A5032772753- Advanced Image and Video Retrieval Techniques
- Multimodal Machine Learning Applications
- Face recognition and analysis
- Advancements in Photolithography Techniques
- Advanced Neural Network Applications
- Video Analysis and Summarization
- Optical Coatings and Gratings
- Software Engineering Research
- Advanced Surface Polishing Techniques
- Electron and X-Ray Spectroscopy Techniques
- Face and Expression Recognition
- AI in cancer detection
- Image Processing Techniques and Applications
- Nuclear Physics and Applications
- Copper Interconnects and Reliability
- Advanced X-ray Imaging Techniques
- Software System Performance and Reliability
- Integrated Circuits and Semiconductor Failure Analysis
- Robotics and Sensor-Based Localization
- CCD and CMOS Imaging Sensors
- Biometric Identification and Security
- Surface Roughness and Optical Measurements
- Nuclear reactor physics and engineering
- Domain Adaptation and Few-Shot Learning
- Software Reliability and Analysis Research
Lawrence Berkeley National Laboratory
2016-2024
Research Applications (United States)
2019-2024
Delhi Technological University
2022-2024
National Institute of Technology Andhra Pradesh
2024
Meenakshi Academy of Higher Education and Research
2024
University of Canberra
2019-2023
University of the South Pacific
2016-2023
Alive Hospice
2020
Guru Kashi University
2016
International Institute of Information Technology Bangalore
2013
This paper investigates the potential of large language models (LLMs) to enhance programming productivity through domain-specific code generation. By integrating domain expertise with advanced machine learning techniques, our approach tailors LLMs generate that aligns closely specialized application requirements. The study outlines a systematic framework for fine-tuning using datasets, enabling automated synthesis accurate and efficient code. Experimental results demonstrate this targeted...
Silicon shortages and supply chain constraints have emerged as critical challenges for organizations deploying artificial intelligence hardware in today’s rapidly evolving technological landscape. This abstract examines the multifaceted impacts of these through lens a program manager tasked with overseeing complex deployment projects. The persistent global deficit semiconductor has created ripple effects throughout industry, resulting extended lead times, increased costs, uncertainty...
In today’s dynamic financial landscape, the integration of quality engineering into Agile and DevOps frameworks has emerged as a transformative approach for claims processing. This research explores paradigm “shifting left” — an early intervention strategy in software development lifecycle that infuses robust checks at every stage development. By incorporating methodologies practices, organizations can accelerate testing, streamline collaboration, ensure is intrinsic component from design to...
Several recent reports have demonstrated that fluorinated analogues of donor/acceptor copolymers surpass nonfluorinated counterparts in terms performance electronic devices. Using a copolymer series consisting fluorinated, partially and benzotriazole, we confirm the addition fluorine substituents beneficially impacts charge transport polymer semiconductors. Transistor measurements factor 5 increase carrier mobilities with degree fluorination backbone. Furthermore, grazing-incidence X-ray...
Many Attention-based deep architectures have been widely adopted for image captioning. Mostly force visual attention to be active word generation. However, spatial region specific generation is not desired always. In this paper, address issue, an adaptive mechanism-based automated Image caption framework proposed. Our model uses Inception-V3 extract various global features and the module helps decide whether attend (and if so, which regions) or sentinel maps. Further, at decoding end, a...
As lithographic nanostructures shrink to critical dimensions less than 10 nm, the semiconductor industry searches for nondestructive metrology techniques extract three-dimensional structural information with subnanometer precision. This study develops a modeling algorithm that extracts unique, complex line profile from x-ray scattering measurements collected in grazing-incidence configuration. represents an important step toward demonstrating utility of as viable tool measuring processing, and more.
Extreme ultraviolet (EUV) lithography is one of the most promising printing techniques for high-volume semiconductor manufacturing at 14-nm half-pitch device node and beyond. However, key challenges around EUV photoresist materials, such as exposure-dose sensitivity or line-width roughness, continue to impede its full adoption into industrial nanofab facilities. Metrology tools are required address these by helping assess impact materials' properties processing conditions along different...
Breast cancer is the most common in women. Classification of cancer/non-cancer patients with clinical records requires high sensitivity and specificity for an acceptable diagnosis test. The state-of-the-art classification model-convolutional neural network (CNN), however, cannot be used such kind tabular data that are represented 1-D format. CNN has been designed to work on a set 2-D matrices whose elements show some correlation neighboring as image data. Conversely, examples vectors-apart...
Image captioning develops a relationship between visual and text information to generate sequence of words as captions. Transformers perform machine translation language comprehension together using encoder decoder structure. With the aim building lightweight production deployment friendly model, we present Lightweight Transformer with GRU integrated for Captioning. In presented number encoders decoders in standard architecture are reduced single decoder. Also, Multi-level rich features from...
Grazing-incidence small-angle X-ray scattering (GISAXS) is an important technique in the characterization of samples at nanometre scale. A key aspect GISAXS data analysis accurate simulation to match measurement. The distorted-wave Born approximation (DWBA) a widely used model for patterns. For certain classes sample such as nanostructures embedded thin films, where electric field intensity variation significant relative size structures, multi-slice DWBA theory more than conventional method....
Abstract Introducing variation in the training dataset through data augmentation has been a popular technique to make Convolutional Neural Networks (CNNs) spatially invariant but leads increased volume and computation cost. Instead of augmentation, feature maps is proposed introduce variations features extracted by CNN. To achieve this, rotation transformer layer called Rotation Invariance Transformer (RiT) developed, which applies transformation augment CNN features. The RiT can be used...
Developing computational algorithms to model the biological vision system has challenged researchers in computer field for several decades. As a result, state-of-the-art such as Convolutional Neural Network (CNN) have emerged image classification and recognition tasks with promising results. CNNs however remain view-specific, producing good results when variation between test train data is small. Making learn invariant features effectively recognise objects that undergo appearance changes...
IEC 61131-3 is an open standard that provides guidelines for programmable logic controllers and control systems. The defines specifications system programming languages, both textual graphical. A application developed using these languages can go through several changes over the course of project's life-cycle. While differences in programs be detected traditional text difference algorithms, there exist no effective means tracking at a granular, semantic level graphical programs. In this...
With the advent of high brightness sources and fast detectors, there is a possibility for combining X-ray acquisition with high-speed data treatment to reach timescale an effective in-line characterization method. We will highlight two recent developments using Small Angle Scattering on nanoscale etched patterns: first inclusion CD-SAXS tool, allowing simulations reconstruct form-factor, inside Xi-cam framework; second development performance Grazing Incidence approach shape line profile....
A large body of physiological findings has suggested the vision system understands a scene in terms its local features such as lines and curves.A highly notable computer algorithm developed that models behaviour is Convolutional Neural Network (CNN).Whilst recognising an object various scales remains trivial for human system, CNNs struggle to achieve same behaviour.Recent are suggesting two new paradigms.Firstly, visual uses both global recognition function.Secondly, brain distributed...