Goddar, Heinz J

The emergence of Sixth Generation (6G) wireless networks presents unprecedented opportunities for revolutionizing smart manufacturing environment through TeraHertz (THz) communication band. However, the associated unique characteristics pose significant challenges for reliable indoor communication. The present disclosure addresses these challenges by providing a context aware spectrum allocation system and method that capture user requirements/intents which are then translated to priority and fairness objectives. Based on mobility of Mobile-edge Entities (MEs), THz channel modelling and dynamic channel state updation are performed and context-based spectrum allocation is then done based on fairness and priority objectives. Further, dynamic mobility and updation of distance of users is captured with respect to a Base Station (BS). Q-learning technique is employed for adaptive decision making based on the reward function. Further, the system integrates Hungarian technique with the Q-learning for global optimization for dynamic sub-band allocation at infrastructure plane level.

The disclosure relates generally to methods and systems for automatic generation of an editable user interface (UI) screen from application screen image. Conventional techniques for automatic generation of the editable user interface screen from application screen image are limited to specific UI page type and challenges with the conversion accuracy. The methods and systems of the present disclosure predict the bounding boxes of different graphical components and their labels present in the given UI image screen using a trained object detection model. Then, the one or more group boxes and the one or more nested group boxes are identified based on their properties and next one or more UI widgets present in both the one or more group boxes and the one or more nested group boxes are identified. Lastly the editable UI screen image is automatically generated from the identified UI widgets.

This disclosure relates generally to a method for automatic knowledge refinement and task execution in robotic agents using a domain-agnostic knowledge base. Conventional Large Language Model (LLM) based methods break down complex tasks into sub-tasks but may generate incorrect steps or refer to unavailable objects or actions. While the robotic agents leverages prior-domain specific-knowledge, the unavailability of such structured knowledge in many practical domains and difficulty in updating data reliably makes agent-based task execution problematic. The integration of the domain-agnostic knowledge base with a Multi-modal Language Models (MLM) or LLM enables the robotic agent to perform tasks that span multiple domains without the constraints imposed by traditional, domain-specific knowledge bases. Combining the domain-agnostic knowledge base and the MLM allows for the generation of refined action sequences that are adaptable to the nuances of complex tasks, thereby enhancing the robotic agent's operational versatility and effectiveness.

This disclosure relates generally to method and system for generating emotional talking head video using disentangled pose and expression-flow guidance. Generating realistic one-shot emotional talking head animation on arbitrary faces is a challenging problem, as it requires realistic emotions, head movements, identity preservation, and accurate lip sync. The method receives a plurality of inputs comprising an identity image, a speech audio input data and an emotion input data to generate emotional talking head video. Further, the plurality of inputs are utilized to generate one or more pose landmarks using the pose generation network, and one or more pose landmarks are generated using the expression generation network. Finally, an image generation network generates emotional talking head video using disentangled optical flow computation for pose-guided and expression-guided motion using one or more expression-invariant pose landmarks and pose-invariant emotion landmarks. Additionally, the method generates accurate facial emotions and emotion accuracy on benchmark datasets.

Recommending an optimal combination of PQC schemes, and the PQC libraries for replacing the classical cryptographic schemes of a cryptographic application is a challenging decision problem. Embodiments herein provide a method for selection of the PQC schemes and the PQC libraries for cryptographic applications. The method first generates one or more candidate PQC libraries, for each of the plurality of cryptographic applications based on a plurality of stringent requirement attributes and a plurality of non-stringent requirement attributes, using a plurality of first compatible PQC libraries, and a plurality of second compatible PQC libraries. Further a schemelibrary pair of the one or more candidate PQC libraries is obtained, for each the plurality of classical cryptographic schemes based on a plurality of performance objective metrics. Then one or more final candidate PQC libraries obtained using the schemelibrary pair which is recommended for the plurality of cryptographic applications of an enterprise.

Computational analysis is essential for efficient design, evaluation, and prediction of processes in industrial systems. However, it requires significant expertise, iterative refinement, and labor-intensive efforts. Embodiments herein provide a method and system for computational analysis of industrial entities using a multimodal Large Language Model (LLM). A plurality of final assumptions required for a plurality of workflows of the computational analysis of the industrial entity are generated by the multimodal LLM. Further the method creates a plurality of prompts based on the generated plurality of final assumptions. A Computer-Aided-Design (CAD) file, a final mesh file, a solver results file, and a plurality of computational analysis result insights are generated using the plurality of prompts, via the multimodal LLM. The method of the present disclosure minimizes user involvement, utilizes a knowledge base to generate relevant outputs, and offers flexibility for seamless integration with various software platforms and tools for the computational analysis.

The present disclosure addresses energy overheads and latency challenges of architecture design of conventional gated recurrent unit (GRU) approaches by providing a Resistive Random-Access Memory In-Memory Computing-based gated recurrent unit (RRAM IMC-based GRU) network architecture for efficient implementation of GRUs. In the present disclosure, the RRAM IMC-based GRU network architecture is used which performs Multiply-Accumulate (MAC) operations using Ohm's law for multiplication and Kirchhoff's current law for accumulation. A plurality of GRU wight parameters are mapped as device conductance in a Resistive Random-Access Memory (RRAM) memristor array structure in a skewed arrangement. An input vector is applied as voltage pulses to wordlines of the RRAM memristor array structure corresponding to values which should be multiplied and accumulated. Outputs of the MAC operation are obtained as bitline currents, which are then sampled and converted to digital values using an Analog-to-Digital Converters for interfacing with other digital post-processing units.

A method of breaking a permanent loop in a multi-protocol and multi-instance data communications network 702 is provided. The multi-protocol data and multi-instance communications network handles a plurality of protocols using a respective plurality of instances of the data communications network. The method includes (i) detecting a permanent loop in data communications for a given destination between a plurality of routers 704A-N in the multi-protocol and multi-instance data communications network, (ii) selecting a first router which is outside of the permanent loop and a second router which is inside the permanent loop, as a rerouting between these routers which will break the permanent loop, and (iii) directing data communications for the given destination from the second router to the first router.

In a competitive retail world, store specific competitive strategies need to be formed by understanding unique characteristics of assortment followed by each specialty competitor within a trade area. Retailers face challenges to get intensive knowledge about store specific competitor assortment due to practical difficulties. Embodiments herein provide a method and system for a generative artificial intelligence (GenAI) based assistant to implement one or more competitive retail strategies associated with specialization of a specialty retail store. Locating competitive position of a specialty retail store and identifying gap in assortment is achieved by capturing unique characteristics of a specialty competitor store which is enabled by applying dynamic enrichment of content in which direction for the enrichment provided by approximating existing patterns that occur among the specialty competitive stores.

A temperature sensor module and a temperature compensation method thereof according to embodiments of the present disclosure are disclosed. The temperature sensor module includes a temperature sensor configured to output a voltage value corresponding to a sensed temperature; a control circuit configured to calculate a first compensation value based on a code value for the voltage value using reference code values at a first temperature and a second temperature higher than the first temperature, and calculate a second compensation value using an average trend line determined as an average value for each temperature in a first temperature interval equal to or lower than the first temperature and a second temperature interval equal to or higher than the second temperature; a 1st compensation circuit configured to apply the first compensation value to the code value to generate a first-stage compensated first code value; and a 2nd compensation circuit configured to apply the second compensation value to the generated first code value to generate a second-stage compensated second code value.