Abstract: Methods, apparatus, systems, and articles of manufacture for conditional classifier chaining in a constrained machine learning environment are disclosed. An example apparatus includes a classification controller to select a first model to be utilized to classify a first feature identified from sensor data. A memory controller is to copy the first model to a memory. A machine learning processor is to apply the first model to the first feature to create a first classification output, the first classification output indicating an identified class. The classification controller is to, in response to a determination that the first classification output identifies a second model to be used for classification, instruct the memory controller to load the second model into the memory. The machine learning processor is to apply the second model to the second feature to create a second classification output.

Abstract: A metric value is determined for each event in a set of events that characterizes a computational communication or object. For example, a metric value could include a length of a URL or agent string in the event. A subset criterion is generated, such that metric values within the subset are relatively separated from a population's center (e.g., within a distribution tail). Application of the criterion to metric values produces a subset. A representation of the subset is presented in an interactive dashboard. The representation can include unique values in the subset and counts of corresponding event occurrences. Clients can select particular elements in the representation to cause more detail to be presented with respect to individual events corresponding to specific values in the subset. Thus, clients can use their knowledge system operations and observance of value frequencies and underlying events to identify anomalous metric values and potential security threats.

Abstract: Described herein are systems and methods to identify and address sources of bias in an end-to-end speech model. In one or more embodiments, the end-to-end model may be a recurrent neural network with two 2D-convolutional input layers, followed by multiple bidirectional recurrent layers and one fully connected layer before a softmax layer. In one or more embodiments, the network is trained end-to-end using the CTC loss function to directly predict sequences of characters from log spectrograms of audio. With optimized recurrent layers and training together with alignment information, some unwanted bias induced by using purely forward only recurrences may be removed in a deployed model.

Abstract: A system and computer-implemented is provided for displaying a configurable metric relating to an environment in a graphical display along with a value of the metric calculated over a configurable time period. The metric is used to identify events of interest in the environment based on processing real time machine data from one or more sources. The configurable metric is selected and a corresponding value is calculated based on the events of interest over the configurable time period. The value of the metric may be continuously updated in real time based on receiving additional real-time machine data and displayed in a graphical interface as time progresses. Statistical trends in the value of the metric may also be determined over the configurable time period and displayed in the graphical interface as well as an indication if the value of the metric exceeds a configurable threshold value.

Abstract: A system and computer-implemented is provided for displaying a configurable metric relating to an environment in a graphical display along with a value of the metric calculated over a configurable time period. The metric is used to identify events of interest in the environment based on processing real time machine data from one or more sources. The configurable metric is selected and a corresponding value is calculated based on the events of interest over the configurable time period. The value of the metric may be continuously updated in real time based on receiving additional real-time machine data and displayed in a graphical interface as time progresses. Statistical trends in the value of the metric may also be determined over the configurable time period and displayed in the graphical interface as well as an indication if the value of the metric exceeds a configurable threshold value.

Abstract: A metric value is determined for each event in a set of events that characterizes a computational communication or object. For example, a metric value could include a length of a URL or agent string in the event. A subset criterion is generated, such that metric values within the subset are relatively separated from a population's center (e.g., within a distribution tail). Application of the criterion to metric values produces a subset. A representation of the subset is presented in an interactive dashboard. The representation can include unique values in the subset and counts of corresponding event occurrences. Clients can select particular elements in the representation to cause more detail to be presented with respect to individual events corresponding to specific values in the subset. Thus, clients can use their knowledge system operations and observance of value frequencies and underlying events to identify anomalous metric values and potential security threats.

Abstract: Methods, systems, articles of manufacture and apparatus to control data acquisition settings in edge-based deployments are disclosed. An example apparatus includes a model generator to transform sensor data to variance data, and differentiate the variance data to generate variance rate of change data. The example apparatus also includes a model analyzer to determine subsets of the variance rate of change data associated with respective data acquisition settings, determine a count of data points corresponding to the rate of change data, and determine an interval spacing value based on the count of the data points and a number of subsets of the variance rate of change data.

Abstract: There is provided the use of compounds of formula I wherein R1, R2, R3 and X have meanings given in the description, for the preparation of a medicament for killing clinically latent microorganisms. There is also provided the use of compounds of formula I for treating microbial infections, as well as certain compounds of formula I per se.

Abstract: A disclosed computer-implemented method includes receiving and indexing the raw data. Indexing includes dividing the raw data into time stamped searchable events that include information relating to computer or network security. Store the indexed data in an indexed data store and extract values from a field in the indexed data using a schema. Search the extracted field values for the security information. Determine a group of security events using the security information. Each security event includes a field value specified by a criteria. Present a graphical interface (GI) including a summary of the group of security events, other summaries of security events, and a remove element (associated with the summary). Receive input corresponding to an interaction of the remove element. Interacting with the remove element causes the summary to be removed from the GI. Update the GI to remove the summary from the GI.

