Abstract: Techniques disclosed herein relate generally to constructing a customized knowledge graph. In one embodiment, entities and relations among entities are extracted from a user dataset based on certain rules to generate a seed graph. Large-scale knowledge graphs are then traversed using a finite state machine to identify candidate entities and/or relations to add to the seed graph. A priority function is used to select entities and/or relations from the candidate entities and/or relations. The selected entities and/or relations are then added to the seed graph to generate the customized knowledge graph.

Abstract: A system for using hash keys to preserve privacy across multiple tasks is disclosed. The system may provide training batch(es) of input observations each having a customer request and stored task to an encoder, and assign a hash key(s) to each of the stored tasks. The system may provide a new batch of input observations with a new customer request and new task to the encoder. The encoder may generate a new hash key assigned to the new customer request and determine whether any existing hash key corresponds with the new hash key. If so, the system may associate the new batch of input observations with the corresponding hash key and update the corresponding hash key such that it is also configured to provide access to the new batch of input observations. If not, the system may generate a new stored task and assign the new hash key to it.

Abstract: An agricultural data collection framework is provided in a system and method for tracking and managing livestock, and for analyzing animal conditions such as health, growth, nutrition, and behavior. The framework uses ultra-high frequency interrogation of RFID tags to collect individual animal data across multiple geographical locations, and incorporates artificial intelligence techniques to develop machine learning base models for statistical process controls around each animal for evaluating the animal condition. The framework provides a determination of normality at an individual animal basis or for a specific location, and generates alerts, predictions, and a targeted processing or application schedule for prioritizing and delivering resources when intervention is needed.

Abstract: Systems, methods, and apparatuses for providing dynamic, prioritized spectrum utilization management. The system includes at least one monitoring sensor, at least one data analysis engine, at least one application, a semantic engine, a programmable rules and policy editor, a tip and cue server, and/or a control panel. The tip and cue server is operable utilize the environmental awareness from the data processed by the at least one data analysis engine in combination with additional information to create actionable data.

Abstract: Disclosed is detecting identification documents in image-borne identification documents and protecting against loss of the image-borne identification documents. A trained deep learning (DL) stack is used to classify production images by inference as containing a sensitive image-borne identification document, with the trained stack configured with parameters determined using labelled ground truth data for the identification documents and examples of other image documents. The trained DL stack is configured to include a first set of layers closer to an input layer and a second set of layers further from the input layer, with the first set pre-trained to perform image recognition before exposing the second set of layers of the stack to the labelled ground truth data for the image-borne identification documents and examples of other image documents, and using the inferred classification of the sensitive image-borne identification document in a DLP system to protect against loss by image exfiltration.

Abstract: A knowledge-based system under uncertainties and/or incompleteness, referred to as augmented knowledge base (AKB) is provided, including constructing, reasoning, analyzing and applying AKBs by creating objects in the form E?A, where A is a rule in a knowledgebase and E is a set of evidences that supports the rule A. A reasoning scheme under uncertainties and/or incompleteness is provided as augmented reasoning (AR).

Abstract: An ensemble of detection techniques are used to identify code that presents intermediate levels of threat. For example, an ensemble of machine learning techniques may be used to evaluate suspiciousness based on binaries, file paths, behaviors, reputations, and so forth, and code may be sorted into safe, unsafe, intermediate, or any similar categories. By filtering and prioritizing intermediate threats with these tools, human threat intervention can advantageously be directed toward code samples and associated contexts most appropriate for non-automated responses.

Abstract: The present disclosure is directed to an automated neural architecture search approach for designing new neural network architectures such as, for example, resource-constrained mobile CNN models. In particular, the present disclosure provides systems and methods to perform neural architecture search using a novel factorized hierarchical search space that permits layer diversity throughout the network, thereby striking the right balance between flexibility and search space size. The resulting neural architectures are able to be run relatively faster and using relatively fewer computing resources (e.g., less processing power, less memory usage, less power consumption, etc.), all while remaining competitive with or even exceeding the performance (e.g., accuracy) of current state-of-the-art mobile-optimized models.

Abstract: Some embodiments provide a method for training a machine-trained (MT) network. The method uses a first set of training inputs to train parameters of the MT network. The method uses a set of validation inputs to measure error for the MT network as trained by the first set of training inputs. The method adds at least a subset of the validation inputs to the first set of training inputs to create a second set of training inputs. The method uses the second set of training inputs to train the parameters of the MT network. The error measurement is used to modify the training with the second set of training inputs.

Abstract: A computerized-method for real-time detection of real concept drift in predictive machine learning models, by processing high-speed streaming data. The computerized-method includes: receiving a real-time data stream having labeled and unlabeled instances. Obtaining a window of ‘n’ instances having a portion of the ‘n’ instances as reliable labels. Computing posterior distribution of the reliable labels; and operating a Drift-Detection (DD) module. The DD module is configured to: operate a kernel density estimation on the computed posterior distribution for sensitivity control of the DD module; operate an error rate function on the estimated kernel density to yield an error value; and train an incremental estimator module, according to the kernel density estimation. When the error value is not above a preconfigured drift threshold repeating operations (i) through (iii), else when the error value is above the preconfigured drift threshold, at least one concept drift related action takes place.

