Edward A. Fox – författare

Electronic Theses and Dissertations examines how electronic publication of theses and dissertations might enhance graduate education. This text clarifies the composition, evaluation, dissemination, and preservation of electronic theses and dissertations (ETDs), and provides a conceptual framework for the development of effective ETD programs. It identifies the main technical concerns related to the adoption of ETD initiatives and contains answers and methods that have proven effective in the longest-running library-group support effort for campus ETDs, making it the foremost guide to the latest innovations, practices, and policies in ETD production, distribution, and institutionalization.

Electronic Theses and Dissertations examines how electronic publication of theses and dissertations might enhance graduate education. This text clarifies the composition, evaluation, dissemination, and preservation of electronic theses and dissertations (ETDs), and provides a conceptual framework for the development of effective ETD programs. It identifies the main technical concerns related to the adoption of ETD initiatives and contains answers and methods that have proven effective in the longest-running library-group support effort for campus ETDs, making it the foremost guide to the latest innovations, practices, and policies in ETD production, distribution, and institutionalization.

Electronic Theses and Dissertations examines how electronic publication of theses and dissertations might enhance graduate education. This text clarifies the composition, evaluation, dissemination, and preservation of electronic theses and dissertations (ETDs), and provides a conceptual framework for the development of effective ETD programs. It id

Electronic Theses and Dissertations examines how electronic publication of theses and dissertations might enhance graduate education. This text clarifies the composition, evaluation, dissemination, and preservation of electronic theses and dissertations (ETDs), and provides a conceptual framework for the development of effective ETD programs. It id

Updated to include three new chapters on transformers, natural language understanding (NLU) with explainable AI, and dabbling with popular LLMs from Hugging Face and OpenAI

Key Features

Leverage ready-to-use recipes with the latest LLMs, including Mistral, Llama, and OpenAI modelsUse LLM-powered agents for custom tasks and real-world interactionsGain practical, in-depth knowledge of transformers and their role in implementing various NLP tasks with open-source and advanced LLMsPurchase of the print or Kindle book includes a free PDF eBook

Book Description

Harness the power of Natural Language Processing to overcome real-world text analysis challenges with this recipe-based roadmap written by two seasoned NLP experts with vast experience transforming various industries with their NLP prowess.You’ll be able to make the most of the latest NLP advancements, including large language models (LLMs), and leverage their capabilities through Hugging Face transformers. Through a series of hands-on recipes, you’ll master essential techniques such as extracting entities and visualizing text data. The authors will expertly guide you through building pipelines for sentiment analysis, topic modeling, and question-answering using popular libraries like spaCy, Gensim, and NLTK. You’ll also learn to implement RAG pipelines to draw out precise answers from a text corpus using LLMs.This second edition expands your skillset with new chapters on cutting-edge LLMs like GPT-4, Natural Language Understanding (NLU), and Explainable AI (XAI)—fostering trust and in your NLP models.By the end of this book, you''ll be equipped with the skills to apply advanced text processing techniques, use pre-trained transformer models, build custom NLP pipelines to extract valuable insights from text data to drive informed decision-making.

What you will learn

Understand fundamental NLP concepts along with their applications using examples in PythonClassify text quickly and accurately with rule-based and supervised methodsTrain NER models and perform sentiment analysis to identify entities and emotions in textExplore topic modeling and text visualization to reveal themes and relationships within textLeverage Hugging Face and OpenAI LLMs to perform advanced NLP tasksUse question-answering techniques to handle both open and closed domainsApply XAI techniques to better understand your model predictions

Who this book is for

This updated edition of the Python Natural Language Processing Cookbook is for data scientists, machine learning engineers, and developers with a background in Python. Whether you’re looking to learn NLP techniques, extract valuable insights from textual data, or create foundational applications, this book will equip you with basic to intermediate skills. No prior NLP knowledge is necessary to get started. All you need is familiarity with basic programming principles. For seasoned developers, the updated sections offer the latest on transformers, explainable AI, and Generative AI with LLMs.

