Industrial AI

Jay Lee, Industrial AI: Applications with Sustainable Performance, Springer, 2020.

This book introduces Industrial AI in multiple dimensions. Industrial AI is a systematic discipline which focuses on developing, validating and deploying various machine learning algorithms for industrial applications with sustainable performance. Combined with the state-of-the-art sensing, communication and big data analytics platforms, a systematic Industrial AI methodology will allow integration of physical systems with computational models. The concept of Industrial AI is in infancy stage and may encompass the collective use of technologies such as Internet of Things, Cyber-Physical Systems and Big Data Analytics under the Industry 4.0 initiative where embedded computing devices, smart objects and the physical environment interact with each other to reach intended goals. A broad range of Industries including automotive, aerospace, healthcare, semiconductors, energy, transportation, mining, construction, and industrial automation could harness the power of Industrial AI to gain insights into the invisible relationship of the operation conditions and further use that insight to optimize their uptime, productivity and efficiency of their operations. In terms of predictive maintenance, Industrial AI can detect incipient changes in the system and predict the remains useful life and further to optimize maintenance tasks to avoid disruption to operations.

In Defense of a Liberal Education

Fareed Zakaria, In Defense of a Liberal Education, W. W. Norton & Company; 1 edition, March 28, 2016.

The liberal arts are under attack. The governors of Florida, Texas, and North Carolina have all pledged that they will not spend taxpayer money subsidizing the liberal arts, and they seem to have an unlikely ally in President Obama. While at a General Electric plant in early 2014, Obama remarked, "I promise you, folks can make a lot more, potentially, with skilled manufacturing or the trades than they might with an art history degree." These messages are hitting home: majors like English and history, once very popular and highly respected, are in steep decline. 

Markov-chain Monte Carlo for disease dynamics modelling

If you search "pandemic markov chain" under Google scholar, you will find many articles on this topic. For example, Introduction to particle Markov-chain Monte Carlo for disease dynamics modellers by A Endo, E van Leeuwen, M Baguelin - Epidemics, 2019 - Elsevier.

馬友友演奏巴哈大提琴組曲

Yo Yo Ma - Bach Six Cello Suites - BBC Proms 2015

「巴哈的大提琴組曲是我一直以來的音樂夥伴。在將近60年裡，給了我寄託、安慰和快樂，無論是在面對壓力、慶祝，抑或是失去的種種生命時刻。即使經過了三百年的今天，這種音樂仍具有這樣強大的力量、仍可以繼續幫助我們度過艱難的時期。與所有文化一樣，音樂可以幫助我們了解我們的環境、彼此以及我們自己。文化可以幫助我們想像美好的未來，將『他們』變成『我們』，使我們團結在一起。而這些事情在現今社會如此重要。」——馬友友

EE104/CME107 Introduction to Machine Learning

Sanjay Lall, EE104/CME107: Introduction to Machine Learning, Stanford University, Spring Quarter, 2021. (quarter: 10 weeks, Julia, YouTube)

Introduction to machine learning. Formulation of supervised and unsupervised learning problems. Regression and classification. Data standardization and feature engineering. Loss function selection and its effect on learning. Regularization and its role in controlling complexity. Validation and overfitting. Robustness to outliers. Simple numerical implementation. Experiments on data from a wide variety of engineering and other disciplines.

共同作者為 S. Boyd，兩位教授的寫作非常清楚，推薦。 

OSQP: an operator splitting solver for quadratic programs

Bartolomeo Stellato, Goran Banjac, Paul Goulart, Alberto Bemporad, and Stephen Boyd, OSQP: an operator splitting solver for quadratic programs, Mathematical Programming Computation, 2020, vol 12, no. 4, pp. 637–672. (Best Paper of the Year)

