Natural Language Analytics with Generative Large-Language Models: A Practical Approach with Ollama and Open-Source Llms
暫譯: 使用生成大型語言模型的自然語言分析：Ollama 與開源 LLM 的實用方法
Marcondes, Francisco S., Gala, Adelino, Magalhães, Renata
相關主題
商品描述
This book explores the application of generative Large Language Models (LLMs) for extracting and analyzing data from natural language artefacts. Unlike traditional uses of LLMs, such as translation and summarization, this book focuses on utilizing these models to convert unstructured text into data that can be processed through the data science pipeline to generate actionable insights.
The content is designed for professionals in diverse fields including cognitive science, linguistics, management, and information systems. It combines insights from both industry and academia to provide a comprehensive understanding of how LLMs can be effectively used for natural language analytics (NLA). The book details practical methodologies for implementing LLMs locally using open-source tools, ensuring data privacy and feasibility without the need for expensive infrastructure.
Key topics include interpretant, mindset and cultural analysis, emphasizing the use of LLMs to derive soft data--qualitative information crucial for nuanced decision-making. The text also outlines the technical aspects of LLMs, including their architecture, token embeddings, and the differences between encoder-based and decoder-based models. By providing a case study and practical examples, the authors show how LLMs can be used to meet various analytical needs, making this book a valuable resource for anyone looking to integrate advanced natural language processing techniques into their data analysis workflows.
商品描述(中文翻譯)
這本書探討了生成式大型語言模型（Large Language Models, LLMs）在從自然語言文獻中提取和分析數據的應用。與傳統的LLMs使用方式（如翻譯和摘要）不同，本書專注於利用這些模型將非結構化文本轉換為可以通過數據科學流程處理的數據，以生成可行的見解。
內容旨在為來自不同領域的專業人士提供，包括認知科學、語言學、管理學和資訊系統。它結合了來自產業和學術界的見解，以提供對LLMs如何有效用於自然語言分析（Natural Language Analytics, NLA）的全面理解。本書詳細介紹了使用開源工具在本地實施LLMs的實用方法，確保數據隱私和可行性，而無需昂貴的基礎設施。
主要主題包括解釋者、心態和文化分析，強調使用LLMs來推導軟數據——對於細緻的決策至關重要的定性信息。文本還概述了LLMs的技術方面，包括其架構、標記嵌入以及基於編碼器和基於解碼器模型之間的差異。通過提供案例研究和實際範例，作者展示了如何使用LLMs來滿足各種分析需求，使本書成為任何希望將先進自然語言處理技術整合到其數據分析工作流程中的人的寶貴資源。
作者簡介
Francisco S. Marcondes is a research collaborator at the Synthetic Intelligence Laboratory (ISLab) at the ALGORITMI Centre at the University of Minho (Braga, Portugal). He has teaching experience in Brazil (including the Pontifical Catholic University of São Paulo) and Portugal (including the University of Minho). He has a degree in Information Systems (2004), a master's degree in Intelligence Technologies and Digital Design (2015) and specialised in Pedagogical Training for Professional Education (2017). He is currently studying for a PhD in Artificial Intelligence and Natural Language Processing.
Adelino Gala is a researcher specializing in Natural Language Processing (NLP), Cognitive Science, Semiotics, and the application of Large and Small Language Models. He holds a Ph.D. in Technologies of Intelligence and Digital Design from the Pontifical Catholic University of São Paulo, Brazil, where his work emphasized learning and cognitive semiotics, and completed his postdoctoral studies at the University of Aveiro, Portugal, focusing on the integration of new digital technologies in communication practices.
Renata Magalhães is a PhD student in Biomedical Engineering at the University of Minho, where she also completed her master's degree (2024) in Digital Humanities. Her thesis explored the topic of emotion detection for school failure prevention. Her research focuses on Natural Language Processing and Artificial Intelligence.
Fernando Perez de Britto has extensive experience in analytics, artificial intelligence and decision support solutions for private organizations, with a master's degree in the area. Reference in the UN concept of Coherence. Founder and CEO of Investment 4 Impact, an investment holding focused on innovation, high technology and impact originating from AI Systems Research (AISR), founded in 2002. He is also responsible for the holding's socio-environmental initiative: "Making Smart Cities". Co-Chair Emeritus of the UNDRR Stakeholder Engagement Mechanism (SEM). Former Vice-Chair of the Global Council of the UNDRR Private Sector Alliance for Disaster Resilient Societies (ARISE) (2020-2023), led the ARISE delegation in the Mid-Term Review of the Implementation of the Sendai Framework for Disaster Risk Reduction 2015- 2030 held in NY (2023), UNDRR ARISE Global Board Member (2017-2023), UNDRR Stakeholder Engagement Mechanism (SEM) Co-Chair (2019-2021) and UNDRR ARISE Advisory Board Member (2015-2017). He received the UN Sasakawa Prize in 2019. Co-author of the book "Smart Cities: why, for whom?" (2016). Master in Intelligence Technologies and Digital Design (Cognitive Sciences) from the Pontifical Catholic University of São Paulo (PUC/SP), specialization in Business and Administration from Fundação Getúlio Vargas de São Paulo (FGV-EAESP) and bachelor's degree in Computer Science from Pontifícia Catholic University of São Paulo (PUC/SP).