Updated to include three new chapters on transformers, natural language understanding (NLU) with explainable AI, and dabbling with popular LLMs from Hugging Face and OpenAIKey FeaturesLeverage ready-to-use recipes with the latest LLMs, including Mistral, Llama, and OpenAI modelsUse LLM-powered agents for custom tasks and real-world interactionsGain practical, in-depth knowledge of transformers and their role in implementing various NLP tasks with open-source and advanced LLMsPurchase of the print or Kindle book includes a free PDF eBookBook DescriptionHarness the power of Natural Language Processing (NLP) to overcome real-world text analysis challenges with this recipe-based roadmap written by two seasoned NLP experts with vast experience transforming various industries with their NLP prowess.You’ll be able to make the most of the latest NLP advancements, including large language models (LLMs), and leverage their capabilities through Hugging Face transformers. Through a series of hands-on recipes, you’ll master essential techniques such as extracting entities and visualizing text data. The authors will expertly guide you through building pipelines for sentiment analysis, topic modeling, and question-answering using popular libraries like spaCy, Gensim, and NLTK. You’ll also learn to implement RAG pipelines to draw out precise answers from a text corpus using LLMs.This second edition expands your skillset with new chapters on cutting-edge LLMs like GPT-4, Natural Language Understanding (NLU), and Explainable AI (XAI)—fostering trust in your NLP models.By the end of this book, you'll be equipped with the skills to apply advanced text processing techniques, use pre-trained transformer models, build custom NLP pipelines to extract valuable insights from text data to drive informed decision-making.What you will learnUnderstand fundamental NLP concepts along with their applications using examples in PythonClassify text quickly and accurately with rule-based and supervised methodsTrain NER models and perform sentiment analysis to identify entities and emotions in textExplore topic modeling and text visualization to reveal themes and relationships within textLeverage Hugging Face and OpenAI LLMs to perform advanced NLP tasksUse question-answering techniques to handle both open and closed domainsApply XAI techniques to better understand your model predictionsWho this book is forThis updated edition of the Python Natural Language Processing Cookbook is for data scientists, machine learning engineers, and developers with a background in Python. Whether you’re looking to learn NLP techniques, extract valuable insights from textual data, or create foundational applications, this book will equip you with basic to intermediate skills. No prior NLP knowledge is necessary to get started. All you need is familiarity with basic programming principles. For seasoned developers, the updated sections offer the latest on transformers, explainable AI, and Generative AI with LLMs.

This is the second book based on the 5S (Societies, Scenarios, Spaces, Structures, Streams) approach to digital libraries (DLs). Leveraging the first volume, on Theoretical Foundations, we focus on the key issues of evaluation and integration. These cross-cutting issues serve as a bridge for those interested in DLs, connecting the introduction and formal discussion in the first book, with the coverage of key technologies in the third book, and of illustrative applications in the fourth book. These two topics have central importance in the DL field, allowing it to be treated scientifically as well as practically. In the scholarly world, we only really understand something if we know how to measure and evaluate it. In the Internet era of distributed information systems, we only can be practical at scale if we integrate across both systems and their associated content. Evaluation of DLs must take place atmultiple levels,so we can address the different entities and their associated measures. Thus, for digital objects, we assess accessibility, pertinence, preservability, relevance, significance, similarity, and timeliness. Other measures are specific to higher-level constructs like metadata, collections, catalogs, repositories, and services.We tie these together through a case study of the 5SQual tool, which we designed and implemented to perform an automatic quantitative evaluation of DLs. Thus, across the Information Life Cycle, we describe metrics and software useful to assess the quality of DLs, and demonstrate utility with regard to representative application areas: archaeology and education. Though integration has been a challenge since the earliest work on DLs, we provide the first comprehensive 5S-based formal description of the DL integration problem, cast in the context of related work. Since archaeology is a fundamentally distributed enterprise, we describe ETANADL, for integrating Near Eastern Archeology sites and information. Thus, we show how 5S-based modeling can lead to integrated services and content. While the first book adopts a minimalist and formal approach to DLs, and provides a systematic and functional method to design and implement DL exploring services, here we broaden to practical DLs with richer metamodels, demonstrating the power of 5S for integration and evaluation.

Digital libraries (DLs) have evolved since their launch in 1991 into an important type of information system, with widespread application. This volume advances that trend further by describing new research and development in the DL field that builds upon the 5S (Societies, Scenarios, Spaces, Structures, Streams) framework, which is discussed in three other DL volumes in this series.While the 5S framework may be used to describe many types of information systems, and is likely to have even broader utility and appeal, we focus here on digital libraries. Drawing upon six (Akbar, Kozievitch, Leidig, Li, Murthy, Park) completed and two (Chen, Fouh) in-process dissertations, as well as the efforts of collaborating researchers, and scores of related publications, presentations, tutorials, and reports, this book demonstrates the applicability of 5S in five digital library application areas, that also have importance in the context of the WWW, Web 2.0, and innovative information systems. By integrating surveys of the state-of-the-art, newresearch, connections with formalization, case studies, and exercises/projects, this book can serve as a textbook for those interested in computing, information, and/or library science. Chapter 1 focuses on images, explaining how they connect with information retrieval, in the context of CBIR systems. Chapter 2 gives two case studies of DLs used in education, which is one of the most common applications of digital libraries. Chapter 3 covers social networks, which are at the heart of work onWeb 2.0, explaining the construction and use of deduced graphs, that can enhance retrieval and recommendation. Chapter 4 demonstrates the value of DLs in eScience, focusing, in particular, on cyber-infrastructure for simulation. Chapter 5 surveys geospatial information in DLs, with a case study on geocoding. Given this rich content, we trust that any interested in digital libraries, or in related systems, will find this volume to be motivating, intellectually satisfying, and useful. We hope it will help move digital libraries forward into a science as well as a practice. We hope it will help build community that will address the needs of the next generation of DLs.

