聯絡人：孫長敏助教

聯絡電話：02-86741111分機66755

刊登時間：2025-04-29 09:43:52

聯絡人：孫長敏助教

聯絡電話：分機66755

刊登時間：2025-04-08 10:41:53

聯絡人：孫長敏助教

聯絡電話：分機66755

刊登時間：2024-11-06 10:24:22

附件：演講摘要

聯絡人：孫長敏助教

聯絡電話：66755

刊登時間：2024-10-15 09:11:19

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題：Optimal Sequence Identification in Order-of-Addition Experiments Using Complete Consecutive Order Pairing (CCOP) Designs

主講人：潘建興 教授 (中央研究院 統計科學所)

時間：115年05月06日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

Order-of-addition (OofA) experiments investigate how the sequence in which components are introduced affects the experimental response, and they have attracted considerable attention in the experimental design literature over the past decade. Recently, a new class of designs, known as Complete Consecutive Order Pairing (CCOP) designs, together with their associated analysis methods, has been proposed to provide a cost-efficient framework for conducting OofA experiments and identifying optimal input sequences when components may have multiple levels. These designs substantially reduce experimental cost while retaining strong inferential efficiency. In this talk, we first review the fundamental principles and advantages of this cost-efficient experimental approach. Two real-world examples are then presented to illustrate how CCOP designs can be applied to identify optimal sequences of practical interest. Finally, we discuss recent extensions of the CCOP framework that broaden its applicability to more complex experimental settings.

~歡迎參加~

國立臺北大學統計學系  敬邀

115.04.29

刊登時間：2026-04-29 14:12:18

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題：Bayesian Joint Modeling with Heteroscedastic Covariance for Longitudinal and Time-to-Event Data

主講人：李國榮 教授 (國立成功大學 統計學系)

時間：115 年 04 月 22 日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

In clinical and epidemiological research, understanding the interplay between longitudinal biomarker measurements and time-to-event outcomes is critical for disease modeling and risk prediction. This paper presents a new Bayesian joint modeling framework that integrates a linear mixed-effects model (LMM) for longitudinal data and a Cox proportional hazards model for survival outcomes. The framework introduces a flexible covariance structure for longitudinal outcomes using hypersphere decomposition within the variance-correlation decomposition (HDVCD) framework. This method ensures positive definiteness while capturing serial correlations and subject-level heterogeneity. We explore three distinct association structures, random-effects model (REM), shared parameter model (SPM), and trajectory model (TM) to capture the relationship between the longitudinal and survival processes. Simulation studies and real-world applications demonstrate the robustness and enhanced predictive performance of the proposed models, particularly in handling complex covariance structures. This framework advances the flexibility and reliability of traditional joint modeling approaches.

~歡迎參加~

國立臺北大學統計學系  敬邀

115.04.09

刊登時間：2026-04-09 16:34:03

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題：Is Complementary-Label Learning Realistic?

主講人：林軒田 教授 (國立臺灣大學 資訊工程學系)

時間：115 年 04 月 08 日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

Complementary-Label Learning (CLL) is a weakly supervised learning paradigm that is claimed to be useful in situations where collecting true labels is expensive. This talk begins by walking the audience through two advances in CLL. The first work discovered that the CLL loss design from Unbiased Risk Estimator (URE) suffers from high variance in gradient estimation. To address this, a novel surrogate complementary loss (SCL) framework is proposed, which reduces variance and improves gradient alignment, mitigating the overfitting issue. The second work introduces a new approach to CLL by reducing the problem to estimating the probability of complementary classes. This framework sidesteps the limitations of traditional methods and improves robustness in noisy environments, offering a broader perspective for both deep and non-deep models. Empirical results demonstrate the efficacy of both approaches in improving CLL performance. Finally, the speaker will share some of the ongoing attempts in making CLL more realistic, including some firsthand experience in collecting real-world datasets and releasing an open-source library.

