Abstract: One embodiment provides for a computer-readable medium storing instructions that cause one or more processors to perform operations comprising determining a per-layer scale factor to apply to tensor data associated with layers of a neural network model and converting the tensor data to converted tensor data. The tensor data may be converted from a floating point datatype to a second datatype that is an 8-bit datatype. The instructions further cause the one or more processors to generate an output tensor based on the converted tensor data and the per-layer scale factor.

Abstract: Disclosed embodiments relate to instructions for fused multiply-add (FMA) operations with variable-precision inputs. In one example, a processor to execute an asymmetric FMA instruction includes fetch circuitry to fetch an FMA instruction having fields to specify an opcode, a destination, and first and second source vectors having first and second widths, respectively, decode circuitry to decode the fetched FMA instruction, and a single instruction multiple data (SIMD) execution circuit to process as many elements of the second source vector as fit into an SIMD lane width by multiplying each element by a corresponding element of the first source vector, and accumulating a resulting product with previous contents of the destination, wherein the SIMD lane width is one of 16 bits, 32 bits, and 64 bits, the first width is one of 4 bits and 8 bits, and the second width is one of 1 bit, 2 bits, and 4 bits.

Abstract: State-of-the-art approaches have concentrated on building solution(s) to match the amplitude of a time series with a user given one. However, these have failed to implement solution(s) which enables searching for pattern(s) that can depict human vision psychology. Embodiments of the present disclosure determine occurrence of pattern of interest in time series data for anomaly detection, wherein time series data is obtained, and first order derivative is computed. Further an angle of change in direction is derived based on a gradient of change in value of the time series data. This angle is further converted to a measurement unit. The time series data is quantized into bins and a weighted finite state transducers diagram (WFSTD) is obtained based on domain knowledge which is then converted to specific pattern. The specific pattern is searched in the bins to determine occurrence/count of the specific pattern for anomaly detection.

Abstract: Industries deploy a plethora of sensors that are attached to a system or human being, respectively. Under multi-sensor environment scenarios, there is a need to detect which sensors are behaving similarly within a time span. Sensor values often vary in range of values yet depict similar time series characteristic and sometimes have a phase difference in operation, thus making it impossible to detect such sensor similarity in a large system where the number of input parameters/sensor observations. Systems and methods of the present disclosure determine similar behavioral pattern between time series data obtained from multiple sensors and cluster the sensors. The system implements a pattern recognition-based approach to find the similarity and then applies a Dynamic Programming-based approach to detect similarity in at least two time series data and cluster the sensors and corresponding time series data into specific cluster(s).

Abstract: Sensor data (or IoT) analytics plays a critical role in taking business decisions for various entities (e.g., organizations, project owners, and the like). However, scaling of such analytical solutions beyond certain point requires adopting to various computing environments which seems to be challenging with the constrained resources available. Embodiments of the present disclosure provide system and method for analysing and executing sensor observational data in computing environments, wherein extract, transform, load (ETL) workflow pipeline created by users in the cloud, can be seamlessly deployed to job execution service available in cloud/edge without any changes in the code/config by end user. The configuration changes are internally handled by the system based on the selected computing environment and queries are executed either in distributed or non-distributed environments to output data frames.

Abstract: Sensor data (or IoT) analytics plays a critical role in taking business decisions for various entities (e.g., organizations, project owners, and the like). However, scaling of such analytical solutions beyond certain point requires adopting to various computing environments which seems to be challenging with the constrained resources available. Embodiments of the present disclosure provide system and method for analysing and executing sensor observational data in computing environments, wherein extract, transform, load (ETL) workflow pipeline created by users in the cloud, can be seamlessly deployed to job execution service available in cloud/edge without any changes in the code/config by end user. The configuration changes are internally handled by the system based on the selected computing environment and queries are executed either in distributed or non-distributed environments to output data frames.

Abstract: State-of-the-art approaches have concentrated on building solution(s) to match the amplitude of a time series with a user given one. However, these have failed to implement solution(s) which enables searching for pattern(s) that can depict human vision psychology. Embodiments of the present disclosure determine occurrence of pattern of interest in time series data for anomaly detection, wherein time series data is obtained, and first order derivative is computed. Further an angle of change in direction is derived based on a gradient of change in value of the time series data. This angle is further converted to a measurement unit. The time series data is quantized into bins and a weighted finite state transducers diagram (WFSTD) is obtained based on domain knowledge which is then converted to specific pattern. The specific pattern is searched in the bins to determine occurrence/count of the specific pattern for anomaly detection.

Abstract: Conventionally, detecting time when a device is going to fail in real time has been a real challenge given the associated constraints and requirements. Due to absence in any supporting information or annotated data, traditional approaches have failed to detection abnormality in devices. Present disclosure provide systems and methods for detecting abnormal behaviour of a device from associated unlabeled sensor observations wherein KPIs are computed based on unlabeled sensor observations of at least two sensor parameters and windowing technique is applied on modified dataset to obtain windowed dataset based on which hyper-parameters of deep learning-based auto-encoder are optimized to obtain set of embeddings. Dimensionality reduction technique is applied on the embeddings to obtain embeddings with reduced dimension. Density based clustering technique with hyper-parameters is applied on embeddings with reduced dimension and cluster(s) for unlabeled sensor observations are obtained.

Abstract: Advanced analytics refers to theories, technologies, tools, and processes that enable an in-depth understanding and discovery of actionable insights in big data, wherein conventional systems and methods may be prone to errors leading to inaccuracies.