by applying a pre-capture image to a trained model. This model was developed using machine learning with multiple pre-capture images as input data and the presence or absence of the component at the feed position as training data. If no defect is detected, the classification section categorizes the pre-capture image into a component-specific image to be used as training data in the machine learning process.

To provide an image data classification apparatus configured to cause a storage unit to keep holding face characteristic data of members of a specific group.SOLUTION: An image data classification apparatus includes: a storage unit which stores multiple pieces of face characteristic data classified into multiple clusters; and a control unit which calculates, for each of the multiple pieces of face characteristic data, a product of weight of one of the multiple pieces of face characteristic data and the number of pieces of face characteristic data of one cluster to which the one piece of face characteristic face data belongs, out of the multiple clusters. The control unit deletes, based on the product for each of the multiple pieces of face characteristic data, at least one of the multiple pieces of face characteristic data, from the storage unit.SELECTED DRAWING: Figure 4

The invention relates to a network flow control method and device, computer equipment, a storage medium and a program product. The method comprises the following steps: generating a classification program; the classification program is an eBPF program running in a kernel mode; controlling the network flow through the classification program; wherein the classification program is configured to classify the network traffic, and if the network traffic belongs to a first category, the network traffic is sent to an auxiliary container; and if the network traffic belongs to the second category, sending the network traffic to a service container. By adopting the method, the communication efficiency can be improved.

The invention discloses a food additive production management and control system and method based on intelligent scheduling, and relates to the technical field of additive production. The method comprises a printing step and an extraction step, and the printing step comprises the following steps that a scheduling unit obtains scheduling records of a production workshop and uniformly uploads the scheduling records to a processing unit, the processing unit stores the scheduling records to a database, and meanwhile, a data classification program is executed according to the data length L of the scheduling records to obtain data blocks and a compression algorithm; the method comprises the steps that a data block is stored in a database, a processing unit obtains a packaging picture from the database, the processing unit executes a picture fusion program according to the data block and a compression algorithm to obtain a fused picture, and the processing unit synchronously stores the fused picture to the database. The distributed storage of the scheduling record can be realized, the data security is improved, the scheduling record is prevented from being manually tampered, and the authenticity of the scheduling record is improved.

Sets of vectors are generated based on respective output signals from at least M+N detectors detecting different respective output modes emitted from one or more linear-optical interferometers (LOIs). A first set of vectors corresponds to an M-mode quantum state emitted from at least one of the LOIs. A second set of vectors corresponds to an N-mode quantum state emitted from at least one of the LOIs. A classifier is trained to classify each of a plurality of pairs of X-dimensional input vectors as a member of one of two or more different classes based on an X-1 dimensional hyperplane determined at least in part by the training, which comprises: receiving training data comprising at least one vector from the first set and at least one vector from the second set, and updating the X-1 dimensional hyperplane based at least in part on the training data (X ≥ M; X ≥N).

An X-ray fitting analysis method and an X-ray fitting analysis system based on a multi-physics variable model. The X-ray fitting analysis method includes: obtaining multiple X-ray measurement signals of an inspection target by an X-ray measurement apparatus; establishing multiple fitting models according to a target architecture; performing an initial fitting procedure to generate multiple initial fitting results and multiple initial parameter ranges; performing a first co-fitting procedure to obtain multiple first fitting results that satisfy a first fitting condition; performing, according to a data type of the first fitting results, a classification procedure to generate multiple classification fitting results; counting to obtain multiple classification parameter ranges; and performing a second co-fitting procedure to obtain multiple second fitting results that satisfy a second fitting condition and respectively correspond to multiple structural parameters.

To more efficiently acquire a work trajectory on which a work vehicle performs work in an operation trajectory on which the work vehicle moves.SOLUTION: An operation trajectory classification method includes: preparing positioning information at a plurality of positioning points (70) obtained by measuring the position of a work vehicle (2) performing prescribed work while moving inside a farm field (9) at a plurality of times (S2); generating operation trajectory information representing an operation trajectory (80) on which the work vehicle moves on the basis of the positioning information (S3); calculating a prescribed parameter representing a state of the work vehicle at the plurality of respective positioning points on the basis of the positioning information (S4); leaving a harvesting work position to move to a discharging position to allow the work vehicle (2) to discharge harvested crops, and classifying a leaving and merging trajectory (85) from the discharging position to a harvesting work position in the operation trajectory (80) on the basis of the parameter (S5); and outputting operation trajectory information so as to make the leaving and merging trajectory (85) identifiable (S6).SELECTED DRAWING: Figure 3

A classification device includes processing circuitry configured to acquire an operation log related to operation information and identify each operation performed by a user using the operation log, create a vector of each operation based on a co-occurrence relationship between operations identified, calculate a similarity between a predetermined number of operations adjacent to each other in chronological order by using the vector of each operation created, determine a division point of operations by using the similarity calculated and divide a time-series operation into operation sets based on the division point, and classify the operation sets divided into classes based on a number of types of operations common to the operation sets.

To automatically generate a label corresponding to each cluster obtained by clustering each sentence of a model document, and to automatically create sentence classification information which is teacher data.SOLUTION: The cluster configuration unit 130 clusters a plurality of sentence units 511 obtained as a result of decomposing the model document 52 into a plurality of clusters. The label setting unit 140 acquires a label setting result in which a label is set to each cluster based on the clustering result. The model learning unit 150 acquires the text classification information 53 in which the text units included in each of the plurality of clusters are associated with the labels based on the label setting result. Here, the text classification information 53 is training data used for learning of the classification model 60 that classifies text units included in the classification target document.SELECTED DRAWING: Figure 1

Provided is a test case classification apparatus including a memory and a processor. The processor includes a test case classification section that performs classification of one test case that is included in test cases describing test contents for system software, by comparing a manual testing execution time representing an amount of time required to manually execute a test based on the one test case and an automated testing execution time representing an amount of time required to automatically execute the test and setting a mode of running the one test case, and a result output section that outputs a result of the classification performed by the test case classification section.