Digital libraries (DLs) have introduced new technologies, as well as leveraging, enhancing, and integrating related technologies, since the early 1990s. These efforts have been enriched through a formal approach, e.g., the 5S (Societies, Scenarios, Spaces, Structures, Streams) framework, which is discussed in two earlier volumes in this series. This volume should help advance work not only in DLs, but also in the WWW and other information systems. Drawing upon four (Kozievitch, Murthy, Park, Yang) completed and three (Elsherbiny, Farag, Srinivasan) in-process dissertations, as well as the efforts of collaborating researchers and scores of related publications, presentations, tutorials, and reports, this book should advance the DL field with regard to at least six key technologies. By integrating surveys of the state-of-the-art, new research, connections with formalization, case studies, and exercises/projects, this book can serve as a computing or information science textbook. It can support studies in cyber-security, document management, hypertext/hypermedia, IR, knowledge management, LIS, multimedia, and machine learning. Chapter 1, with a case study on fingerprint collections, focuses on complex (composite, compound) objects, connecting DL and related work on buckets, DCC, and OAI-ORE. Chapter 2, discussing annotations, as in hypertext/hypermedia, emphasizes parts of documents, including images as well as text, managing superimposed information. The SuperIDR system, and prototype efforts with Flickr, should motivate further development and standardization related to annotation, which would benefit all DL and WWW users. Chapter 3, on ontologies, explains how they help with browsing, query expansion, focused crawling, and classification. This chapter connects DLs with the Semantic Web, and uses CTRnet as an example. Chapter 4, on (hierarchical) classification, leverages LIS theory, as well as machine learning, and is important for DLs as well as the WWW. Chapter 5,on extraction from text, covers document segmentation, as well as how to construct a database from heterogeneous collections of references (from ETDs); i.e., converting strings to canonical forms. Chapter 6 surveys the security approaches used in information systems, and explains how those approaches can apply to digital libraries which are not fully open. Given this rich content, those interested in DLs will be able to find solutions to key problems, using the right technologies and methods. We hope this book will help show how formal approaches can enhance the development of suitable technologies and how they can be better integrated with DLs and other information systems.

This is the second book based on the 5S (Societies, Scenarios, Spaces, Structures, Streams) approach to digital libraries (DLs). Leveraging the first volume, on Theoretical Foundations, we focus on the key issues of evaluation and integration. These cross-cutting issues serve as a bridge for those interested in DLs, connecting the introduction and formal discussion in the first book, with the coverage of key technologies in the third book, and of illustrative applications in the fourth book. These two topics have central importance in the DL field, allowing it to be treated scientifically as well as practically. In the scholarly world, we only really understand something if we know how to measure and evaluate it. In the Internet era of distributed information systems, we only can be practical at scale if we integrate across both systems and their associated content. Evaluation of DLs must take place atmultiple levels,so we can address the different entities and their associated measures. Thus, for digital objects, we assess accessibility, pertinence, preservability, relevance, significance, similarity, and timeliness. Other measures are specific to higher-level constructs like metadata, collections, catalogs, repositories, and services.We tie these together through a case study of the 5SQual tool, which we designed and implemented to perform an automatic quantitative evaluation of DLs. Thus, across the Information Life Cycle, we describe metrics and software useful to assess the quality of DLs, and demonstrate utility with regard to representative application areas: archaeology and education. Though integration has been a challenge since the earliest work on DLs, we provide the first comprehensive 5S-based formal description of the DL integration problem, cast in the context of related work. Since archaeology is a fundamentally distributed enterprise, we describe ETANADL, for integrating Near Eastern Archeology sites and information. Thus, we show how 5S-based modeling can lead to integrated services and content. While the first book adopts a minimalist and formal approach to DLs, and provides a systematic and functional method to design and implement DL exploring services, here we broaden to practical DLs with richer metamodels, demonstrating the power of 5S for integration and evaluation.