~歡迎參加~

國立臺北大學統計學系  敬邀

115.04.07

刊登時間：2026-04-07 19:29:18

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題：Generalized Method of Moments for Semiparametric Transformation Models of Recurrent Events with Informative Censoring

主講人：蔡長宇 教授 (國立高雄大學 統計學研究所)

時間：115 年 04 月 01 日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

In this talk, we consider semiparametric transformation models for recurrent events with a shared frailty variable, allowing for dependence between the event process and censoring. Unlike standard shared frailty proportional rate models, our framework relaxes the proportionality assumption and allows the rate functions to vary flexibly across covariates over time. Motivated by the decomposition of the rate function into shape and size components (Wang and Huang, 2014), we develop an inverse-rate weighting approach and a generalized method of moments framework that combines information from different components of the rate function to improve efficiency. We establish large-sample properties, evaluate finite-sample performance through simulations, and demonstrate practical utility using a real dataset.  

~歡迎參加~

國立臺北大學統計學系  敬邀

115.03.29

刊登時間：2026-03-29 21:54:42

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題：統計學-跨域融合與典範轉移：AI 驅動下的資料科學未來展望

主講人：謝邦昌教授 (輔仁大學學術特聘講座教授)

時間：115年3月11日 (星期三，14：00~16：00)

地點：三峽校區商學院3F13教室

Abstract

在人工智慧快速崛起與數位科技全面滲透的時代，資料科學正經歷一場深刻的典範轉移。統計學作為資料分析與科學推論的基礎學科，正由傳統的分析工具轉化為連結跨領域知識、推動創新與決策的重要核心。本演講將從宏觀視角探討資料科學在 AI 驅動下的發展趨勢，回顧統計學、機器學習與深度學習交織演進的歷程，並說明計算能力突破、海量資料累積與演算法創新如何重塑科學研究與產業發展的模式。隨著「AI for Science」浪潮興起，人工智慧正逐步成為科學發現的重要推動力量，從生命科學、醫療健康到材料設計與智慧產業，皆展現出跨域融合所帶來的巨大潛力。面對未來，本演講亦將從產官學多元視角探討資料科學的關鍵趨勢與挑戰，包括多模態資料整合、人機協作新模式，以及倫理治理與社會責任等議題，期望引領聽眾重新思考統計學與資料科學在智慧時代中的角色與使命，共同展望以資料與智慧驅動的未來創新格局。

~歡迎參加~

國立臺北大學統計學系  敬邀

115.03.09

刊登時間：2026-03-09 00:40:58

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題：Volatility-of-Volatility Aligned Uncertainty and Return Predictability 

主講人：陳德峰 副教授(香港理工大學會計及金融學院) 

時間：115年03月04日 (星期三，13：10~15：00) 

地點：三峽校區商學院1F01教室 

Abstract

We propose a novel approach to forecasting market returns by constructing economic uncertainty indices aligned with future volatility-of-volatility (VOV). We employ the partial least squares (PLS) method to synthesize information from economic policy uncertainty (EPU) indices and macro-financial uncertainty measures into indices that best predict future VOV. This VOV alignment significantly enhances return predictability, delivering out-of-sample R² up to 13% for U.S. equities and generating substantial economic gains for mean-variance investors. Predictability extends beyond equities to hedge fund returns ( R² up to 18.3%) and global markets ( R² up to 18.7%). Our results clarify which uncertainty components—such as financial uncertainty, sovereign debt crises, and regulatory risk—drive predictive power, offering an interpretable and theoretically grounded tool for improving asset allocation across diverse asset classes and geographies. 

