Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

We will review some of the major results in random graphs and some of the more challenging open problems. We will cover algorithmic and structural questions. We will touch on newer models, including those related to the WWW.

Random graphs

The topic of dynamic pricing and learning has received a considerable amount of attention in recent years, from different scientific communities. We survey these literature streams: we provide a brief introduction to the historical origins of quantitative research on pricing and demand estimation, point to different subfields in the area of dynamic pricing, and provide an in-depth overview of the available literature on dynamic pricing and learning. Our focus is on the operations research and management science literature, but we also discuss relevant contributions from marketing, economics, econometrics, and computer science. We discuss relations with methodologically related research areas, and identify directions for future research.

Dynamic pricing and learning: Historical origins, current research, and new directions

http://www.cs.nott.ac.uk/~pszgxk/papers/corsportsbib.pdf

Invited Review: Scheduling in sports: An annotated bibliography

カナダ鉱山冶金学会 (Canadian Institute of Mining and Metallurgy) の憲章と定款

We study a family of stylized assortment planning problems, where arriving customers make purchase decisions among offered products based on maximizing their utility. Given limited display capacity and no a priori information on consumers' utility, the retailer must select which subset of products to offer. By offering different assortments and observing the resulting purchase behavior, the retailer learns about consumer preferences, but this experimentation should be balanced with the goal of maximizing revenues. We develop a family of dynamic policies that judiciously balance the aforementioned trade-off between exploration and exploitation, and prove that their performance cannot be improved upon in a precise mathematical sense. One salient feature of these policies is that they “quickly” recognize, and hence limit experimentation on, strictly suboptimal products.

/pdf/optimal-dynamic-assortment-planning-with-demand-learning-126nx6hsva.pdf

Optimal Dynamic Assortment Planning with Demand Learning

Summary Background By July 14, 2021, 81·3 % of adults (aged ≥18 years) in Chile had received a first SARS-CoV-2 vaccine and 72·3% had received a second SARS-CoV-2 vaccine, with the majority of people given Sinovac's inactivated CoronaVac vaccine (75·3% of vaccines dispensed) or Pfizer–BioNTech's mRNA BNT162b2 vaccine (20·9% of vaccines dispensed). Due to the absence of simultaneous real-world data for these vaccines, we aimed to compare SARS-CoV-2 IgG positivity between vaccines using a dynamic national monitoring strategy. Methods From March 12, 2021, 28 testing stations for SARS-CoV-2 IgG detection were installed in hotspots based on cellular-phone mobility tracking within the most populated cities in Chile. Individuals voluntarily approaching the testing stations were invited to do a lateral flow test by finger prick and respond to a questionnaire on sociodemographic characteristics, vaccination status (including type of vaccine if one was received), variables associated with SARS-CoV-2 exposure, and comorbidities. We compared the proportion of individuals testing positive for anti-SARS-CoV-2 IgG across sites by week since vaccination between recipients of CoronaVac and BNT162b2. Unvaccinated participants served as a control population and were matched to vaccinated individuals on the basis of date of presentation to the testing station, gender, and age group. Individuals were excluded from the analysis if they were younger than 18 years, had no declared gender, had an invalid IgG test result, had previously tested positive for SARS-CoV-2 infection on PCR, could not recall their vaccination status, or had been immunised against COVID-19 with vaccines other than CoronaVac or BNT162b2. Here, we report data collected up to July 2, 2021. Findings Of 64 813 individuals enrolled, 56 261 were included in the final analysis, of whom 33 533 (59·6%) had received at least one dose of the CoronaVac vaccine, 8947 (15·9%) had received at least one dose of the BNT162b2 vaccine, and 13 781 (24·5%) had not received a vaccine. SARS-CoV-2 IgG positivity during week 4 after the first dose of CoronaVac was 28·1% (95% CI 25·0–31·2; 220 of 783 individuals), reaching a peak of 77·4% (75·5–79·3; 1473 of 1902 individuals) during week 3 after the second dose. SARS-CoV-2 IgG positivity during week 4 after the first dose of the BNT162b2 vaccine was 79·4% (75·7–83·1; 367 of 462 individuals), increasing to 96·5% (94·9–98·1; 497 of 515 individuals) during week 3 after the second dose and remaining above 92% until the end of the study. For unvaccinated individuals, IgG seropositivity ranged from 6·0% (4·4–7·6; 49 of 810 individuals) to 18·7% (12·5–24·9; 28 of 150 individuals) during the 5 month period. Regression analyses showed that IgG seropositivity was significantly lower in men than women and in people with diabetes or chronic diseases for CoronaVac vaccine recipients (p Interpretation IgG seropositivity was lower after CoronaVac than after BNT162b2 and declined over time since vaccination for CoronaVac recipients but not BNT162b2 recipients. Prolonged IgG monitoring will allow further evaluation of seropositivity overtime, providing data, in conjunction with effectiveness studies, for possible future re-assessment of vaccination strategies. Funding Instituto Sistemas Complejos de Ingenieria and Ministerio de Salud Chile. Translation For the Spanish translation of the abstract see Supplementary Materials section.

Dynamic IgG seropositivity after rollout of CoronaVac and BNT162b2 COVID-19 vaccines in Chile: a sentinel surveillance study.

Since 2005, Chile's professional soccer league has used a game-scheduling system that is based on an integer linear programming model. The Chilean league managers considered several operational, economic, and sporting criteria for the final tournaments' scheduling. Thus, they created a highly constrained problem that had been, in practice, unsolvable using their previous methodology. This led to the adoption of a model that used some techniques that were new in soccer-league sports scheduling. The schedules they generated provided the teams with benefits such as lower costs, higher incomes, and fairer seasons. In addition, the tournaments were more attractive to sports fans. The success of the new scheduling system has completely fulfilled the expectations of the Asociacion Nacional de Futbol Profesional (ANFP), the organization for Chilean professional soccer.

/pdf/scheduling-the-chilean-soccer-league-by-integer-programming-3tlqxev3em.pdf

Scheduling the Chilean Soccer League by Integer Programming

Scheduling the South American Qualifiers to the 2018 FIFA World Cup by integer programming

We consider an online retailer facing heterogeneous customers with initially unknown product preferences. Customers are characterized by a diverse set of demographic and transactional attributes. T...

Denis Sauré

Papers

Optimal Dynamic Assortment Planning with Demand Learning

Dynamic IgG seropositivity after rollout of CoronaVac and BNT162b2 COVID-19 vaccines in Chile: a sentinel surveillance study.

Scheduling the Chilean Soccer League by Integer Programming

Scheduling the South American Qualifiers to the 2018 FIFA World Cup by integer programming

A Dynamic Clustering Approach to Data-Driven Assortment Personalization