Dalila Durães is Assistant Professor of Computer Sciences at the Department of Informatics, University of Minho, Braga, Portugal. Researcher at the ALGORITMI Centre and LASI (Intelligent Systems Associate Laboratory) at the group ISlab - Synthetic Intelligence. She graduated in Electronic Engineering and Informatics in 1995 and, in 2004, completed her Master's Degree in Industrial Electronic Engineering in Automation and Robotics. She holds two international PhDs: one in Educational Sciences, in Teachers, Curricula and Educational Institutions, from the University of Granada, Spain, completed in 2012, and another in Artificial Intelligence by the Polytechnic University of Madrid, Spain, completed in 2018. She started her career on 2015 developing scientific research in the field of Intelligent Systems/Artificial Intelligence (AI), namely in Human-Computer Interaction, Machine and Deep Learning, Behaviour Analysis, with particular attention to the detection of violence, sentiment analysis, intelligent tutoring and recognition of human actions. Her interests, in the last years, were absorbed by the broad, yet closely related, concepts of Intelligent Environments, Data Fusion, Digital Assistants, Artificial Intelligent applying to Education, and the incorporation of AI methods and techniques in these fields. Its main research objective is to make systems smarter, sensible, and reliable. She is the author and co-author of over 80 book chapters, journal papers, conference and workshop papers, and books (with peer review). She is also a member of the editorial board of several international journals. During the past few years, she has served as an expert/reviewer for several conferences and journals. She has participated in several research projects sponsored by Portuguese and European public and private Institutions. She has supervised several Ph.D. and 15 M.Sc. students. She is Vice-chair of the Portuguese Association for Artificial Intelligence (APPIA) and Vice-chair of IEEE, Computational Intelligence Society, Portuguese Chapter. She has 2 Awards and Distinctions - Recognition by the Scientific Community: Best Paper in 2nd International Conference on sustainable Smart Cities and Territories International 2023, and Best Paper Application Award in 16th International Conference on Distributed Computing and Artificial Intelligence 2019. International research collaboration with different universities, resulting on the production several papers for journals with high impact factor, several conferences with peer-reviews and some book chapter.
Paulo Novais is Full Professor in the Department of Computer Science and researcher at the ALGORITMI Centre, in the School of Engineering at the University of Minho. Coordinator of the Associate Laboratory of Intelligent Systems (LASI). His main research goal is to design systems that are more sensitive to human presence, a little smarter and more reliable. He is Director of the PhD Programme in Informatics, co-founder and Deputy Director of the Master's Degree in Law and Informatics at the University of Minho and served as President of the Portuguese Association for Artificial Intelligence (APPIA) between 2016 and 2019 and current Chairman of the Supervisory Board. Portuguese representative at IFIP (International Federation for Information Processing) - Artificial Intelligence Technical Group. He is chair of the Computational Intelligence Society Portuguese Chapter and Senior member of the IEEE (Institute of Electrical and Electronics Engineers), a member of the executive committee of IBERAMIA (IberoAmerican Society of Artificial Intelligence) and has worked as an expert for various institutions such as the European Commission, FCT, A3ES, ANI and the Calouste Gulbenkian Foundation, among others.
