The technology disclosed provides a so-called “pointer sentinel mixture architecture” for neural network sequence models that has the ability to either reproduce a token from a recent context or produce a token from a predefined vocabulary. In one implementation, a pointer sentinel-LSTM architecture achieves state of the art language modeling performance of 70.9 perplexity on the Penn Treebank dataset, while using far fewer parameters than a standard softmax LSTM.

Due to the high language use variability in real-life, manual construction of semantic resources to cover all synonyms is prohibitively expensive and may result in limited coverage. Described herein are systems and methods that automate the process of synonymy resource development, including both formal entities and noisy descriptions from end-users. Embodiments of a multi-task model with hierarchical task relationship are presented that learn more representative entity/term embeddings and apply them to synonym prediction. In model embodiments, a skip-gram word embedding model is extended by introducing an auxiliary task “neighboring word/term semantic type prediction” and hierarchically organize them based on the task complexity. In one or more embodiments, existing term-term synonymous knowledge is integrated into the word embedding learning framework. Embeddings trained from the multi-task model embodiments yield significant improvement for entity semantic relatedness evaluation, neighboring word/term semantic type prediction, and synonym prediction compared with baselines.

A method for machine learning based artifact rejection is provided. The method may include applying a machine learning model to identify artefactual independent components in transcranial magnetic stimulation electroencephalogram data collected during a transcranial magnetic stimulation procedure. Clean transcranial magnetic stimulation electroencephalogram data is generated by removing, from the transcranial magnetic stimulation electroencephalogram data, the artefactual independent components. Real-time adjustments to parameters of the transcranial magnetic stimulation procedure may be performed based on the clean transcranial magnetic stimulation electroencephalogram data. Related systems and articles of manufacture, including computer program products, are also provided.

A method for machine learning based artifact rejection is provided. The method may include applying a machine learning model to identify artefactual independent components in transcranial magnetic stimulation electroencephalogram data collected during a transcranial magnetic stimulation procedure. Clean transcranial magnetic stimulation electroencephalogram data is generated by removing, from the transcranial magnetic stimulation electroencephalogram data, the artefactual independent components. Real-time adjustments to parameters of the transcranial magnetic stimulation procedure may be performed based on the clean transcranial magnetic stimulation electroencephalogram data. Related systems and articles of manufacture, including computer program products, are also provided.

Techniques are described for generating HLSD for a textual format source code, which, when rendered, causes a display of visual content. The rendering of the source code generates a tree hierarchy of visual source elements, which logically is possible to map to any graph tree. In an embodiment, visual source elements of the source code are classified to higher-level semantic data (HLSD) labels based on their property(s) and/or the property(s) of neighbor visual source element(s) in the tree hierarchy (context). The HLSD labels indicate the type of HLSD widget mapped to the visual source elements. Techniques further include determining features and a layout arrangement for HLSD widgets and generating a template thereof for the visual content.

A method, software, database and system for attribute partner identification and social network based attribute analysis are presented in which attribute profiles associated with individuals can be compared and potential partners identified. Connections can be formed within social networks based on analysis of genetic and non-genetic data. Degrees of attribute separation (genetic and non-genetic) can be utilized to analyze relationships and to identify individuals who might benefit from being connected.

System monitors and methods of monitoring a system are disclosed. In one arrangement a system monitor predicts a future state of a system. A data receiving unit receives system data representing a set of one or more measurements performed on the system. A first statistical model is fitted to the system data. The first statistical model is compared to each of a plurality of dictionary entries in a database. Each dictionary entry comprises a second statistical model. The second statistical model is of the same general class as the first statistical model and obtained by fitting the second statistical model to data representing a set of one or more previous measurements performed on a system of the same type as the system being monitored and having a known subsequent state. A prediction of a future state of the system being monitored is output based on the comparison. The first statistical model and the second statistical model are each a stochastic process or approximation to a stochastic process.

Determining a quality score for a part manufactured by an additive manufacturing machine based on build parameters and sensor data without the need for extensive physical testing of the part. Sensor data is received from the additive manufacturing machine during manufacture of the part using a first set of build parameters. The first set of build parameters is received. A first algorithm is applied to the first set of build parameters and the received sensor data to generate a quality score. The first algorithm is trained by receiving a reference derived from physical measurements performed on at least one reference part built using a reference set of build parameters. The quality score is output via the communication interface of the device.

Determining a quality score for a part manufactured by an additive manufacturing machine based on build parameters and sensor data without the need for extensive physical testing of the part. Sensor data is received from the additive manufacturing machine during manufacture of the part using a first set of build parameters. The first set of build parameters is received. A first algorithm is applied to the first set of build parameters and the received sensor data to generate a quality score. The first algorithm is trained by receiving a reference derived from physical measurements performed on at least one reference part built using a reference set of build parameters. The quality score is output via the communication interface of the device.

Systems and methods for managing Application Programming Interfaces (APIs) are disclosed. Systems may involve automatically generating a honeypot. For example, the system may include one or more memory units storing instructions and one or more processors configured to execute the instructions to perform operations. The operations may include receiving, from a client device, a call to an API node and classifying the call as unauthorized. The operation may include sending the call to a node-imitating model associated with the API node and receiving, from the node-imitating model, synthetic node output data. The operations may include sending a notification based on the synthetic node output data to the client device.