Abstract: This disclosure relates generally to a method and system for latency optimized heterogeneous deployment of convolutional neural network (CNN). State-of-the-art methods for optimal deployment of convolutional neural network provide a reasonable accuracy. However, for unseen networks the same level of accuracy is not attained. The disclosed method provides an automated and unified framework for the convolutional neural network (CNN) that optimally partitions the CNN and maps these partitions to hardware accelerators yielding a latency optimized deployment configuration. The method provides an optimal partitioning of the CNN for deployment on heterogeneous hardware platforms by searching network partition and hardware pair optimized for latency while including communication cost between hardware. The method employs performance model-based optimization algorithm to optimally deploy components of a deep learning pipeline across right heterogeneous hardware for high performance.

Abstract: State of the art approaches used to address security aspects in serverless platforms perform workflow validations on an end to end flow, however, this cannot prevent attacks targeted at intermediate function calls in the workflow. Further, the existing systems store policy data in insecure manner, which causes security issues. The disclosure herein generally relates to serverless clouds, and, more particularly, to a method and system for privacy-preserving workflow validations in serverless clouds. The system stores policy data in a secured/encrypted manner. The system also performs validations at different levels, at a first level to allow/deny access at an ingress point, and at a second level to allow/deny access at critical intermediate points. This approach thus provides safety against attacks that may have been initiated post initial validation, and offers added data security.

Abstract: Traditionally, counterweight fork type autonomous mobile robots (AMR) have been used for any kind of pallet. But the challenge is it occupies lot more maneuvering space while making turns, which cannot work in narrow operating zones. Hence fork over AMR is preferred. However, these fork over AMR have extended parts always touching the ground surface and thus are not suitable for pallets with a wooden plank at the bottom of the fork opening in the pallet. To overcome the above technical problems, an Adjustable Counterweight-based Fork Type Autonomous Mobile Robot (ACFTAMR) is provided that includes chassis assembly and vertical mast unit, a horizontal cross slide mechanism and forks. The chassis assembly is provided counterweight assembly and counterbalance shafts that move forward and backward during pickup and release of payload when the vertical mast unit moves in upward/downward direction, thus providing better stability and counterbalance to the ACFTAMR.

Abstract: This disclosure relates to systems and methods for shapelet decomposition based recognition using radar. State-of-the-art solutions involve use of standard machine learning classification techniques for gesture recognition which suffer with problem of dependency on collected data. The present disclosure overcome the limitations faced by the state-of-the-art solutions by obtaining a plurality of time domain signal using a radar system comprising three vertically arranged radars and one or more sensors, identifying one or more gesture windows to obtain one or more shapelets corresponding to one or gestures which are further decomposed into a plurality of sub shapelets. Further, at least one of (i) a positive or (i) a negative time delay is applied to each of the plurality of sub shapelets to obtain a plurality of composite shapelets which are further mapped with a plurality of trained shapelets to recognize gestures comprised in one or more activities performed by a subject.

Abstract: Wiper is utilized to clean side-view mirror of a vehicle. Existing wiper attached with the side-view mirror obstructs visibility of a driver while cleaning the side-view mirror in a rainy season or in a dusty environment, which leads to an accident. A vehicle mirror wiping unit is provided. The vehicle mirror wiping unit includes a mirror; a mirror shutter assembly; a mirror casing; a mirror mounting unit, consisting of: the mirror, a mirror spring, a spring retaining cap, and a mirror spring holder; a pin pass through a cut-out of the spring retaining cap, and a cut-out in the mirror spring holder; and a cleaning elastomer is triggered by the mirror shutter motor, and the mirror shutter moves from the first end of the mirror to the second end of the mirror. A frontal surface of the mirror shutter is a reflective surface functioning as a mirror.

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: One of the biggest challenges faced by oil and gas companies is to monitor such long pipelines for leak events and generate false leak event alarms during routine pipe maintenance. A data associated with a first sensing unit is processed to obtain an instant timing information (T0) of a leak event in a conduit at a test environment. A data associated with a second sensing unit is processed to obtain a transient signal associated with the leak event at a specific band. An accelerometer data is filtered to obtain a band passed filtered accelerometer signal (Accelbpf). The Accelbpf is truncated in a time domain from the T0 to a duration Td of the leak event to obtain a temporal template signal (Acceltemplate). A leak event of a real-time conduit is dynamically detected at a physical environment based on Acceltemplate when a cross-correlation value is greater than a threshold value (?).