作者簡介(中文翻譯)
Francisco S. Marcondes 是葡萄牙米尼奧大學 (University of Minho) ALGORITMI 中心的合成智慧實驗室 (ISLab) 研究合作者。他在巴西（包括聖保羅教皇天主教大學）和葡萄牙（包括米尼奧大學）擁有教學經驗。他於 2004 年獲得資訊系統學位，2015 年獲得智慧技術與數位設計碩士學位，並於 2017 年專攻職業教育的教學訓練。目前，他正在攻讀人工智慧和自然語言處理的博士學位。
Adelino Gala 是一位專注於自然語言處理 (NLP)、認知科學、符號學以及大型和小型語言模型應用的研究員。他擁有巴西聖保羅教皇天主教大學的智慧技術與數位設計博士學位，研究重點在於學習和認知符號學，並在葡萄牙阿維羅大學完成了博士後研究，專注於新數位技術在溝通實踐中的整合。
Renata Magalhães 是米尼奧大學生物醫學工程的博士生，她也在該校完成了數位人文學的碩士學位（2024）。她的論文探討了情感檢測在預防學校失敗中的應用。她的研究重點是自然語言處理和人工智慧。
Fernando Perez de Britto 在分析、人工智慧和決策支持解決方案方面擁有豐富的經驗，並在該領域獲得碩士學位。他是聯合國一致性概念的參考人物。Investment 4 Impact 的創始人兼首席執行官，該投資控股公司專注於創新、高科技和源自人工智慧系統研究 (AISR) 的影響，成立於 2002 年。他還負責該控股公司的社會環境倡議：“打造智慧城市”。他是聯合國減災辦公室 (UNDRR) 利益相關者參與機制 (SEM) 的名譽共同主席。曾任聯合國減災辦公室私營部門聯盟 (ARISE) 全球委員會副主席（2020-2023），在 2023 年於紐約舉行的 2015-2030 年仙台減災框架中期評估中領導 ARISE 代表團，並於 2017-2023 年擔任 UNDRR ARISE 全球董事會成員，2019-2021 年擔任 UNDRR 利益相關者參與機制 (SEM) 共同主席，2015-2017 年擔任 UNDRR ARISE 諮詢委員會成員。他於 2019 年獲得聯合國佐佐木獎。是書籍《智慧城市：為什麼，為誰？》（2016）的共同作者。擁有聖保羅教皇天主教大學 (PUC/SP) 的智慧技術與數位設計（認知科學）碩士學位，聖保羅蓋圖利奧·瓦爾加斯基金會 (FGV-EAESP) 的商業與管理專業，及聖保羅教皇天主教大學 (PUC/SP) 的計算機科學學士學位。
Dalila Durães 是葡萄牙米尼奧大學資訊系的助理教授，並在 ALGORITMI 中心和 LASI（智能系統協作實驗室）擔任研究員，隸屬於 ISlab - 合成智慧小組。她於 1995 年獲得電子工程與資訊學學位，並於 2004 年完成工業電子工程（自動化與機器人）碩士學位。她擁有兩個國際博士學位：一個是西班牙格拉納達大學的教育科學博士，專注於教師、課程和教育機構，於 2012 年完成；另一個是西班牙馬德里理工大學的人工智慧博士，於 2018 年完成。她於 2015 年開始在智能系統/人工智慧 (AI) 領域開展科學研究，特別是在人機互動、機器學習和深度學習、行為分析等方面，特別關注暴力檢測、情感分析、智能輔導和人類行為識別。近幾年來，她的興趣集中在智能環境、數據融合、數位助理、應用於教育的人工智慧以及在這些領域中融入人工智慧方法和技術。她的主要研究目標是使系統更智能、敏感和可靠。她是超過 80 篇書籍章節、期刊論文、會議和研討會論文及書籍（經過同行評審）的作者和合著者。她還是幾本國際期刊的編輯委員會成員。在過去幾年中，她擔任了多個會議和期刊的專家/審稿人。她參與了多個由葡萄牙和歐洲公共及私人機構贊助的研究項目，並指導了多位博士生和 15 位碩士生。她是葡萄牙人工智慧協會 (APPIA) 的副主席，以及 IEEE 計算智能學會葡萄牙分會的副主席。她獲得了 2 項獎項和榮譽 - 科學界的認可：2023 年第二屆可持續智慧城市與領域國際會議最佳論文獎，以及 2019 年第十六屆分散式計算與人工智慧國際會議最佳應用論文獎。與不同大學的國際研究合作，產出多篇高影響因子的期刊論文、數個經過同行評審的會議論文及一些書籍章節。
Paulo Novais 是米尼奧大學計算機科學系的全職教授，並在 ALGORITMI 中心擔任研究員。他是智能系統協作實驗室 (LASI) 的負責人。他的主要研究目標是設計對人類存在更敏感、更智能和更可靠的系統。他是資訊學博士課程的主任，米尼奧大學法律與資訊碩士學位的共同創始人及副主任，並於 2016 至 2019 年擔任葡萄牙人工智慧協會 (APPIA) 的會長，目前是監事會的主席。作為 IFIP（國際資訊處理聯合會）人工智慧技術小組的葡萄牙代表，他是計算智能學會葡萄牙分會的主席，也是 IEEE（電氣和電子工程師學會）的資深會員，並擔任 IBERAMIA（伊比利亞美洲人工智慧學會）執行委員會成員，曾為歐洲委員會、FCT、A3ES、ANI 和卡盧斯特·古爾本基安基金會等多個機構擔任專家。
