879 results about "Learning abilities" patented technology

The current invention is directed to a system for adaptive medical decision support. The invented system provides a system that allows users to efficiently enter, access, and analyze medical information, without disrupting patient-doctor interactions or medical facility course of business; which assists in all stages of medical assessment and treatment; and which is tailored to the particular medical practice or specialty and taking into account the developing habits, preferences, performance, and individual patient histories, of an individual user. The invention provides a learning capacity configured to learn previously presented data and decisions and predict data or decisions based on data that it receives from the user, thereby adapting its operations to the developing habits, preferences, performance, and individual patient histories of an individual user. The system may also provide a “virtual specialist” feature, whereby the system can be instructed to produce the probable actions or recommendations of particular medical specialists.

Systems and methods disclosed herein relates to building an intelligent assistant that can take in human requests / commands in simple text form, especially in natural language format, and perform tasks for users. Systems and methods are disclosed herein in which knowledge of how to interpret users' requests and carry out tasks, including how to find and manipulate information on the Internet, can be learned from users by the designed assistant, and the knowledge can be subsequently used by the assistant to perform tasks for users. Using the disclosed methods, the designed assistant enables a user to teach the assistant, by actually performing the task manually through the provided user interface, and / or by referring to some knowledge that the assistant already knows; the designed assistant may generate more generic knowledge based on what it learns, and can apply the more generic knowledge to serve requests that it has never seen and never directly learned, and can revise / improve the knowledge according to execution result / feedback. The methods and systems being disclosed here are useful for building an intelligent assistant, especially a universal personal assistant and an intelligent search assistant.

The present invention relates to the field of deep learning techniques, in particular to a deep convolution neural network training method and a device. The deep convolution neural network training method and the device comprise the steps of a, pretraining the DCNN on a large scale data set, and pruning the DCNN; b, performing the migration learning on the pruned DCNN; c, performing the model compression and the pruning on the migrated DCNN with the small-scale target data set, In the process of migrating learning of large-scale source data set to small-scale target data set, the model compression and the pruning are performed on the DCNN by the migration learning method and the advantages of model compression technology, so as to improve the migration learning ability to reduce the risk of overfitting and the deployment difficulty on the small-scale target data set and improve the prediction ability of the model on the target data set.

The invention relates to an image retrieval method based on deep learning and Hash. According to the image retrieval method, on the basis of powerful learning capacity of a deep convolutional neural network, deep features of images are extracted, and the problems of weak feature expression capacity and low retrieval precision caused by use of lower features of the images in the prior art are solved; a Hash layer is introduced for construction of a Hash function, learning of the deep features of the images and the construction of the Hash function are completed in the same process, an internal relation of the image features and the Hash function is explored, and the accuracy rate of the image retrieval is greatly increased; quantization error loss is added to a loss layer of the deep convolutional neural network, the expression capacity of Hash codes is enhanced, by means of a Softmax classifier loss module and a quantization error loss module, quantization errors caused by binaryzation in the Hash function are effectively reduced, and the accuracy rate of the image retrieval is further increased.

The invention is directed to a method for training at least one learning-capable system comprising the steps of providing a predetermined training data set corresponding to a predetermined number of subjects comprising a predetermined input data set and a predetermined outcome data set, augmenting the input data set and / or the outcome data set, and training each learning-capable system using the augmented input data set and / or the augmented outcome data set.

A graphical user interface is configured to display information related to diagnosis and / or repair of various objects, such as automobiles. The GUI displays relevant reference information, of any type, next to procedure steps of a selected diagnostic or repair procedure. Advantageously, reference information from multiple sources may be automatically selected for display in the GUI according to a current task that is selected by the user. Accordingly, a technician viewing the GUI does not need to access any other reference information, valuable man hours are spared, and efficiency increases. In one embodiment, the layout and immediate availability of relevant reference information may also increase the learning capabilities of the viewing technician. Thus, the layout of information in the GUI as described herein may facilitate increased employee productivity and increased employee learning.

The invention discloses a reinforcing learning initial method of a mobile robot based on an artificial potential field and relates to a path planning Q-learning initial method of the mobile robot. The working environment of the robot is virtualized to an artificial potential field. The potential values of all the states are confirmed by utilizing priori knowledge, so that the potential value of an obstacle area is zero, and a target point has the biggest potential value of the whole field; and at the moment, the potential value of each state of the artificial potential field stands for the biggest cumulative return obtained by following the best strategy of the corresponding state. Then a Q initial value is defined to the sum of the instant return of the current state and the maximum equivalent cumulative return of the following state. Known environmental information is mapped to a Q function initial value by the artificial potential field so as to integrate the priori knowledge into a learning system of the robot, so that the learning ability of the robot is improved in the reinforcing learning initial stage. Compared with the traditional Q-learning algorithm, the reinforcing learning initial method can efficiently improve the learning efficiency in the initial stage and speed up the algorithm convergence speed, and the algorithm convergence process is more stable.

