Tata Consultancy Services Patente

Agriculture is impacted due to various parameters that results into heavy losses to the crops. Existing methods rely on manual inventories from fields and statistical results are prone to errors thus classification of crops may not be accurate. Present disclosure provides systems and methods for detection and prediction of disguised and potential non-performing assets (NPAs) in agriculture. Disguised NPAs detection involves, obtaining satellite data from sources and natural calamity, detecting of crop for past years, crop growth, estimating yields, market prices for individual years, AC zones, etc. wherein a generate performance score is generated for each field for classifying crop as disguised NPA. Predicting potential NPAs includes obtaining current season satellite data, data received from the fields via sensors, determining crop protocol being followed, forecast weather, estimating cumulative growth, feedback received from fields, projected yield, estimating crop losses, and then performance score is generated for predicting crops as potential NPA.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

State of art techniques relate to telepresence robots expressing internal states using emotional expressions and corresponding text messages are based on the perceived input from the co-located user or the participant, and not related to the teleoperator. Embodiments of the present disclosure provide a method and system for expressing telepresence robot internal states using combination of multiple modalities. The telepresence robot expresses its own internal states in the form of a plurality of emotional expressions and corresponding text messages to the teleoperator, using a robot internal state predictor model. Unlike the state of art techniques which express an emotion as its internal state, the disclosed method expresses the internal state as the emotional expression with respect to a task processing of the telepresence robot.

Planning cardiac surgery is a complex task. Current heart models are mostly static and provide no functional information or personalization apart from structural information from imaging modalities like MRI, CT etc. The embodiments of present disclosure addresses unresolved problem of generating a personalized four dimensional cardiac model that integrates blood flow velocity, pressure profile and volume on three-dimensional cardiac structure, providing a comprehensive tool for pre or post-surgical planning and analysis. Unlike existing techniques, which often rely on separate assessments of cardiac anatomy and hemodynamics, using invasive methods, the present disclosure provides a 4D cardiac model that captures the person specific dynamic interaction between cardiac structures and blood flow throughout the cardiac cycle. Further, the 4D cardiac model enables clinicians to visualize and quantify complex flow patterns and pressure distributions within the region of interest in the heart, facilitating precise surgical planning and outcome prediction especially in pediatric heart surgery.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

This disclosure relates generally to system and method for molecular property prediction. The method utilizes a set-pooling aggregation operator to derive a graph-level representation of a complete input molecular graph to assist in inductive learning tasks. The method includes iteratively down-sampling the molecular graph into a coarsened molecular graph, and determining adjacency matrix and feature matrix of the coarsened molecular graph. The method then includes computing an average of the hidden state node attributes of the coarsened graph obtained after preforming the iterations to obtain a graph level representation vector of the molecular graph. Using a linear layer from the graph level representation vector the molecular properties are determined.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

This disclosure relates generally to method and system for structure-based drug design using a multi-modal deep learning model. The method processes a target protein for designing at least one optimized molecule by using a multi-modal deep learning model. The GAT-VAE module obtains a latent vector of at least one active site graph comprising of key amino acid residues from the target protein. The SMILES-VAE module obtains at least one latent vector from the target protein. Further, the conditional molecular generator concatenates the active site graph with the latent vector to generate a set of molecules. The RL framework is iteratively performed on the concatenated latent vector to optimize at least one molecule by using the drug-target affinity (DTA) predictor module to predict an affinity value for the set of molecules towards the target protein. Further, at least one optimized molecule is designed with an affinity of the target protein.