Abstract: The present disclosure provides a technique to evaluate encrypted Machine Learning (ML) models. Conventional methods are unable to provide a holistic approach to evaluate encrypted ML models. Initially, the system receives an encrypted ML model. The ML model can be an unencrypted ML model trained with encrypted data or an encrypted ML model trained with encrypted data, or an encrypted ML model trained with unencrypted data. Further, a plurality of evaluation functions pertaining to the ML model to be calibrated are identified using a pattern matching technique. Further, an approximated function is generated for each of the plurality of evaluation functions using a corresponding approximation technique. After generating a plurality of approximated functions, an Expected Calibration Error (ECE) value is computed based on the plurality of approximated functions. Finally, the ML model is calibrated based on the computed ECE value. The ML model is perfectly calibrated if the computed ECE value is equal to zero.

Abstract: Re-usage of part of object or object is highly important in manufacturing industry as it can drastically reduce cost and time spent on manufacturing. However, lack of proper information about availability of similar parts leads to redesigning of similar part. Existing databases for engineering drawings do not store categorized information due to which performing feature-based search is not possible. Present application provides systems and methods for extracting, digitizing, and using engineering drawing data. The system receives engineering drawing document and extracts text data present in each cell of table provided in document. Once table data is extracted, isometric views and views other than isometric views that are present in document are identified by the system using pretrained machine learning based model. The system further extract view labels and coordinate information from identified views.

Abstract: The present disclosure is of a method for automated authoring of purposive models from Natural Language (NL) documents. Conventional model extractors to automatically extract models from NL documents are specific to a metamodel, do not consider document structure and are not configurable. Initially, the system receives a plurality of Natural Language (NL) documents, a metamodel, a plurality of pattern trees corresponding to the metamodel, and a configurable domain dictionary. Each of the plurality of pattern trees includes a plurality of pattern elements. Further, a document information is generated from each of the plurality of NL documents using a document information reading technique. Finally, a plurality of purposive models are generated for each of the plurality of NL documents by interpreting a corresponding document information based on the plurality of pattern trees and the metamodel using a pattern interpretation technique.

Abstract: This disclosure relates generally to method and system for jointly pruning and hardware acceleration of pre-trained deep learning models. The present disclosure enables pruning a plurality of DNN models layers using an optimal pruning ratio. The method processes a pruning request to transform the plurality of DNN models and the plurality of hardware accelerators into a plurality of pruned hardware accelerated DNN models based on at least one user option. The first pruning search option executes a hardware pruning search technique to perform search on each DNN model and each processor based on at least one of a performance indicator and an optimal pruning ratio. The second pruning search option executes an optimal pruning search technique, to perform search on each layer with corresponding pruning ratio.

Abstract: Stubble burning is a serious problem resulting in pollution attributable to smog, loss of nutrients in the top soil, and risk of fires spreading out of control. Existing methodologies have attempted to predict burning areas, but have failed to do so because of inefficient sensing mechanism. Present disclosure proposes a system and method to compute burning index score pertaining to crops by detecting harvest season and predicting probable areas of burning by combining current year's crop area map along with harvesting period and historical hot spot information. Computation of the burning index score is accomplished based on inputs received from at least one of satellite imaging, multi-spectral drone based sensing devices and crowdsourcing information. This will help to prioritize the area for taking corrective measures such as training of farmers, availability of resources, optimizing the resources schedule, etc.

Abstract: Embodiments provide a method and system for audio generation from contextual text input is provided. The disclosure gives due importance to the granularity of the content. The system allows the user to specify the properties of the audio to be generated. Here, context is used to identify the importance of a particular sound over the others and thus automatic adjustments of the audio output to give a more realistic feel. The system generates dataset for training audio models. The user can give input query in natural language and the audio requested will be generated for training and developing the necessary classification or other necessary audio models. The system provides a feature of automated fine-tuning of the model parameters to suit the new automatically collected training data. Furthermore, the system provides a pre-trained inbuilt model repository with audio models belonging to the main categories of noises.