The invention belongs to the field of image processing, aims at restoring a high-resolution depth image and utilizing the great learning capacity of convolutional neural networks to solve the defects that the conventional algorithm is high in computational complexity and high in actual application cost and cannot effectively extract features, and provides the technical scheme of a depth map super-resolution reconstruction method based on the convolutional neural networks. The convolutional neural networks (CNN) combining a convolutional layer and a deconvolutional layer is utilized to extract the depth image features of low-resolution sample depth image block and a high-resolution sample depth image block, and then the nonlinear mapping relation between the depth image features is learnt so as to restore the high-resolution depth image. The depth map super-resolution reconstruction method based on the convolutional neural networks is mainly applied to the occasion of image processing.

A graphical user interface is configured to display information related to diagnosis and / or repair of various objects, such as automobiles. The GUI displays relevant reference information, of any type, next to procedure steps of a selected diagnostic or repair procedure. Advantageously, reference information from multiple sources may be automatically selected for display in the GUI according to a current task that is selected by the user. Accordingly, a technician viewing the GUI does not need to access any other reference information, valuable man hours are spared, and efficiency increases. In one embodiment, the layout and immediate availability of relevant reference information may also increase the learning capabilities of the viewing technician. Thus, the layout of information in the GUI as described herein may facilitate increased employee productivity and increased employee learning.

The invention provides a robot reinforced learning initialization method based on a neural network. The neural network has the same topological structure as a robot working space, and each neuron corresponds to a discrete state of a state space. The method comprises the following steps of: evolving the neural network according to the known partial environmental information till reaching a balance state, wherein at the moment, the output value of each neuron represents maximum cumulative return acquired when the corresponding state follows the optimal strategy; defining the initial value of a Q function as the sum of the immediate return of the current state and the maximum converted cumulative return acquired when the subsequent state follows the optimal strategy; and the mapping the known environmental information into the initial value of the Q function by the neural network. Therefore, the prior knowledge is fused into a robot learning system, and the learning capacity of the robot at the initial stage of reinforced learning is improved; and compared with the conventional Q learning algorithm, the method has the advantages of effectively improving the learning efficiency of the initial stage and increasing the algorithm convergence speed.

The invention provides a short-term traffic flow prediction method based on a convolutional neural network. The short-term traffic flow prediction method comprises the steps that firstly, the formats of input matrixes are determined according to the number of upstream and downstream road sections and the number of historical flow data predicted to be used; secondly, a structure of a convolutional neural network prediction model is determined according to the formats of input matrixes, and model training is completed by using the historical flow data of predicted road sections and the upstream and downstream road sections of the predicted road sections; finally, prediction is performed by using the trained model. The method utilizes the convolutional neural network having powerful characteristic learning capability to accurately predict short-term traffic flow, considers the flows of the predicted road sections and the upstream and downstream road sections of the predicted road sections simultaneously, and enables input data to be expanded to two dimensions so as to conform to the input format of the convolutional neural network. In addition, information of the road sections relevant with the predicted road sections is also provided to enable the prediction model to learn more flow characteristics, and accordingly the prediction accuracy is improved.

The invention discloses an ReLU convolutional neutral network-based image denoising method. The method comprises the following steps of building an ReLU convolutional neutral network model, wherein the ReLU convolutional neutral network model comprises a plurality of convolutional layers and active layers after the convolutional layers, wherein the active layers are ReLU functions; selecting a training set, and setting training parameters of the ReLU convolutional neutral network model; training the ReLU convolutional neutral network model by taking a minimized loss function as a target according to the ReLU convolutional neutral network model and the training parameters of the ReLU convolutional neutral network model to form an image denoising neural network model; and inputting a to-be-processed image to the image denoising neural network model, and outputting a denoised image. According to the ReLU convolutional neutral network-based image denoising method disclosed by the invention, the learning ability of the neural network is greatly enhanced, accurate mapping from noisy images to clean images is established, and real-time denoising can be realized.