Digital libraries (DLs) have evolved since their launch in 1991 into an important type of information system, with widespread application. This volume advances that trend further by describing new research and development in the DL field that builds upon the 5S (Societies, Scenarios, Spaces, Structures, Streams) framework, which is discussed in three other DL volumes in this series.While the 5S framework may be used to describe many types of information systems, and is likely to have even broader utility and appeal, we focus here on digital libraries. Drawing upon six (Akbar, Kozievitch, Leidig, Li, Murthy, Park) completed and two (Chen, Fouh) in-process dissertations, as well as the efforts of collaborating researchers, and scores of related publications, presentations, tutorials, and reports, this book demonstrates the applicability of 5S in five digital library application areas, that also have importance in the context of the WWW, Web 2.0, and innovative information systems. By integrating surveys of the state-of-the-art, newresearch, connections with formalization, case studies, and exercises/projects, this book can serve as a textbook for those interested in computing, information, and/or library science. Chapter 1 focuses on images, explaining how they connect with information retrieval, in the context of CBIR systems. Chapter 2 gives two case studies of DLs used in education, which is one of the most common applications of digital libraries. Chapter 3 covers social networks, which are at the heart of work onWeb 2.0, explaining the construction and use of deduced graphs, that can enhance retrieval and recommendation. Chapter 4 demonstrates the value of DLs in eScience, focusing, in particular, on cyber-infrastructure for simulation. Chapter 5 surveys geospatial information in DLs, with a case study on geocoding. Given this rich content, we trust that any interested in digital libraries, or in related systems, will find this volume to be motivating, intellectually satisfying, and useful. We hope it will help move digital libraries forward into a science as well as a practice. We hope it will help build community that will address the needs of the next generation of DLs.

Digital libraries (DLs) have introduced new technologies, as well as leveraging, enhancing, and integrating related technologies, since the early 1990s. These efforts have been enriched through a formal approach, e.g., the 5S (Societies, Scenarios, Spaces, Structures, Streams) framework, which is discussed in two earlier volumes in this series. This volume should help advance work not only in DLs, but also in the WWW and other information systems. Drawing upon four (Kozievitch, Murthy, Park, Yang) completed and three (Elsherbiny, Farag, Srinivasan) in-process dissertations, as well as the efforts of collaborating researchers and scores of related publications, presentations, tutorials, and reports, this book should advance the DL field with regard to at least six key technologies. By integrating surveys of the state-of-the-art, new research, connections with formalization, case studies, and exercises/projects, this book can serve as a computing or information science textbook. It can support studies in cyber-security, document management, hypertext/hypermedia, IR, knowledge management, LIS, multimedia, and machine learning. Chapter 1, with a case study on fingerprint collections, focuses on complex (composite, compound) objects, connecting DL and related work on buckets, DCC, and OAI-ORE. Chapter 2, discussing annotations, as in hypertext/hypermedia, emphasizes parts of documents, including images as well as text, managing superimposed information. The SuperIDR system, and prototype efforts with Flickr, should motivate further development and standardization related to annotation, which would benefit all DL and WWW users. Chapter 3, on ontologies, explains how they help with browsing, query expansion, focused crawling, and classification. This chapter connects DLs with the Semantic Web, and uses CTRnet as an example. Chapter 4, on (hierarchical) classification, leverages LIS theory, as well as machine learning, and is important for DLs as well as the WWW. Chapter 5,on extraction from text, covers document segmentation, as well as how to construct a database from heterogeneous collections of references (from ETDs); i.e., converting strings to canonical forms. Chapter 6 surveys the security approaches used in information systems, and explains how those approaches can apply to digital libraries which are not fully open. Given this rich content, those interested in DLs will be able to find solutions to key problems, using the right technologies and methods. We hope this book will help show how formal approaches can enhance the development of suitable technologies and how they can be better integrated with DLs and other information systems.

The International Conference on Asian Digital Libraries 2005 (ICADL 2005), held in Bangkok, Thailand, 12-15 December 2005, was the 8th in a series of annual internationalconferences organizedby digital libraryresearchersin Asia. ICADL was set up to facilitate and stimulate the exchange of digital library research, information, and technology in the Asian region while also functi- ing as a sister conference to its North American and European counterparts, i.e., JCDL and ECDL. The theme of ICADL 2005 was "Digital Libraries: - plementing Strategies and Sharing Experiences," covering strategies, implem- tation experiences, systems, techniques, management, and applications. ICADL 2005 provided a forum for sharing experiences and exchanging research results, innovative ideas, and state-of-the-art developments among researchers, edu- tors, practitioners, and policy makers from a variety of disciplines such as c- puter science,libraryand informationscience,archivalandmuseum studies,and knowledge management. ICADL 2005 also marked the special occasion of the 50th birthday anniv- sary of Her Royal Highness Princess Maha Chakri Sirindhorn, Patron of the Thai Library Association.ICADL 2005 received 164 paper submissions from 17 countries, including a good number from outside the Asian region. From these, 40 full papers, 15 short papers, and 15 poster papers were accepted and are included in these proceedings.Eachpaper wasreviewedby the ProgramCommittee membersand additional co-reviewers. The technical program comprised one day of tutorials, followed by keynote and invited speeches, paper and poster sessions as well as a post-conference workshop on metadata and Dublin Core.