Images are being generated at an ever-increasing rate by sources such as defence and civilian satellites, military reconnaissance surveillance flights, fingerprinting mug-shot-capturing devices, scientific experiments, biomedical imaging, home entertainment systems. For example, NASA's Earth Observing System will generate about 1 terabyte of image data per day when fully operational. A content-based retrieval (CBIR) system is required to effectively efficiently use information from these...

Natural Language Processing (NLP) is a research field where language in consideration processed to understand its syntactic, semantic, and sentimental aspects. The advancement the NLP area has helped solve problems domains such as Neural Machine Translation, Name Entity Recognition, Sentiment Analysis, Chatbots, name few. topic of broadly consists two main parts: representation input text (raw data) into numerical format (vectors or matrix) design models for processing data. This paper...

Similarity-based retrieval of images is an important task in many image database applications. A major class users' requests requires retrieving those the that are spatially similar to query image. We propose algorithm for computing spatial similarity between two symbolic images. a logical representation original where objects uniquely labeled with names. Spatial relationships represented as edges weighted graph referred spatial-orientation graph. then quantified terms number of, well extent...

Effective search and retrieval are enabling technologies for realizing the full potential of Web. The authors examine relevant issues, including methods representing document content. They also compare available tools suggest improving effectiveness.

Despite some key problems, big data could fundamentally change scientific research methodology and how businesses develop products provide services.

The advent of Big Data created a need for out-of-the-box horizontal scalability data management systems. This ushered in an array choices under the umbrella term NoSQL. In this paper, we provide taxonomy and unified perspective on NoSQL Using perspective, compare contrast various systems using multiple facets including system architecture, model, query language, client API, scalability, availability. We group current into seven broad categories: Key-Value, Table-type/Column, Document, Graph,...

Though the issues of data quality trace back their origin to early days computing, recent emergence Big Data has added more dimensions. Furthermore, given range applications, potential consequences bad can be for disastrous and widespread. This paper provides a perspective on in context. it also discusses integration that arise biological databases attendant issues.

Big data validation and system verification are crucial for ensuring the quality of big applications. However, a rigorous technique such tasks is yet to emerge. During past decade, we have developed called CMA investigating classification biological cells based on cell morphology which captured in diffraction images. includes collection scientific software tools, machine learning algorithms, large-scale image repository. In order ensure CMA, framework rigorously validating massive scale as...

Deep learning is an important technique for extracting value from big data. However, the effectiveness of deep requires large volumes high quality training In many cases, size data not enough effectively a classifier. Data augmentation widely adopted approach increasing amount But augmented may be questionable. Therefore, systematic evaluation critical. Furthermore, if noisy, it necessary to separate out noise automatically. this paper, we propose classifier automatically separating good...

Deep learning has been widely used for extracting values from big data. As many other machine algorithms, deep requires significant training Experiments have shown both the volume and quality of data can significantly impact effectiveness value extraction. In some cases, is not sufficiently large effectively a model. high enough to achieve optimal performance. Many approaches proposed augmenting mitigate deficiency. However, whether augmented are “fit purpose” still question. A framework...

Large language models (LLMs), consisting of billions and trillions parameters, have demonstrated exceptional ability in natural understanding (NLU) generation (NLG) tasks. Increases their numbers parameters model sizes resulted better performance accuracy. However, with such enormous incur significant computational costs resources, making them challenging to fine tune adapt a specific downstream task. Several parameter-efficient fine-tuning (PEFT) techniques been proposed address this issue....

Traditional lectures espousing software engineering principles hardly engage studentspsila attention due to the fact that students often view as mere academic concepts without a clear understanding of how they can be used in practice. Some issues contribute this perception include lack experience writing and large programs, opportunities for inspecting maintaining code written by others. To address these issues, we have worked on project whose overarching goal is teach subset basic using...

Big data requirements are motivating new database-management models that can process billions of requests per second, and established relational changing to keep pace. The authors provide practical tools for navigating this shifting product landscape finding candidate systems best fit a manager's application needs.

The articles in this special section focus on cognitive computing systems. Humans are arguably the most intelligent entities known universe; objective of is to understand and replicate essence human intelligence. Autonomous systems self-contained self-regulated that continuously evolve real time response changes their environment. Fundamental evolution learning development. Cognition basis for autonomous Human cognition refers processes enable humans perform both mundane specialized tasks....

A spatial similarity algorithm assesses the degree to which relationships among domain objects in a database image conform those specified query image. In this paper, we propose geometry-based structure for representing images and an associated algorithm. The proposed recognizes both translation, scale, rotation variants of image, generated by arbitrary composition transformations. has /spl Theta/(n log n) time complexity terms number common images. retrieval effectiveness is evaluated using...