The invention discloses an intensive learning based urban intersection passing method for a driverless vehicle. The method includes a step 1 of collecting vehicle continuous running state informationand position information through a photographing method, the vehicle continuous running state information and position information including speed, lateral speed and acceleration value, longitudinal speed and acceleration value, traveling track curvature value, accelerator opening degree and brake pedal pressure; a second step of obtaining characteristic motion track and the velocity quantity of actual data through clustering; a step 3 of processing original data by an exponential weighting moving average method; a step 4 of realizing the interaction passing method by utilizing an NQL algorithm. The NQL algorithm of the invention is obviously superior to a Q learning algorithm in learning ability when handling complex intersection scenes and a better training effect can be achieved in shorter training time with less training data.

The invention provides a ship intelligent identification tracking method. For the ship intelligent identification tracking method, based on a deep learning algorithm based on computer vision, a basicclassification network structure and a multi-scale target prediction method in traditional deep learning are improved, and a Darknet network and a YOLOv3 algorithm are combined to track a ship and detect and identify the type of the ship in real time. According to the ship intelligent identification tracking method, the idea of a residual error network is introduced; a full convolution structure is adopted; the network depth is increased; the data feature learning capability is improved; and local feature interaction between the feature maps is achieved in a convolution kernel mode through a YOLOv3 algorithm, and matching and positioning of targets are conducted, and target area prediction and category prediction are integrated into a single neural network model on the basis, so that global information of the images is recognized as the targets. Experimental results show that compared with a traditional method, the algorithm provided by the invention not only has better real-time performance and accuracy, but also has better robustness for various environmental changes.

The invention discloses an aspect-level emotion classification model and method based on dual-memory attention, belonging to the technical field of text emotion classification. The model of the invention mainly comprises three modules: an encoder composed of a standard GRU loop neural network, a GRU loop neural network decoder introducing a feedforward neural network attention layer and a Softmaxclassifier. The model treats input statements as a sequence, based on the attention paid to the position of the aspect-level words in the sentence, Two memory modules are constructed from the originaltext sequence and the hidden layer state of the encoder respectively. The randomly initialized attention distribution is fine-tuned through the attention layer of the feedforward neural network to capture the important emotional features in the sentences, and the encoder-decoder classification model is established based on the learning ability of the GRU loop neural network to the sequence to achieve aspect-level affective classification capabilities. The invention can remarkably improve the robustness of the text emotion classification and improve the classification accuracy.

The invention discloses a pile-up noise reduction own coding network bearing fault diagnosis method based on particle swarm optimization. The bearing fault diagnosis method provides an improved pile-up noise reduction own coding network SADE bearing fault diagnosis method, SDAE network hyper-parameters, such as cyber hidden layer nodes, sparse parameters, input data random zero setting ratio, are selected adaptively by particle swarm optimization PSO, a SADE network structure is determined, top character representation of malfunction inputting a soft-max classifier is obtained and a classification of defects is discerned. The bearing fault diagnosis method has better feature in learning capacity and more robustness than feature of learning of ordinary sparse own coding device, and builds a SDAE diagnostic model having multi-hidden layer by optimizing the hyper-parameters of noise reduction own coding network deepness network structure with the particle swarm optimization, accuracy of the classification of defects is improved at last.

The invention provides a test system of a stimulus information cognition ability value, comprising a stimulus information providing device, a sight tracking device, a feedback data acquisition device, a cognition index value analysis device, a cognition accuracy degree analysis device and a cognition ability value computing device. The stimulus information providing device shows vision and audition stimulus information for a tester; the sight tracking device tracks and records the motion state of eyes of a tester in the process of acquiring and processing the vision stimulus information; the feedback data acquisition device is used for acquiring feedback data input by the tester responding to the stimulus information; the cognition index value analysis device acquires a cognition index value of the tester in the process of processing the vision stimulus information according to the record of the sight tracking device through a visual motion analysis method; the cognition accuracy degree analysis device computes the cognition accuracy degree of the tester according to the content of the feedback data; and the cognition ability value computing device computes obtained standard scores for representing the comprehensive cognition ability of the tester according to the cognition index value and the cognition accuracy degree through a statistic method. The test system can more accurately record the cognition process of the tester and the difference of the cognition ability among testers and can also be widely applied in the fields of talent selection, job placement, learning ability diagnosis, advertising effect, safety training of drivers, and the like.

