商品描述
This book presents the first unified, practical framework for continuous-time series analysis using state-of-the-art neural architectures. Moving beyond traditional discrete-time methods, it directly addresses real-world challenges such as irregular sampling, asynchronous observations, and hidden system dynamics through Neural ODEs, SDEs, and CDEs. Covering both foundational and advanced models -- RNNs, Transformers, graph networks, and emerging quantum-hybrid approaches -- the book bridges classical time-series theory with modern deep learning. It emphasizes probabilistic forecasting, uncertainty quantification, and cutting-edge generative techniques, including diffusion models and VAEs, equipping readers with tools for robust, interpretable predictions. Recent Trends in Modelling the Continuous Time Series using Deep Learning tackles core issues such as long-range dependencies, multivariate interactions, dimensionality reduction, and spatiotemporal coherence, while providing structured evaluation frameworks and benchmarking protocols tailored to continuous-time settings. Through rich case studies in healthcare (EHR analytics, wearable monitoring), finance (volatility forecasting, high-frequency trading), and IoT systems (sensor fusion, predictive maintenance), the book demonstrates how continuous-time models enable personalized insights, constraint-aware learning, and more reliable decision-making. Designed for researchers, engineers, and practitioners, this book is a definitive resource for applying continuous-time neural methods to complex, real-world environments.
商品描述(中文翻譯)
本書提出了首個統一的、實用的連續時間序列分析框架，使用最先進的神經架構。超越傳統的離散時間方法，它直接解決了現實世界中的挑戰，例如不規則取樣、非同步觀測和隱藏系統動態，通過神經常微分方程（Neural ODEs）、隨機微分方程（SDEs）和連續時間動態系統（CDEs）。
本書涵蓋了基礎和進階模型——遞迴神經網絡（RNNs）、變壓器（Transformers）、圖形網絡以及新興的量子混合方法——將經典時間序列理論與現代深度學習相結合。它強調概率預測、不確定性量化和尖端生成技術，包括擴散模型（diffusion models）和變分自編碼器（VAEs），為讀者提供穩健且可解釋的預測工具。
《使用深度學習建模連續時間序列的最新趨勢》探討了核心問題，如長期依賴性、多變量互動、降維和時空一致性，同時提供針對連續時間環境的結構化評估框架和基準協議。
通過在醫療保健（電子健康紀錄分析、可穿戴監測）、金融（波動性預測、高頻交易）和物聯網系統（傳感器融合、預測性維護）中的豐富案例研究，本書展示了連續時間模型如何實現個性化見解、約束感知學習和更可靠的決策。該書專為研究人員、工程師和實務工作者設計，是將連續時間神經方法應用於複雜現實環境的權威資源。
作者簡介
Dr. Mansura Habiba is a seasoned technology leader with over 15 years of experience in AI, machine learning, and cloud architecture. As Principal Platform Architect at IBM, she has led major AI and high-performance computing (HPC) initiatives, delivering advanced solutions for global banks, automotive leaders, and major energy enterprises. Her expertise includes designing scalable AI systems, optimizing cloud infrastructure, and driving digital transformation across IBM Cloud, AWS, and Azure. She is the author of multiple books and numerous peer-reviewed publications that advance the AI and cloud computing community. Driven by a vision to accelerate AI adoption across industries, Dr. Habiba focuses on enabling organizations to harness AI for efficiency, innovation, and sustainable growth. Her commitment to ethical AI, data privacy, and strategic leadership makes her a trusted advisor in today's rapidly evolving digital landscape.
Professor Dr. Barak Pearlmutter is a distinguished researcher and professor in machine learning, neural computation, and theoretical neuroscience. He is a faculty member in the Department of Computer Science at Maynooth University in Ireland, where he conducts pioneering research in automatic differentiation, deep learning, and the brain's computational processes. He is well known for his contributions to automatic differentiation (AD), a foundational technique for optimizing machine learning algorithms by enabling efficient computation of derivatives. His work has significantly influenced computational methods in both theoretical neuroscience and artificial intelligence, and he has also made important contributions to the study of dynamical systems and the development of sophisticated neural computation models. An advocate for interdisciplinary research, Professor Pearlmutter frequently collaborates with neuroscientists, physicists, and engineers to bridge the gap between biological and artificial neural systems. His numerous publications and presentations in top-tier journals and conferences have established him as a leading figure in computational science.
Dr. Mehrdad Maleki is an AI and machine learning expert with a strong mathematical background and a PhD in theoretical computer science. He specializes in developing deep learning models, large language models, and AI-driven solutions for pattern prediction across industries. He has a deep understanding of complex systems and excels at tackling the most challenging problems with a diverse set of advanced tools and techniques. In addition to his AI expertise, Dr. Maleki has a solid foundation in cryptography and quantum computing, and holds several patents in these fields. His work brings together cutting-edge AI methods with secure, scalable solutions applied across a range of domains.
作者簡介(中文翻譯)
曼蘇拉·哈比巴博士是一位資深的技術領導者，擁有超過15年的人工智慧（AI）、機器學習和雲端架構經驗。作為IBM的首席平台架構師，她主導了多項重要的AI和高效能計算（HPC）計畫，為全球銀行、汽車領導者和主要能源企業提供先進的解決方案。她的專業領域包括設計可擴展的AI系統、優化雲端基礎設施，以及推動IBM Cloud、AWS和Azure的數位轉型。她是多本書籍和多篇經過同行評審的出版物的作者，這些作品促進了AI和雲端計算社群的發展。哈比巴博士的願景是加速各行業的AI採用，專注於幫助組織利用AI實現效率、創新和可持續增長。她對倫理AI、數據隱私和戰略領導的承諾，使她成為當今快速變化的數位環境中的可信顧問。
巴拉克·皮爾穆特教授博士是一位在機器學習、神經計算和理論神經科學領域的傑出研究者和教授。他是愛爾蘭梅努斯大學計算機科學系的教職員，進行自動微分、深度學習和大腦計算過程的開創性研究。他因對自動微分（AD）的貢獻而聞名，這是一種基礎技術，用於通過實現導數的高效計算來優化機器學習算法。他的研究對理論神經科學和人工智慧中的計算方法產生了重大影響，並且他在動態系統研究和複雜神經計算模型的發展方面也做出了重要貢獻。作為跨學科研究的倡導者，皮爾穆特教授經常與神經科學家、物理學家和工程師合作，彌合生物神經系統和人工神經系統之間的差距。他在頂尖期刊和會議上的多篇出版物和演講使他成為計算科學領域的領軍人物。
梅赫達德·馬雷基博士是一位擁有強大數學背景和理論計算機科學博士學位的AI和機器學習專家。他專注於開發深度學習模型、大型語言模型和用於各行業模式預測的AI驅動解決方案。他對複雜系統有深入的理解，並擅長使用多種先進工具和技術解決最具挑戰性的問題。除了AI專業知識外，馬雷基博士在密碼學和量子計算方面也有堅實的基礎，並在這些領域擁有多項專利。他的工作將尖端的AI方法與安全、可擴展的解決方案相結合，應用於多個領域。