We present a general-purpose solver for convex quadratic programs based on the alternating direction method of multipliers, employing a novel operator splitting technique that requires the solution of a quasi-definite linear system with the same coefficient matrix at almost every iteration. Our algorithm is very robust, placing no requirements on the problem data such as positive definiteness of the objective function or linear independence of the constraint functions. It can be configured to be division-free once an initial matrix factorization is carried out, making it suitable for real-time applications in embedded systems. In addition, our technique is the first operator splitting method for quadratic programs able to reliably detect primal and dual infeasible problems from the algorithm iterates. The method also supports factorization caching and warm starting, making it particularly efficient when solving parametrized problems arising in finance, control, and machine learning. Our open-source C implementation OSQP has a small footprint, is library-free, and has been extensively tested on many problem instances from a wide variety of application areas. It is typically ten times faster than competing interior-point methods, and sometimes much more when factorization caching or warm start is used. OSQP has already shown a large impact with tens of thousands of users both in academia and in large corporations.

Software

Users can call OSQP from C, C++, Fortran, Python, Matlab, R, Julia, Ruby and Rust, and via parsers such as CVXPY [1,26], JuMP [33], and YALMIP [65]. 

使用 Python 呼叫 Gurobi 解背包問題 (Knapsack problem)

背包問題 (Knapsack problem) 是一種組合最佳化的 NP 完全問題。問題可以描述為：給定一組物品，每種物品都有自己的重量和價格，在限定的總重量內，我們如何選擇，才能使得物品的總價格最高。

Vattenfall Optimizes Offshore Wind Farm Design

Martina Fischetti, Jesper Runge Kristoffersen, Thomas Hjort, Michele Monaci, and David Pisinger,  Vattenfall Optimizes Offshore Wind Farm Design, INFORMS Journal on Applied Analytics, 2020, Vol. 50, No. 1, pp. 80–94.

In this paper, we describe the use of operations research for offshore wind farm design in Vattenfall, one of the world’s leading companies in the generation of offshore wind energy. We focus on two key aspects that Vattenfall must address in its wind farm design process. The first is determining where to locate the turbines. This aspect is important because the placement of each turbine creates interference on the neighboring turbines, causing a power loss at the overall farm level. The optimizers must minimize this interference based on the wind conditions; however, they must also consider the other costs involved, which depend on factors such as the water depth or soil conditions at each position. The second aspect involves determining how to interconnect the turbines with cables (i.e., cable optimization). This requires Vattenfall to consider both the immediate costs and long-term costs connected with the electrical infrastructure. We developed mixed-integer programming models and matheuristic techniques to solve the two problems as they arise in practical applications. The resulting tools have given Vattenfall a competitive advantage at multiple levels. They facilitate increased revenues and reduced costs of approximately 10 million euros of net present value (NPV) per farm, while ensuring a much faster, more streamlined, and efficient design process. Considering only the sites that Vattenfall has already acquired using our optimization tools, the company experienced NPV gains of more than 150 million euros. This has contributed substantially to its competitiveness in offshore tenders and made green energy cheaper for its end customers. The tools have also been used to design the first wind farms that will be constructed subsidy-free.

Martina Fischetti and David Pisinger, Mathematical Optimization and Algorithms for Offshore Wind Farm Design: An Overview, Business & Information Systems Engineering, 2019, Vol.61, No. 4, pp. 469-485. (Further details)

M. Fischetti, Mixed-integer models and algorithms for wind farm layout optimization. Master’s thesis, University of Padova, 2014. (Stochastic programming for wake effect)

Fischetti M, Fischetti M (2016) Matheuristics. Mart´ı P, Panos P, Resende MG, eds. Handbook of Heuristics (Springer International Publishing, Cham, Switzerland), 1–33.

Grokking Bitcoin

Kalle Rosenbaum, Grokking Bitcoin, Manning, April 2019.

If you think Bitcoin is just an alternative currency for geeks, it's time to think again. Grokking Bitcoin opens up this powerful distributed ledger system, exploring the technology that enables applications both for Bitcoin-based financial transactions and using the blockchain for registering physical property ownership. With this fully illustrated, easy-to-read guide, you'll finally understand how Bitcoin works, how you can use it, and why you can trust the blockchain.

Pitfalls and protocols of data science in manufacturing practice

Chia-Yen Lee & Chen-Fu Chien, Pitfalls and protocols of data science in manufacturing practice, Journal of Intelligent Manufacturing, 2020.