 

~歡迎參加~

國立臺北大學統計學系  敬邀

115.03.02

刊登時間：2026-03-02 19:30:40

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題： Nature-inspired Metaheuristics as a General-Purpose Optimization Tool in statistical Research

主講人：Professor Weng Kee Wong (Department of Biostatistics, Fielding School of Public Health, University of California at Los Angeles)

時間：114年12月10日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

Nature-metaheuristics have been widely used in engineering and computer science to address various types of optimization problems for decades and are now increasingly used across disciplines. They are increasingly popular in industry and academia for tackling all kinds of complex and high-dimensional optimization problems. Interestingly, metaheuristics seems to be still relatively underused in the statistical research community. I present an overview of nature-inspired metaheuristics and some of their applications in statistics. The main appealing features of these algorithms are their speed, flexibility, availability of codes in different platforms, and ease of implementation and usage. Above all, they are virtually assumptions free, which allows us to apply them to solve a huge range of high-dimensional optimization tasks. I will enumerate the advantages of nature-inspired metaheuristic algorithms over existing optimization algorithms and illustrate their diverse applications in biostatistics, and beyond. In particular, I will show how nature-inspired algorithms can find more flexible and computationally challenging designs for early-phase clinical trials. Keywords: Design Efficiency, Early Phase Trials, Optimal Experiment Designs, Particle Swarm Optimization.

Bio of Wong Bio

Professor Wong is a Professor at UCLA since 1990 and over the years, he has done collaborative work in dentistry, environment health science, rheumatology, and various domains in oncology, including in the design and analysis of cancer control and prevention trials for controlling Hepatitis B among Asians, colorectal cancer for Hispanics, and fighting obesity and promotion of minority health at workplace. His main methodology research area is in constructing model-based optimal experimental designs for biostatistical applications. His recent interest is in the applications of nature-inspired metaheuristics to tackle challenging problems in designs for toxicology and other areas of statistics. He has delivered more than 230 presentations globally, including recent short courses at Seoul National University, UCLA's Pediatric Dentistry Program, and at the Toxicology Center in TU Dortmund University in Germany. Professor Wong has received several R01 grant awards from NIH, NSF and private foundations as a principal investigator. He is fellow of the American Statistical Association, the Institute of Mathematical Statistics, the American Association for the Advancement of Science, an elected member of the International Statistical Institute and a full member of the Sigma Xi - The Scientific Research Honor Society. He currently holds a 3-year Yushan Scholarship Award from the Ministry of Education in Taiwan.

 

~歡迎參加~

國立臺北大學統計學系  敬邀

114.12.03

刊登時間：2025-12-04 15:48:47

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題： Machine Learning to Advance Genome-Wide Association Studies for Predicting Progression of Pediatric Asthma

主講人：張書瑋 教授 (長庚大學健康數據科學研究所)

時間：114年12月3日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

Due to the challenges in diagnosing the progression of pediatric asthma, persistence, exacerbation, and rehabilitation in particular, we developed a predictive framework that integrated whole-genome genotyping data with questionnaire and clinical information to classify disease outcomes. Through multiple combinations of different feature screening methods, sampling strategies, and machine learning and deep learning models, we successfully applied high-dimensional genotyping data to model training. Approxiately 6.27 million SNP features in the original data were reduced to a couple of hundreds to less than two thousand, which still achieved excellent predictive performance (ROC AUC from 0.95 to 0.98). We also integrated the results of all training combinations to compare the performance of different models to identify the key predictive features of the best model. The machine learning model we trained was proved to effectively predict the occurrence and persistence of asthma in children.

 

~歡迎參加~

國立臺北大學統計學系  敬邀

114.11.26

刊登時間：2025-12-02 20:39:04

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題： Time series forecasting: Exploring hybrid strategies with singular spectrum analysis

主講人：Prof. Paulo Canas Rodrigues

時間：114年11月26日 (星期三，13：00~15：00)

地點：三峽校區商學院1F02教室

Abstract

Time series forecasting plays a key role in areas such as energy, environment, economy, and finances. Hybrid methodologies, combining the results of statistical, mathematical, and machine learning methods, have become popular for time series analysis and forecasting, as they allow researchers to compensate for the limitations of one approach with the strengths of the other and combine them into new frameworks while improving forecasting accuracy. In this class of methods, algorithms for time series forecasting are applied sequentially, i.e., the second algorithm is applied to the residuals that were not captured by the first one. In this talk, I will discuss several hybrid strategies for time series forecasting that use singular spectrum analysis, classical time series models, and recurrent neural networks, with application to several areas of research.