Abstract: This disclosure relates generally to system and method for forecasting location of target in monocular first person view. Conventional systems for location forecasting utilizes complex neural networks and hence are computationally intensive and requires high compute power. The disclosed system includes an efficient and light-weight RNN based network model for predicting motion of targets in first person monocular videos. The network model includes an auto-encoder in the encoding phase and a regularizing layer in the end helps us get better accuracy. The disclosed method relies entirely just on detection bounding boxes for prediction as well as training of the network model and is still capable of transferring zero-shot on a different dataset.

Abstract: Conventionally, manufacturing of molded parts using composite materials has led to poor dimensional accuracy and tensile strength due to improper curing thus resulting in rejection or early/premature failure of composite part. Embodiments of the present disclosure provide simulation-based systems and methods for manufacturing/generating molded parts using reinforced composite materials. The optimized cure cycle is computed for a given component without carrying out numerous experiments. The present disclosure implements multiscale method and surrogate modeling in virtual testing for more accurate and faster manufacturing of molded parts. Process parameters for specified qualities (e.g., minimum residual stresses, minimum deformation, etc.) required for a part are determined along with least process manufacturing time. The resulting optimized time dependent cure cycle for each thermal zone of the heated mold is transferred to a master controller (e.g.

Abstract: There is huge demand for automation in manufacturing, logistics, postal, distribution centers, ecommerce, retail. Typical scissor lift designs can carry large payload but are larger in size. There is a need for a compact fork with more payload carrying capacity. The dual side actuated scissor fork type lift unit is provided. The dual side actuated scissor fork type lift unit includes a top plate, a bottom plate, one or more linear motion (LM) blocks mounted on one or more linear guides, one or more mounting blocks mounted on the one or more LM blocks, and slot on a first end of at least one actuation link is connected to the bottom plate through a pin. The one or more LM blocks is free to slide on the one or more linear guides. The one or more mounting blocks moves along with a motion of the one or more LM blocks.

Abstract: In retail, macro space optimization is carried out at individual stores to allocate optimum space for each category. Each retailer has many stores and macro space optimization is experimented with different objectives such as expand space, reduce space and constant space of a store individually. Thus, the number of space recommendations analyzed at corporate level increases extremely high and making it difficult to bring key inferences out of these recommendations and creating challenges in implementation of results such as creation of planograms and floor plans. Embodiments of the present disclosure provide a method and system for identifying underlying patterns that reside in space recommendations across stores and creating drastically reduced number of floor plans and planograms in accordance with the identified set of patterns unlike large number of floor plans or planograms generated by state of the art space planning systems.

Abstract: Accuracy of transcript is of foremost importance in Automatic Speech Recognition (ASR). State of the art system mostly rely on spelling correction based contextual improvement in ASR, which is generally a static vocabulary based biasing approach. Embodiments of the present disclosure provide a method and system for visual context aware ASR. The method provides biasing using shallow fusion biasing approach with a modified beam search decoding technique, which introduces a non-greedy pruning strategy to allow biasing at the sub-word level. The biasing algorithm brings in the visual context of the robot to the speech recognizer based on a dynamic biasing vocabulary, improving the transcription accuracy. The dynamic biasing vocabulary, comprising objects in a current environment accompanied by their self and relational attributes, is generated using a bias prediction network that explicitly adds label to objects, which are detected and captioned via a state of the art dense image captioning network.

Abstract: The present disclosure detects a pupil of an eye using circle formation based scoring method. The conventional approaches fail to provide an accurate and reliable biometric authentication due to the usage of simple thresholding based statistical methods and iris dependent segmentation methods. The present disclosure utilizes a circle plotting approach and selects the optimum circle using several parameters. The present disclosure can generate a pupil boundary that fits the pupil region inside an iris perfectly. Initially, the system receives an input image of an eye. After removing reflections, a core point of the reflection free image is identified. Further, a plurality of points are obtained based on a sudden gradient change. and a plurality of circles are plotted. Further, an optimum circle is identified using a score based optimum circle selection method. Finally, the pupil associated with the input image is identified based on the optimum circle.