The invention discloses a nasopharyngeal-carcinoma (NPC) lesion automatic-segmentation method and nasopharyngeal-carcinoma lesion automatic-segmentation systems based on deep learning. The method comprises: carrying out registration on a PET (Positron Emission Tomography) image and a CT (Computed Tomography) image of nasopharyngeal carcinoma to obtain a PET image and a CT image after registration;and inputting the PET image and the CT image after registration into a convolutional neural network to carry out feature representation and scores map reconstruction to obtain a nasopharyngeal-carcinoma lesion segmentation result graph. The method carries out registration on the PET image and the CT image of the nasopharyngeal carcinoma, obtains a nasopharyngeal-carcinoma lesion by automatic segmentation through the convolutional neural network, and is more objective and accurate as compared with manual segmentation manners of doctors; and the convolutional neural network in deep learning isadopted, consistency is better, feature learning ability is higher, the problems of dimension disasters, easy falling into a local optimum and the like are solved, lesion segmentation can be carried out on multi-modal images of the PET-CT images, and an application range is wider. The method can be widely applied to the field of medical image processing.

An apparatus having a memory and circuit is disclosed. The memory may (i) assert a first signal in response to detecting a conflict between at least two addresses requesting access to a block at a first time, (ii) generate a second signal in response to a cache miss caused by an address requesting access to the block at a second time and (iii) store a line fetched in response to the cache miss in another block by adjusting the first address by an offset. The second time is generally after the first time. The circuit may (i) generate the offset in response to the assertion of the first signal and (ii) present the offset in a third signal to the memory in response to the assertion of the second signal corresponding to reception of the first address at the second time. The offset is generally associated with the first address.

A system and method for the analysis and visualization of normal and abnormal tissues, objects and structures in digital images generated by medical image sources is provided. The present invention utilizes principles of Iterative Transformational Divergence in which objects in images, when subjected to special transformations, will exhibit radically different responses based on the physical, chemical, or numerical properties of the object or its representation (such as images), combined with machine learning capabilities. Using the system and methods of the present invention, certain objects, such as cancerous growths, that appear indistinguishable from other objects to the eye or computer recognition systems, or are otherwise almost identical, generate radically different and statistically significant differences in the image describers (metrics) that can be easily measured.

The invention discloses a personalized teaching resource recommendation system based on a knowledge map and ability evaluation. The personalized teaching resource recommendation system comprises a knowledge map resource module, a student learning ability evaluation module, a student portrait module, a teacher portrait module and a personalized recommendation module, wherein the knowledge map resource module is used for constructing a knowledge map and a resource map of a resource library; the student learning ability evaluation module is used for evaluating learning ability of a student on the basis of the knowledge map resource module to obtain a student learning ability level; the student portrait module is used for drawing a student portrait by combining the student learning ability level with a student information library, and clustering student information; the teacher portrait module is used for drawing a teacher portrait based on a teacher information library; and the personalized recommendation module is used for recommending resources to the student or the teacher. The personalized teaching resource recommendation system is based on heterogeneous teaching resources, and can recommend teaching resources to learners with different learning abilities and teachers teaching different learners with relatively high precision.

A method and apparatus for detecting abnormal behavior of enterprise software applications is disclosed. A profile that represents the behavior of the function is created for each service and error function integrated in an enterprise software application. This profile is based on input measurements, such as response time, throughput, and non-availability. For each such input measurement, the expected behavior is determined, as well as the upper and lower bounds on that expected behavior. The invention further monitors the behavior of service and error functions and produces an exception if at least one of the upper or lower bounds is violated. The detection scheme disclosed is dynamic, adaptive, and has self-learning capabilities.

The present invention relates to a short-term load prediction method based on a particle swarm optimization least squares support vector machine. Aiming at the deficiency of a single kernel function least squares support vector machine model, the Gaussian kernel function and the Polynomial kernel function are combined to obtain a new hybrid kernel function so as to improve the learning ability and the generalization ability of the least squares support vector machine model; the particle swarm optimization algorithm based on double populations is employed to optimize parameters of the least squares support vector machine of the hybrid kernel function, the particle swarm optimization algorithm based on double populations has advantages of good global search and local search performances, and a strategy having dynamic accelerated factors is employed so as to greatly increase the variety of particles and prevent the search from being caught in a local extremum. The short-term load prediction method based on the particle swarm optimization least squares support vector machine maximally utilizes information in computation, and in the process of selecting the optimal parameter value, the average mean square error of load data and actual data is employed as the adaptation value of the particle swarm optimization algorithm so as to improve the short-item load prediction accuracy value.