Bio: Paulo Canas Rodrigues is a Professor of Statistics and Data Science at the Federal University of Bahia and the Director of the Statistical Learning Laboratory (SaLLy; www.SaLLy.ufba.br). Paulo completed his Ph.D. in Statistics at the Nova University of Lisbon, Portugal (2012), and his Habilitation in Mathematics, with a specialization in Statistics and Stochastic Processes, at the Lisbon University, Portugal (2019). His research in time series forecasting, statistical learning, artificial intelligence, statistics, and data science resulted in more than 120 scientific papers in collaboration with more than 180 co-authors from 86 universities in 31 countries and delivered more than 180 invited talks at conferences and scientific seminars. He is an ISI Elected Member. Among other activities, he is the current President of the International Association for Statistical Computing, the Past-President of the International Society for Business and Industrial Statistics, a Member of the Representative Council of the International Biometric Society, and a Council Member of the International Statistical Institute. Website: www.paulocanas.org; www.SaLLy.ufba.br.

 

~歡迎參加~

國立臺北大學統計學系  敬邀

114.11.12

刊登時間：2025-11-25 12:12:05

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題： Inference for the Parameters of Kumaraswamy Distribution

主講人：曾耀霆 教授 (Department of Statistics and Data Science, Tamkang University)

時間：114年11月12日 (星期三，13：00~15：00)

地點：三峽校區商學院3F13教室

Abstract

This study investigates parameter inference for the Kumaraswamy distribution under progressively type-II censored samples. Maximum likelihood estimators are derived, and their existence and uniqueness are established. Exact confidence intervals and joint confidence regions are constructed using pivotal quantities. The performance of the proposed methods is evaluated through simulation studies. Finally, an analysis of a real data example is conducted to demonstrate the applicability of the proposed estimation methods.

Keywords: Confidence interval; Joint confidence region; Maximum likelihood estimation; Pivot; Progressive type-II censoring.

 

~歡迎參加~

國立臺北大學統計學系  敬邀

114.11.02

刊登時間：2025-11-04 21:49:20

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題：  Robust Estimation of Population Attributable Fractions in the Presence of Multiple Ordered Mediators

主講人：戴安順 教授 (Institute of Statistics and Data Science, National Tsing Hua University)

時間：114年11月05日 (星期三，13：00~15：00)

地點：三峽校區商學院3F13教室

Abstract

Population Attributable Fractions (PAFs) are critical tools for quantifying the proportion of an outcome in a population that can be attributed to specific exposures. In scenarios involving multiple ordered mediators, accurately estimating PAFs becomes complex due to potential mediator-outcome relationships, mediator-mediator interactions, and confounding. This study introduces a robust estimation framework for PAFs in the presence of multiple ordered mediators, called mediated PAFs, leveraging multiply robust estimators that remain consistent under various model misspecifications. The proposed method accommodates the ordered structure of mediators and addresses challenges such as mediator collinearity and indirect effects, ensuring reliable and interpretable estimates. Through extensive simulations and application to real-world data, we demonstrate the utility of our method in uncovering the pathways and population-level impact of exposures mediated by ordered intermediate variables. This work offers a versatile and practical tool for researchers in epidemiology and public health, enabling more nuanced analyses of mediation effects in complex systems.