Driven by ongoing migration for Industry 4.0, the increasing adoption of artificial intelligence, big data analytics, cloud computing, Internet of Things, and robotics have empowered smart manufacturing and digital transformation. However, increasing applications of machine learning and data science (DS) techniques present a range of procedural issues including those that involved in data, assumptions, methodologies, and applicable conditions. Each of these issues may increase difficulties for implementation in practice, especially associated with the manufacturing characteristics and domain knowledge. However, little research has been done to examine and resolve related issues systematically. Gaps of existing studies can be traced to the lack of a framework within which the pitfalls involved in implementation procedures can be identified and thus appropriate procedures for employing effective methodologies can be suggested. This study aims to develop a five-phase analytics framework that can facilitate the investigation of pitfalls for intelligent manufacturing and suggest protocols to empower practical applications of the DS methodologies from descriptive and predictive analytics to prescriptive and automating analytics in various contexts.

Targeting Prospective Customers: Robustness of Machine-Learning Methods to Typical Data Challenges

Duncan Simester, Artem Timoshenko, and Spyros I. Zoumpoulis, Targeting Prospective Customers: Robustness of Machine-Learning Methods to Typical Data Challenges, Management Science, 2020, Vol. 66, No. 6, Pages: 2495–2522.

We investigate how firms can use the results of field experiments to optimize the targeting of promotions when prospecting for new customers. We evaluate seven widely used machine-learning methods using a series of two large-scale field experiments. The first field experiment generates a common pool of training data for each of the seven methods. We then validate the seven optimized policies provided by each method together with uniform benchmark policies in a second field experiment. The findings not only compare the performance of the targeting methods, but also demonstrate how well the methods address common data challenges. Our results reveal that when the training data are ideal, model-driven methods perform better than distance-driven methods and classification methods. However, the performance advantage vanishes in the presence of challenges that affect the quality of the training data, including the extent to which the training data captures details of the implementation setting. The challenges we study are covariate shift, concept shift, information loss through aggregation, and imbalanced data. Intuitively, the model-driven methods make better use of the information available in the training data, but the performance of these methods is more sensitive to deterioration in the quality of this information. The classification methods we tested performed relatively poorly. We explain the poor performance of the classification methods in our setting and describe how the performance of these methods could be improved.

Machine Learning Under a Modern Optimization Lens

Dimitris Bertsimas and Jack Dunn, Machine Learning Under a Modern Optimization Lens, Dynamic Ideas LLC, 2019.

The book provides an original treatment of machine learning (ML) using convex, robust and mixed integer optimization that leads to solutions to central ML problems at large scale that can be found in seconds/minutes, can be certified to be optimal in minutes/hours, and outperform classical heuristic approaches in out-of-sample experiments.

Efficient Large-Scale Internet Media Selection Optimization for Online Display Advertising

Courtney Paulson, Lan Luo, and Gareth M. James, “Efficient Large-ScaleInternet Media Selection Optimization for Online Display Advertising,” Journal of Marketing Research, vol. 55, 2018, pp. 489-506. 

In today's digital market, the number of websites available for advertising has ballooned into the millions. Consequently, firms often turn to ad agencies and demand-side platforms (DSPs) to decide how to allocate their Internet display advertising budgets. Nevertheless, most extant DSP algorithms are rule-based and strictly proprietary. This article is among the first efforts in marketing to develop a nonproprietary algorithm for optimal budget allocation of Internet display ads within the context of programmatic advertising. Unlike many DSP algorithms that treat each ad impression independently, this method explicitly accounts for viewership correlations across websites. Consequently, campaign managers can make optimal bidding decisions over the entire set of advertising opportunities. More importantly, they can avoid overbidding for impressions from high-cost publishers, unless such sites reach an otherwise unreachable audience. The proposed method can also be used as a budget-setting tool, because it readily provides optimal bidding guidelines for a range of campaign budgets. Finally, this method can accommodate several practical considerations including consumer targeting, target frequency of ad exposure, and mandatory media coverage to matched content websites.

Algorithm: Coordinate descent algorithm for budget optimization problem (7).