The invention discloses an image de-noising method based on a convolution pair neural network, comprising the following steps: building a convolution pair neural network model, wherein the convolution pair neural network model includes multiple convolution pairs and corresponding activation layers; selecting a training set, and setting the training parameters of the convolution pair neural network model; according to the convolution pair neural network model and the training parameters thereof, training the convolution pair neural network model with the goal of loss function minimizing to form an image de-noising neural network model; and inputting a to-be-processed image to the image de-noising neural network model, and outputting a de-noised image. Through the image de-noising method based on a convolution pair neural network disclosed by the invention, the learning ability of the neural network is enhanced greatly, accurate mapping from noisy images to clean images is established, and real-time de-noising is realized.

The invention relates to infant formula milk powder and a production method thereof. The infant formula milk powder contains nutrition constituents such as protein, fat, carbohydrate, vitamins, mineral elements and the like with content regulated by the national standards on infant foods. Moreover, whey peptides, collagen peptides and casein phosphopeptides are added in the milk powder. Since the three kinds of binding peptides are added, the invention has the advantages that the immunity of cesarean infants is improved and the disease resistance of the infants is enhanced; the digestive function of the infants is improved, the intestinal mucosa proliferation is promoted and the digestive function is improved; and the development of the nervous system is promoted, the nutrition supply to the nervous system is guaranteed and the memory and the learning ability of the infants are improved.

The invention discloses a gait recognition method based on deep learning. The gait recognition method based on deep learning is characterized that identity of a person in a video is recognized according to gaits of the person through weight-shared two-channel convolutional neural networks by utilizing strong learning ability of the convolutional neural networks in a deep learning mode. The method has strong robustness on gait changes crossing large view angles, effectively solves the problem that accuracy is low when existing gait recognition technologies are used for recognizing gaits crossing view angles. The method can be widely applied to scenes with video monitoring, such as safety monitoring in airports and supermarkets, personnel recognition and criminal detection.

The invention relates to a cell image segmentation method based on automatic feature learning. As a method for learning features of cell images is very good in feature learning capacity, the cell segmentation accuracy can be greatly improved, and meanwhile, a random forest classifier does not need to select the features, so that the method is capable of well solving the confronted problems of feature extraction and selection in a recognition process. The cell image segmentation method based on the automatic feature learning comprises the following steps: 1, preprocessing: preprocessing initial cell images in a training set and a test set; (2) training a feature extractor; (3) performing recognition by utilizing the random forest classifier; and (4) postprocessing.

The invention discloses linguistic model training method and system based on distributed neural networks. The method comprises the following steps: splitting a large vocabulary into a plurality of small vocabularies; corresponding each small vocabulary to a neural network linguistic model, each neural network linguistic model having the same number of input dimensions and being subjected to the first training independently; merging output vectors of each neural network linguistic model and performing the second training; obtaining a normalized neural network linguistic model. The system comprises an input module, a first training module, a second training model and an output model. According to the method, a plurality of neural networks are applied to training and learning different vocabularies, in this way, learning ability of the neural networks is fully used, learning and training time of the large vocabularies is greatly reduced; besides, outputs of the large vocabularies are normalized to realize normalization and sharing of the plurality of neural networks, so that NNLM can learn information as much as possible, and the accuracy of relevant application services, such as large-scale voice identification and machine translation, is improved.

The embodiment of the invention provides a method for information recommendation based on motion identification. The method comprises following steps: acquiring information associated with motion; identifying a motion type of a user based on information associated with motion; and recommending information to the user based on information associated with motion and the motion type.The invention further provides a device for information recommendation based on motion identification.By adoption of the above mode, the motion type can be identified based on multi-modal motion data of sensors and motion information of communication data of the user.According to the technical scheme, the method and device for information recommendation based on motion identification have following beneficial effects: a personalized identification library with the self-learning capability can be set up in order to further improve accuracy of identification of the motion type; and based on user features of the personalized identification library, targeted diversified information is recommended to the user.

The invention discloses a pH (potential of hydrogen) value predicting method of a BP (back propagation) neutral network based on a simulated annealing (SA) algorithm optimization. The pH value predicting method comprises the following steps: step one, selecting a sample according to a sample selecting strategy and inputting; step two, according to the BP theorem, determining the structure of the BP neutral network; step three, according to a network training strategy, applying the simulated annealing algorithm to optimize the BP network weight parameter; training the BP network by using the input sample, and determining the optimal weight and optimal hidden node number of the BP network; step four, according to the well trained BP neutral network, structuring a predicting model of the pH value. The pH value predicting method overcomes the randomness of the BP network in terms of weight selection, improves the rate of convergence and study ability of the BP neutral network. Besides, the method optimizes the selection of the training sample and the network hidden neutral element number, and improves the generalization ability of the BP neutral network. Moreover, the pH value predicting method is high in predicting accuracy of pH value and good in nonlinear fitting ability.