Abstract: A dual-diaphragm loudspeaker driver assembly can include a multiple pole magnet structure, and first and second pole piece assemblies can be provided on opposite first and second sides of the multiple pole magnet structure. In an example, each pole piece assembly defines an airgap over a polarity transition region on a respective side of the magnet structure. First and second voice coils can be provided in respective ones of the airgaps, wherein each of the voice coils is coupled to a respective diaphragm assembly, and at least one acoustic tuning port can be configured to provide a damped acoustic communication path between first and opposite second sides of each diaphragm assembly.

Abstract: A disclosed computer-implemented method includes receiving and indexing the raw data. Indexing includes dividing the raw data into time stamped searchable events that include information relating to computer or network security. Store the indexed data in an indexed data store and extract values from a field in the indexed data using a schema. Search the extracted field values for the security information. Determine a group of security events using the security information. Each security event includes a field value specified by a criteria. Present a graphical interface (GI) including a summary of the group of security events, other summaries of security events, and a remove element (associated with the summary). Receive input corresponding to an interaction of the remove element. Interacting with the remove element causes the summary to be removed from the GI. Update the GI to remove the summary from the GI.

Abstract: Aspects of a micromirror spectrophotometer assembly are described. In one example case, an instrument includes a diffraction grating to disperse broadband light over a range of wavelengths, a detector, a digital micromirror device (DMD) configured to scan through and reflect at least a portion of the range of wavelengths toward the detector, and a base platform having a number of integrally formed assembly mounts. The assembly mounts are formed to align and secure the diffraction grating, the detector, the DMD, and other optical components of the instrument in a predetermined arrangement. The instrument can also include a reference paddle having a reference material for calibration of the instrument, and a rotatable sample tray to rotate a sample placed on the sample tray for measurement.

Abstract: An apparatus is disclosed to compress packets, the apparatus comprising; a data analyzer to identify a new destination address and a protocol identifier of an input packet corresponding to a new destination node and a communication system between the new destination node and a source node; a compression engine to utilize a plurality of compression functions based on the new destination address and the protocol identifier and reduce a size of the input packet; a compression analyzer to identify a reduced packet and a compression function identifier corresponding to the reduced packet, the compression function identifier associated with one of the compression functions; and a source modifier to construct a packet to include the compression function identifier by modifying unregistered values of a protocol identifier by a difference associated with the compression function identifier, the packet to inform the new destination node of a compression function.

Abstract: Methods, apparatus, systems, and articles of manufacture for conditional classifier chaining in a constrained machine learning environment are disclosed. An example apparatus includes a classification controller to select a first model to be utilized to classify a first feature identified from sensor data. A memory controller is to copy the first model to a memory. A machine learning processor is to apply the first model to the first feature to create a first classification output, the first classification output indicating an identified class. The classification controller is to, in response to a determination that the first classification output identifies a second model to be used for classification, instruct the memory controller to load the second model into the memory. The machine learning processor is to apply the second model to the second feature to create a second classification output.

Abstract: Methods, apparatus, systems and articles of manufacture to improve accuracy of a fog/edge-based classifier system are disclosed.

Abstract: The present invention relates to novel aminopyridyloxypyrazole compounds that inhibit the activity of transforming growth factor beta receptor 1 (TGF?R1), pharmaceutical compositions comprising the compounds, and methods of using the compounds to treat cancer, preferably colon cancer, melanoma, hepatocellular carcinoma, renal cancer, glioblastoma, pancreatic cancer, myelodysplastic syndrome, lung cancer, and gastric cancer, and/or fibrosis, preferably liver fibrosis and chronic kidney disease.

Abstract: The present invention relates to novel aminopyridyloxypyrazole compounds that inhibit the activity of transforming growth factor beta receptor 1 (TGF?R1), pharmaceutical compositions comprising the compounds, and methods of using the compounds to treat cancer, preferably colon cancer, melanoma, hepatocellular carcinoma, renal cancer, glioblastoma, pancreatic cancer, myelodysplastic syndrome, lung cancer, and gastric cancer, and/or fibrosis, preferably liver fibrosis and chronic kidney disease.

Abstract: The present invention relates at least in part to methods for the manufacture of shape memory polymer (SMP) materials. Particularly, although not exclusive, the present invention relates to processes for the formation of complex shaped devices composed of shape memory polymer.

Abstract: The present invention relates to compositions comprising shape memory polymer (SMP) materials and uses thereof. Particularly, although not exclusively, the present invention relates to biocompatible shape memory polymer (SMP) materials and uses thereof in the medical field.

Abstract: The present invention relates at least in part to surgical devices which comprise a shape memory polymer material composition. Particularly, although not exclusively, the present invention relates to a fixation device e.g. an anchor device e.g. a suture anchor which comprises a shape memory material. Included in the present invention are anchor devices e.g. suture anchors which are formed entirely of a shape memory polymer material. Embodiments of the present invention comprise hybrid suture anchors, particularly suture anchors which are formed from a shape memory polymer material and a non-shape memory material. Methods of securing an anchor in a bone or tissue are also included in the present invention.