 

~歡迎參加~

國立臺北大學統計學系  敬邀

114.10.29

刊登時間：2025-11-04 21:49:39

附件：演講摘要

國立臺北大學統計學系

專題演講

 

講題： AI, BI & SI—Artificial, Biological and Statistical Intelligences (Part 2)

主講人：林共進 特聘教授 (Department of Statistics, Purdue University)

時間：114年10月30日 (星期四，15：00~17：00)

地點：三峽校區商學院3F13教室

Abstract

Artificial Intelligence (AI) is clearly one of the hottest subjects these days. Basically, AI employs a huge number of inputs (training data), super-efficient computer power/memory, and smart algorithms to perform its intelligence. In contrast, Biological Intelligence (BI) is a natural intelligence that requires very little or even no input. This talk will first discuss the fundamental issue of input (training data) for AI. After all, not-so-informative inputs (even if they are huge) will result in a not-so-intelligent AI. Specifically, three issues will be discussed: (1) input bias, (2) data right vs. right data, and (3) sample vs. population. Finally, the importance of Statistical Intelligence (SI) will be introduced. SI is somehow in between AI and BI. It employs important sample data, solid theoretically proven statistical inference/models, and natural intelligence. In my view, AI will become more and more powerful in many senses, but it will never replace BI. After all, it is said that “The truth is stranger than fiction, because fiction must make sense.” The ultimate goal of this study is to find out “how can humans use AI, BI, and SI together to do things better.” Dr. Dennis K. J. Lin is a Distinguished Professor of Statistics at Purdue University. He served as the Department Head during 2020-2022. Prior to this current job, he was a University Distinguished Professor of Supply Chain Management and Statistics at Penn State, where he worked for 25 years. His research interests are data quality, industrial statistics, statistical inference, and data science. He has published nearly 300 SCI/SSCI papers in a wide variety of journals. He currently serves or has served as an associate editor for more than 10 professional journals and was a co-editor for Applied Stochastic Models for Business and Industry. Dr. Lin is an elected fellow of ASA, IMS, ASQ, & RSS, an elected member of ISI, and a lifetime member of ICSA. He is an honorary chair professor for various universities, including Fudan University, and National Taiwan Normal University and a Chang-Jiang Scholar at Renmin University of China,. His recent awards include, the Youden Address (ASQ, 2010), the Shewell Award (ASQ, 2010), the Don Owen Award (ASA, 2011), the Loutit Address (SSC, 2011), the Hunter Award (ASQ, 2014), the Shewhart Medal (ASQ, 2015), and the SPES Award (ASA, 2016). He was the Deming Lecturer Award at 2020 JSM. His most recent award is “The 2022 Distinguished Alumni Award” (National Tsing Hua University, Taiwan). 

 

~歡迎參加~

國立臺北大學統計學系  敬邀

114.10.27

刊登時間：2025-10-28 23:26:39

附件：演講摘要

國立臺北大學統計學系

專題演講

講題： 從方法正確到決策可行：在 LLM 與 Agentic Workflow 時代重構資料分析師角色

主講人：劉嘉凱 執行長 (智庫驅動股份有限公司)

時間：114年10月15日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

當大型語言模型（LLM）與代理人工作流（Agentic Workflow）正逐步重塑企業的決策模式，資料分析師的價值也從「方法正確」轉向「決策可行」。 本演講將帶領聽眾思考：在分析自動化與智慧代理快速普及的時代，統計與資料科學的專業訓練，該如何轉化為「能被採用的洞察」？ 我們將從學術研究所追求的顯著性與模型精準度，延伸至實務場域更關注的決策採用率、投資報酬率（ROI）與責任鏈（Accountability Chain），並探討資料分析師如何進化為「智能決策設計師（Intelligent Decision Designer）」——能定義問題、設計決策流程，並監督人機協作的可信任運作。透過真實案例與互動討論，參與者將理解如何在 LLM 與 Agentic Workflow 時代，打造兼具理論嚴謹與實務影響力的專業競爭力。

 

~歡迎參加~

國立臺北大學統計學系 敬邀
114.10.08

刊登時間：2025-10-14 13:49:15

附件：演講摘要

 

國立臺北大學統計學系

專題演講

講題：AI Agents as Data Partners: Elevating the Role of the Analyst

主講人：詹欣諭 Senior Technical Trainer (Microsoft)

時間：114年10月08日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

As data complexity and real-time insight demands accelerate, traditional analytics workflows face mounting bottlenecks. This talk posits a new paradigm: making AI agents partners in data analysis. Rather than passive tools, these agents interpret intent, plan workflows, invoke auxiliary tools, detect errors, and self-adjust—supporting the full pipeline from exploration and modeling to insight generation. Through this lens, we ask: how might analysts evolve? What design challenges and trade-offs emerge in agentic systems? And how might future data analysis place greater emphasis on strategy, verification, and human–agent collaboration?