Embodiments herein provide automated bin packing of heterogeneous objects to fill maximum number of objects in an available bin space in an efficient manner. Conventional methods are less efficient in packing heterogeneous objects in a space and time efficient manner. The apparatus receives a plurality of 3D objects through a primary conveyor and diverts each of the plurality of 3D objects to an identified cuboidal bin through a corresponding secondary conveyor. Each of the plurality of 3D objects are diverted based on a tracking code associated with each of the plurality of 3D objects and a computed available bin space of the identified cuboidal bin corresponding to each of the plurality of 3D objects. Here the available bin space is computed based on a bin space information received from a corresponding bin vision sensor and a corresponding state space of the identified cuboidal bin.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Unlike visual similarity, visual compatibility is a complex concept. Existing approaches for outfit compatibility prediction does not focus on methods with personalization. The present disclosure proposes a novel approach to model the user's preference for different styles. The outfit compatibility prediction module is a critical component of an outfit recommendation system. An outfit is said to be compatible if all the items are visually compatible and match the user's preferences. The present disclosure represents the outfit as a graph and uses Graph Neural Network (GNN) with attention mechanism to capture the inter-relationship between the items. A graph readout layer generates the final outfit embedding. The proposed approach efficiently models the preferences of the users for different styles. Finally, the outfit compatibility score is generated by computing the similarity between the outfit embedding and the user embedding.

Providing the right mechanism to apply blockchain technology with flexible or dynamic credit disbursement while also addressing valuation of the agricultural assets that are dynamic in nature is required. A method and a system for estimating flexible credit eligibility and disbursement schedule using NFTs of agricultural assets is disclosed. A detailed blockchain based approach is provided for evaluating credit collateral of the NFTs representing the agricultural assets, also interchangeably referred to as physical farm assets or agri-assets. Valuation of the credit eligibility is based on flexibility provided to the farmer on deciding the farm assets to be used, current status of the farmer selected agricultural assets and credit equivalent value of the each of NFTed asset selected by the farmer for the season/year. The dynamic disbursement schedule is based on the personalized or customized parameters associated with the current status of farm and farm assets linked to the credit collateral.

Zusammenfassung wird geladen …

State of the art telescope designs require increasing number of sub-apertures for optimum performance, however, with the increasing number of sub-apertures, number, and amplitudes of the sidelobes increase along with that of the primary maxima, resulting in a trade-off of the imaging quality. Disclosed herein are three configurations using a central sub-aperture and a plurality of peripheral sub-apertures, encompassing the central sub-aperture. Size of the central sub-aperture and the plurality of peripheral sub-apertures is in a proportionate relationship. Further, the plurality of the peripheral sub-apertures forms at least two concentric zones, wherein each concentric zone has equal number peripheral sub-apertures from among the plurality of peripheral sub-apertures, and the sizes of the peripheral sub-apertures in each two adjacent concentric zones have a proportionate relationship. This way there is significant side lobe suppression compensating for the imaging performance loss due to reduced aperture area.

Post pandemic, retailers are adopting more contactless services for shopper's checkout such as self-checkout, hybrid checkout and mobile checkout, and these touchpoints have become the potential areas for fraudulent activity during check out process. For detecting fraud carried out by a customer at the time of self-checkout, existing approaches require respective customer identity and his purchase history. Embodiments of the present disclosure do not require customer identity and information about his historical shopping carts and provide a method and system for approving a user shopping cart for self-checkout from the items picked by the customers in real time.

Zusammenfassung wird geladen …

This disclosure relates generally to identification and mitigation of bias while training deep learning models. Conventional methods do not provide effective methods for bias identification, and they require pre-defined concepts and rules for bias mitigation. The embodiments of the present disclosure train an auto-encoder to produce a generalized representation of an input image by decomposing into a set of latent embedding. The set of latent embedding are used to learn the shape and color concepts of the input image. The feature specialization is done by training an auto-encoder to reconstruct the input image using the shape embedding modulated by color embedding. To identify the bias, permutation invariant neural network is trained for classification task and attribution scores corresponding to each concept embedding are computed. The method also performs de-biasing the classifier by training it with a set of counterfactual images generated by modifying the latent embedding learned by the auto-encoder.