Abstract: Approaches for multitask learning as question answering include a method for training that includes receiving a plurality of training samples including training samples from a plurality of task types, presenting the training samples to a neural model to generate an answer, determining an error between the generated answer and the natural language ground truth answer for each training sample presented, and adjusting parameters of the neural model based on the error. Each of the training samples includes a natural language context, question, and ground truth answer. An order in which the training samples are presented to the neural model includes initially selecting the training samples according to a first training strategy and switching to selecting the training samples according to a second training strategy. In some embodiments the first training strategy is a sequential training strategy and the second training strategy is a joint training strategy.

Abstract: A sensor transformation attention network (STAN) model including sensors, attention modules, a merge module and a task-specific module is provided. The attention modules calculate attention scores of feature vectors corresponding to the input signals collected by the sensors. The merge module calculates attention values of the attention scores, and generates a merged transformation vector based on the attention values and the feature vectors. The task-specific module classifies the merged transformation vector.

Abstract: The present disclosure relates to a data processing method, a medical term processing system, a medical diagnostic system, a medical intelligent triage apparatus and a computer readable storage medium. The method comprises: acquiring statement information inputted by a user; dividing the statement information into a plurality of information segments by means of a predetermined algorithm, the plurality of information segments comprising a first information segment; establishing a Gaussian distribution of each of the information segments; calculating the similarity between the first information segment and words in a database by means of a similarity model; and obtaining, from words in the database, at least one second information segment for describing the first information segment based on the calculated similarity. The present disclosure can provide professional terms, such as medical terms, corresponding to colloquial expressions, to facilitate users' diagnosis and treatment.

Abstract: An information processing apparatus includes a processor that acquires measurement information measured by a device put on a task performer; detects a predetermined operation in the measurement information; extracts feature vectors from an operation section in which the predetermined operation has been detected; divides the feature vectors into clusters by division methods; obtains ratios at which the feature vectors are classified into the respective clusters in a predetermined time; performs learning by using the ratios as an input and using, as an output label, information indicating whether the task performer performs a predetermined task in the predetermined time; generates a task estimation model by determining weighting for the clusters based on a result of the learning; and estimates, by using the task estimation model, whether the task performer is in a process of performing the predetermined task.

Abstract: The present disclosure is directed to a novel system for generating expansion artificial intelligence (AI) based decision making engines to improve overall AI solution fairness, reduce bias and reduce discrimination in outcomes. Embodiments perform a restriction stage that includes generating an initial AI solution, perform an expansion stage by generating a plurality of artificial intelligence expansion engines by modifying the starting state to determine a new starting state for each; modifying the set of criteria to determine a new set of criteria, using the new starting state and the new set of criteria to generate an expansion AI solution; and joining the initial AI solution of the restriction engine with the plurality of expansion AI solutions to generate an ensemble AI solution. This solution may be refined by a restriction stage and the expansion and/or restriction stages may be reiterated as desired.

Abstract: An identification of an item that was misclassified by a classification model constructed in accordance with a machine learning technique is received. One example of such a machine learning technique is a random forest. A subset of training data, previously used to construct the model, and that is associated with the misclassified item is identified. At least a portion of the identified subset is provided as output.

Abstract: A risk evaluation method is disclosed. A machine learning is conducted by using a neural network by inputting training data. A data distance corresponding to a permission level is calculated based on restoration data and the training data, the permission level being among a plurality of layers of the neural network, the restoration data being generated by using at least one weight of a plurality of permission level weights at the permission level, the plurality of permission level weights being among a plurality of weights of synapses at the plurality of layers, the plurality of weights being generated by the machine learning.

Abstract: Systems, methods, and apparatuses for providing dynamic, prioritized spectrum utilization management. The system includes at least one monitoring sensor, at least one data analysis engine, at least one application, a semantic engine, a programmable rules and policy editor, a tip and cue server, and/or a control panel. The tip and cue server is operable utilize the environmental awareness from the data processed by the at least one data analysis engine in combination with additional information to create actionable data.

Abstract: The disclosure presents herein a method to train a classifier in a machine learning using more than one simultaneous sample to address class imbalance problem in any discriminative classifier. A modified representation of the training dataset is obtained by simultaneously considering features based representations of more than one sample. A modification to an architecture of a classifier is needed into handling the modified date representation of the more than one samples. The modification of the classifier directs same number of units in the input layer as to accept the plurality of simultaneous samples in the training dataset. The output layer will consist of units equal to twice the considered number of classes in the classification task, therefore, the output layer herein will have four units for two-class classification task. The disclosure herein can be implemented to resolve the problem of learning from low resourced data.

Abstract: In one embodiment, a computer-implemented method performed by a data processing (DP) accelerator, includes receiving, at the DP accelerator, first data representing a set of training data from a host processor; receiving, at the DP accelerator, a watermark kernel from the host processor; and executing the watermark kernel within the DP accelerator on an artificial intelligence (AI) model. The watermark kernel, when executed, is configured to: generate a watermark, train the AI model using the set of training data, and implant the watermark within the AI model during training of the AI model. The DP accelerator then transmits second data representing the trained AI model having the watermark implanted therein to the host processor. In an embodiment, the method further includes receiving a pre-trained AI model and the training is performed for the pre-trained AI model.