 

 

~歡迎參加~

國立臺北大學統計學系  敬邀

114.10.05

刊登時間：2025-10-08 11:54:54

附件：演講摘要

 

國立臺北大學統計學系

專題演講

講題： Nonparametric Control Charts with the Exceedance Probability Criterion: Advances from Univariate to Multivariate Methods

主講人：楊宏基 教授 (東華大學應數系)

時間：114年10月01日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

Statistical process monitoring plays a crucial role in ensuring product quality and process stability. However, traditional control chart methods often face limitations when sample sizes are small or when distributional assumptions are difficult to justify. To address these challenges, this talk introduces a new class of nonparametric control charts built upon the Exceedance Probability Criterion (EPC). For univariate settings, fractional order statistics (FOS) are employed to construct control limits, offering both rigorous theoretical guarantees and practical applicability under small to moderate sample sizes. The framework is further extended to multivariate processes by integrating depth functions with the EPC, thereby enhancing sensitivity to multidimensional anomalies and subtle process shifts. Simulation studies and real data applications demonstrate that the proposed approaches outperform existing methods in terms of coverage probability, detection power, and robustness. This talk highlights the theoretical foundation, methodological innovations, and practical applications of EPC-based nonparametric control charts, presenting a unified and flexible framework for modern process monitoring.

 

 

~歡迎參加~

國立臺北大學統計學系  敬邀

114.09.18

刊登時間：2025-09-19 13:03:43

附件：演講摘要

 

國立臺北大學統計學系

專題演講

講題： AI 的思維與創新

主講人：張志勇 教授 (淡江大學資工系)

時間：114年9月17日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

This lecture focuses on “AI Literacy and Thinking,” beginning with a review of the development of artificial intelligence—from electrification and data-driven technologies to the rise of generative AI—explaining how it imitates and extends human cognition, perception, and reasoning. It then highlights AI applications in fields such as image processing, natural language, healthcare, finance, and transportation, showcasing examples from smart factories, financial analysis, medical diagnosis, and social media to underscore both the value and challenges of AI in industry. The lecture further emphasizes the emergence and impact of generative AI. Generative AI can learn from vast amounts of text, audio, and visual data, demonstrating remarkable capabilities in writing, code generation, and image and video creation, greatly enhancing industrial efficiency. At the same time, it warns of challenges brought by generative AI, such as hallucinations, responsibility attribution, and over-reliance on expertise, urging individuals to cultivate proper literacy and critical thinking. Finally, the lecture stresses that AI cannot replace human emotion, trust, and creativity; only by leveraging new tools and integrating them with professional expertise can we create greater value in collaboration with AI.

 

 

~歡迎參加~

國立臺北大學統計學系  敬邀

114.09.12

刊登時間：2025-09-15 15:36:50

附件：演講摘要

國立臺北大學統計學系

專題演講

講題： 企業 AI 應用以及導入實作

主講人：葉盈伶 資深經理

時間：114年9月10日 (星期三，13：10~15：00)

地點：三峽校區商學院3F13教室

Abstract

本演講以企業 2023 年內部引入 AI 實例為出發點，說明企業導入 AI 的心法；說明 2025 年 LLM 的最新發展與應用場景，並探討 AI 對未來工作的影響與 AI 工程師必備技能。內容由實務經驗延伸至技術趨勢與未來挑戰，提供了企業端的觀察視角。 

~ ~ 歡迎參加 ~ ~

台北大學統計學系

114.09.10

刊登時間：2025-09-15 15:35:38

附件：演講摘要