Small and compact Deep Learning models are required for embedded AI in several domains. In many industrial use-cases, there are requirements to transform already trained models to ensemble embedded systems or re-train those for a given deployment scenario, with limited data for transfer learning. Moreover, the hardware platforms used in embedded application include FPGAs, AI hardware accelerators, System-on-Chips and on-premises computing elements (Fog / Network Edge). These are interconnected through heterogenous bus / network with different capacities. Method of the present disclosure finds how to automatically partition a given DNN into ensemble devices, considering the effect of accuracy - latency power - tradeoff, due to intermediate compression and effect of quantization due to conversion to AI accelerator SDKs. Method of the present disclosure is an iterative approach to obtain a set of partitions by repeatedly refining the partitions and generating a cascaded model for inference and training on ensemble hardware.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

This disclosure relates generally to system and method for molecular property prediction. The disclosed method includes mapping node embeddings of a molecular graph to a graph-level embedding characterizing the molecular graph. The graph level representation is acquired by pooling characteristics of hidden states of the nodes in the molecular graph by performing an iterative content based attention in a plurality of iterations. The content based attention is performed by considering an edge information fused transformed hidden state vector of the nodes of the molecular graph. The graph level embedding is fed through the linear projection to predict the molecular properties of the molecular graphs.

Conventional task planners assume that the task-plans provided are executable, hence these are not task-aware. Present disclosure alleviates the downward refinability assumption, that is, planning can be decomposed separate symbolic and continuous planning steps by introducing bi-level planning, a plan which is a series of actions that the robot needs to take to achieve the goal task is curated. Firstly, abstract symbolic actions are converted to continuous vectors and used therein to enable interaction with an environment. Images of objects placed in the environment are captured and concepts are learnt from the captured images and attributes of objects are detected. A hierarchical scene graph is generated from the concepts and attributes wherein the graph includes interpretable sub-symbolic representations and from these interpretable symbolic representations are obtained for identifying goal task. Anomalies are detected from the scene graph and robotic actions are generated to correct the detected anomalies.

Diagnosis of rare human diseases using DNA sequencing is a fast growing area of research. Conventional methods carries a risk of incorrect phenotype interpretation. However, obtaining a correct genotype and phenotype matching is challenging. A system for matching phenotype descriptions and pathogenic variants provides a one to one mapping of the phenotype and genotypes of a plurality of subjects under test. Initially, a plurality of phenotypes and a plurality of genome sequences are segmented based on metadata. A phenotype driven gene prioritization and a variant prioritization is applied on the segmented data method. A similarity score is calculated between the phenotype driven gene prioritization output and the variant prioritization output. The similarity score is further utilized to obtain a one to one matching of the plurality of phenotypes and the plurality of genotype sequences of the plurality of subjects under test.

Zusammenfassung wird geladen …

The quality of user experience (UX) of an information visualization depends on multiple diverse aspects. These include qualitative, quantitative, and contextual parameters that are unmeasurable and incomparable. Hence, measuring the UX of a visualization is challenging. The disclosure herein relates to a system and method that collects, processes, and analyzes a multiple diverse parameters to measure and profile the UX of a visualization. To accomplish this, the system collects data regarding the usage, effectiveness, and user perception of the visualization. The system creates a quantitative and comparable version of all these parameters to measure holistically the UX of the visualization. This involves quantifying qualitative values, considering context by objectively computing the complexity of charts, complexity based benchmarking, summarizing individual parameters into hedonic and pragmatic scores, plotting the visualization on a scatter chart, dividing the scatter chart based on low-high hedonic and pragmatic scores, labeling each section and labeling the visualization.

Zusammenfassung wird geladen …

Coarse resolution satellite data has limited information content of land captured in the images and small area field boundary detection is technically challenging. Embodiments of the present disclosure provide a method and system for detection of field boundaries using coarse resolution satellite data and GIS. Multi-layer crop segmentation approach is used over RoI, wherein agricultural land related segments are derived from a soil layer, a temporal stack for a current season and a temporal change layer from the coarse resolution satellite data of the RoI. Segmentation based on different aspects captures various details at each segmentation layer. The segments are then combined, which aggregates all information related to the land to accurately detect the field boundaries in the RoI. The field boundaries are further refined using GIS topology operations. Baseline cadastral maps are updated using information from maps generated using refined field boundaries identified in real time for each season.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …