Agenda Day Two
Event Home| Additional Information | Agenda | Attendee List | CPD | Hotel Information | Registration | Sponsors | Sponsor this Event
Agenda | Agenda Day Two | Agenda Day Three
Track Legend
B/I – Beginner/Implementer
M/S – Manager/Supervisor
AP – Advanced Practitioner
| Thursday, September 20 |
|---|
|
6:00 a.m. – 4:00 p.m.
|
|
6:00 a.m. – 7:00 a.m.
Brief summaries of progress and goals for the Predictive Analytics and Futurism section (PAF) plus a welcome to the Symposium and overview. Session Coordinator(s) Facilitator(s) |
|
7:00 a.m. – 8:15 a.m.
Presentation(s): View Presentation Competency: Strategic Insight and Integration Moderator(s): David L. Snell, ASA, MAAA Presenter(s): Chuck Bloss, FSA, MAAA, FCA, SVP & Chief Actuary; Dr. Philip Smalley, MD, FRCPC, Chief Medical Director Panelists will explore the emergence of data analytics within health care. Genetics/Genomics represent a high potential health medical advance. This session will cover how companies apply them to help the insured live longer with less morbidity and mortality, and to help insurers spend less on prescription drugs. Precision medicine can reduce the number of prescriptions necessary for covered participants, providing both better health results for the participant and lower costs for the health insurance provider. Track: All Session Coordinator(s) Facilitator(s) |
|
8:30 a.m. – 9:45 a.m.
Presentation(s): View Presentation Competency: Results-Oriented Solutions Moderator(s): Dorothy Andrews, ASA, MAAA Presenter(s): Jack T. Kerbeshian, FSA, MAAA The Delphi method is a non-quantitative forecasting approach that is most useful when other forecasting techniques appear to have limited value. The Society of Actuaries (SOA) has had some impressive results from Delphi studies (even quoted in the Wall Street Journal for their uncanny accuracy on long-range predictions). Attend this session to learn about this effective, yet underutilized method of leveraging the wisdom of crowds. Track: Manager/Supervisor Session Coordinator(s) Minyu Cao, FSA, CERA Facilitator(s) Competency: External Forces & Industry Knowledge Moderator(s): Michael David Hoyer, FSA, MAAA Presenter(s): Chris E. Stehno This session will cover new data sources such as electronic health records, social media and the Internet of Things (IoT), and how they are dramatically changing the way we get our information for predictive analytics (PA). Track: Beginner/Implementer Session Coordinator(s) David L. Snell, ASA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Relationship Management Moderator(s): Bradley Joseph Lipic Presenter(s): Eileen Sheila Burns, FSA, MAAA; Jeevan Duggempudi; Qichun Xu, FSA This session will cover building a Data Science team including hiring and retaining talent, and bringing different skillsets to the table such as Data Engineers, Data Scientists and Data Storytellers. Where actuaries fit into this picture, how actuaries work with data scientists and how to get a job in this new area will also be discussed. Track: Advanced Practitioner Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Michael Cletus Niemerg, FSA, MAAA Presenter(s): Talex Diede This session will cover clustering techniques generally (k-means, hierarchical), and how k-means have been used to segment policyholders using third party features such as credit scores to assess customer profitability. Track: All Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Kimberly M. Steiner, FSA, MAAA Presenter(s): Julia Romero, FSA, CERA This topic will cover how to integrate traditional actuarial profit models with data science/predictive models, so one could use traditional data science techniques and modeling approaches to evaluating profitability and customer value. This session will cover rethinking actuarial profit models as an end-to-end model in a data oriented world. Track: All Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Jason Hiquet FSA,CERA Presenter(s): Marjorie A. Rosenberg, FSA This session will cover a case study to illustrate a novel way of clustering individuals to create similar groups where covariates are all categorical. This method adapts to categorical data where there is no inherent order in the variable, like race. Data from the National Health Interview Survey (NHIS) was used to form the clusters and apply them for prediction purposes to the Medical Expenditures Panel Study (MEPS). This approach considers the person-weighting of surveys to produce clusters and expenditure estimates per cluster that are representative of the US adult civilian non-institutionalized population. For the clustering method, we apply the K-Medoids approach with an adapted version of the Goodall dissimilarity index. This approach is validated on independent NHIS/MEPS data from a different panel. The results indicate the robustness of the clusters across years and indicate the ability to distinguish them for the predictability of expenditures. Track: Beginner/Implementer Session Coordinator(s) Stuart Klugman, FSA, CERA Facilitator(s) |
|
10:00 a.m. – 11:15 a.m.
Presentation(s): View Presentation Competency: Strategic Insight and Integration Moderator(s): Stuart Klugman, FSA, CERA, Ph.D. Presenter(s): Chris E. Stehno; Francois Millard, FSA, FIA, MAAA Two signature issues today are the big data and behavioral science revolutions, however they are seldom discussed together. This session will present a framework in which data science and behavioral (“nudge”) science function as two parts of a greater whole. Behavioral science provides an expanded toolset for operationalizing predictive models; it also serves as a wealth of ideas for getting the most out of big digital technology to create innovative, pro-social products and business models. This presentation will sketch basic principles of behavioral economics and choice architecture, and provide examples of innovative data science/behavioral science mash-ups. Track: Manager/Supervisor Session Coordinator(s) Xiaojie Wang, FSA, CERA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Dennis Kapylou, ASA Presenter(s): Kyle Nobbe, FSA, MAAA; Vikram Kamath For many, predictive modeling is simply feeding data into a “black box” and getting output. However, understanding the model development process is critical when it comes to validating model results and implementing learnings in a business setting. This session will cover the process of building a simple model then demonstrating key ways actuaries should validate results. Track: Beginner/Implementer
Session Coordinator(s) Rosmery Cruz Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Rosmery Cruz Presenter(s): Jeff T. Heaton, Ph.D. Bring your laptop to experience this deep learning neural network tool offered free from Google. Track: Advanced Practitioner Session Coordinator(s) Rosmery Cruz Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Randal Olson, Ph.D. Presenter(s): Benjamin Copeland, ASA, MAAA; Joseph Charles Long The growth of huge social networks like Facebook, Twitter and LinkedIn have sparked innovation in graph analytics. Graph analytics is not about data visualization; it is instead about modeling the connections between people. Graph analytics is used to suggest new friends to connect to and what stories you might want to read in these social networks. These same analytic frameworks can be used for novel applications in health care. For example, inferred referral patterns can be used to build a social network graph of health care providers. From there, interesting visualizations and inferences around network leakage or post-acute care can be generated. This session will introduce the concepts behind social network analysis and provide case studies of its application to health care. Track: All Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Competency: Technical Skills & Analytical Problem Solving Moderator(s): Jason Hiquet FSA,CERA Presenter(s): Boyi Xie When we talk about data, one usually thinks of numerical data in a tabular format, however, the development of text mining techniques made it possible for the analyses on unstructured, textual data. Natural Language Processing is a field in Artificial Intelligence (AI) that focuses on data relevant to human languages, such as words and documents. In this session, we will introduce the development of Natural Language Processing, and its application to the insurance industry, such as information extraction from submission documents, identifying events and trends from news articles and classifying lawsuits into loss scenarios for experience studies. Track: Beginner/Implementer Session Coordinator(s) Xiaojie Wang, FSA, CERA Facilitator(s) Presentation(s): View Presentation Competency: Strategic Insight and Integration Moderator(s): Taylor Fay Presenter(s): Rajiv Shah This session focuses on how automated artificial intelligence (AI) and machine learning (ML) are transforming the insurance industry. Through case studies, this session will cover how AI is able to understand otherwise unknown processes better, make it transparent and increase insurability (e.g. quantifying mortality risk of customers with complex medical history or insuring high-frequency claimants through granular risk segmentation). We will also discuss why model transparency and validations are critical components of AI and ML workflow - for both the success of the insurers as well as the well-being of the consumers. Track: Manager/Supervisor Session Coordinator(s) David L. Snell, ASA, MAAA Facilitator(s) |
|
11:15 a.m. – 12:15 p.m.
Join your fellow meeting attendees at this networking lunch. Catch up with colleagues and friends, discuss previous sessions and make new acquaintances. Session Coordinator(s) Facilitator(s) |
|
12:30 p.m. – 1:20 p.m.
Competency: Technical Skills & Analytical Problem Solving Session Coordinator(s) Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Jeff T. Heaton, Ph.D. Presenter(s): Dihui Lai, ASA The world has witnessed great advances in machine learning (ML)/artificial intelligence (AI) technologies. However, most AI systems are developed for general purposes (e.g. face recognition, language translation, voice recognition etc.). AI designed for insurance is less common. This session will demonstrate an AI system customized specifically for underwriting purposes and how it improves the efficiency of underwriting workflow. Track: Beginner/Implementer Session Coordinator(s) Rosmery Cruz Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Gary A Hatfield FSA, MAAA, CERA Presenter(s): Geoffrey R. Hileman, FSA, MAAA; Norman M. Storwick, FSA, MAAA Fuzzy logic is conceptually interesting but can it add value? This session will review potential uses for fuzzy logic and how it can improve actual operations. At the end of this overview, we will demonstrate an actual solution and compare the value of fuzzy logic to other analytic approaches. Track: Advanced Practitioner Session Coordinator(s) Xiaojie Wang, FSA, CERA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Dennis Kapylou, ASA Presenter(s): Chris Wells Python has become the most popular programming language among data scientists while having popularity in many other disciplines as well. Attend this session to learn how to get started in this amazingly versatile language. Track: Beginner/Implementer Session Coordinator(s) Kevin J. Pledge, FSA, FIA Facilitator(s) Competency: External Forces & Industry Knowledge, Results-Oriented Solutions Moderator(s): Michael Cletus Niemerg, FSA, MAAA Presenter(s): Gershon Henoch Firestone, FSA Digitilization of Underwriting has three primary impacts: 1) changing the way business is done due to automation; 2) better understanding of the business; 3) change the way data is used. This session will cover some or all of the following: the process of impact of the three listed above, case studies and the comparison of data available today and in the future and how data utilization will be impacted by tomorrow's data. Track: Manager/Supervisor Session Coordinator(s) Xiaojie Wang, FSA, CERA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Michael Blakeney, FSA, MAAA Presenter(s): Rosmery Cruz Simple mathematical equations outperform very complex ones in predictive analytics (PA) situations. This session will cover why and how to guard against that and other items that might not be intuitive. Track: Beginner/Implementer Session Coordinator(s) Rosmery Cruz Facilitator(s) |
|
1:35 p.m. – 2:25 p.m.
Competency: Results-Oriented Solutions Moderator(s): Anthony Cappelletti, FSA, FCAS, FCIA Presenter(s): Boyi Xie This session will present a predictive analytics (PA) framework for a professional liability portfolio. The business line is facing two challenges: 1) the underwriting procedure is complicated – underwriters need to look at many underwriter triggers that have little indication of future risk; 2) the portfolio loses clients because of an over-complicated application process, (e.g. more than five pages of forms). We rely on predictive analytics (PA) to develop an accelerated underwriting process to better quantify risk. It will help refine underwriting triggers to improve efficiency so underwriters can touch little on low risk policies. We also aim to simplify the application process by asking minimum questions while maintaining accurate risk estimation. Track: Manager/Supervisor Session Coordinator(s) Xiaojie Wang, FSA, CERA Facilitator(s) Competency: Technical Skills & Analytical Problem Solving Moderator(s): Anders Larson, FSA, MAAA Presenter(s): Qichun Xu, FSA This session will cover how with an education background and business experience, an actuary can quickly learn the basics of data science and perform the important modeling work in insurance. Track: Beginner/Implementer Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Presenter(s): Yuanjin Liu, ASA The language R is terse, powerful, free and able to run on several different machine platforms and operating systems. Fluency in R may soon become a selection factor in getting your next job. This session will demonstrate some of the predictive power you can have beyond spreadsheets. It will cover how to get started in the new world of predictive analytics (PA). This session will be a beginner-level tutorial of how to fit, validate and interpret a logistic regression model with R. Track: Beginner/Implementer Session Coordinator(s) David L. Snell, ASA, MAAA Facilitator(s) Competency: Technical Skills & Analytical Problem Solving Moderator(s): Michael Blakeney, FSA, MAAA Presenter(s): Adnan Mohammed Haque A number of different companies have explored different statistical and machine learning techniques for looking at their mortality experience. What have they learned? How did they choose which tools to use and how do they validate this approach vs. the traditional methods of the past? Track: Beginner/Implementer Session Coordinator(s) David L. Snell, ASA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Professional Values Moderator(s): Jordan Durlester Presenter(s): Bradley Joseph Lipic This session will discuss strategy and tactics for becoming good stewards of consumer/customer data. Topics will include various techniques, such as anonymization, pseudonymization, tokenization and the merging of disparate, de-identified data source. Track: Manager/Supervisor Session Coordinator(s) Rosmery Cruz Facilitator(s) Presentation(s): View Presentation Competency: Results-Oriented Solutions Moderator(s): Rosmery Cruz Presenter(s): Jeff T. Heaton, Ph.D. The human race is great at devising many different ways to accomplish the same task; predictive modeling is no different. Data scientists can choose from a wide variety of languages to prepare data, fit models and verify results. The Rosetta Stone, which showed writing composed in three languages side by side, was instrumental in translating unfamiliar Egyptian hieroglyphics into the more common Greek language. This presentation follows a similar approach by showing the same predictive model implemented in the following languages: R, Python, SQL, Excel, MatLab, SAS and Julia. Attendees will compare and contrast each of these languages andbuild a greater understanding of those languages less familiar to them. Track: Advanced Practitioner Session Coordinator(s) Rosmery Cruz Facilitator(s) |
|
2:25 p.m. – 2:40 p.m.
|
|
2:40 p.m. – 3:30 p.m.
Presentation(s): View Presentation Competency: Results-Oriented Solutions Moderator(s): Dorothy Andrews, ASA, MAAA Presenter(s): Cheng-Sheng Peter Wu, ASA, FCAS, MAAA This session will discuss how insurers are applying predictive analytics (PA) to improve their business which will include examples from general Insurance such as auto, home and umbrella. Track: Manager/Supervisor Session Coordinator(s) David L. Snell, ASA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Anders Larson, FSA, MAAA Presenter(s): Michael Cletus Niemerg, FSA, MAAA Interpreting sophisticated models like GBMs or Neural Networks can be a daunting task. This session will focus on ways to help interpret and understand the output of complex models. Track: Beginner/Implementer Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Kimberly M. Steiner, FSA, MAAA Presenter(s): Nicholas Scott Hanewinckel, FSA, CERA This session will cover interactive demonstrations of how to get started (or how to move to the next level) in predictive modeling. Track: Beginner/Implementer Session Coordinator(s) David L. Snell, ASA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Jason Hiquet FSA,CERA Presenter(s): Rajiv Shah Automation of predictive analytics (PA) is changing analytics. This workshop lets the audience try automated analytics. The effects on accuracy, transparency and lowering the barriers for analytic usage will be seen. The presenter will also share how organizations are responding and transforming around automated analytics. Track: Beginner/Implementer Session Coordinator(s) Xiaojie Wang, FSA, CERA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Erica Rode, ASA, MAAA Presenter(s): Sheamus Kee Parkes, FSA, MAAA The cloud inspires awe and fear in equal parts these days; there is a need to learn more, but there’s much information to learn and it may be difficult to know what to learn next. In this session, we’ll carve off a small corner of the cloud and make it your own. Setting up a Data Science Virtual Machine on Digital Ocean will be demonstrated. It’s the fastest way to get a data science development environment setup for your own use. Many of the topics and ideas covered will apply to almost all other cloud offerings. What use cases Digital Ocean is good for and what use cases you should look elsewhere for will also be discussed. Track: Beginner/Implementer Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Results-Oriented Solutions Moderator(s): Rosmery Cruz Presenter(s): Edmond Deuser; Jeff T. Heaton, Ph.D. Most predictive models are intended to be used by more people than the data scientists who produced them, but integrating a predictive model created in R or Python into a larger system presents a challenge. An application programming interface (API) allows models to be easily accessed by other systems, both within and outside of an organization. When using an API, it is critical to consider aspects such as security, accounting, support, service level agreements (SLA) and the access patterns of clients. This presentation will explore these considerations and discuss technical details involved, such as cloud deployment, high availably and data exchange formats. Track: Advanced Practitioner Session Coordinator(s) Facilitator(s) |
|
3:30 p.m. – 3:45 p.m.
|
|
3:45 p.m. – 4:35 p.m.
Presentation(s): View Presentation Competency: Results-Oriented Solutions Moderator(s): Jeevan Duggempudi Presenter(s): Gregory P. Heck, Vincent J. Granieri, FSA, MAAA, EA Predictive models have disrupted the traditional underwriting process. In this session, we will review a decade in the life underwriting of a disruptor – the confluence of actuaries, underwriters and clinicians. In addition, perspectives of other stakeholders will be explored. Track: Manager/Supervisor Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Michael David Hoyer, FSA, MAAA Presenter(s): Michael Cletus Niemerg, FSA, MAAA This session will walk the user through a simple case study to evaluate the practical aspects of building a regression model. Topics that will be covered include: stepwise selection, penalized regression and extensions such as generalized linear models and general additive models. Track: Beginner/Implementer Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Professional Values Moderator(s): Stuart Klugman, FSA, CERA, Ph.D. Presenter(s): Rohan Noel Alahakone, ASA, MAAA; Dorothy L. Andrews, ASA, MAAA Actuarial tools now include both open and closed source software; both need a strong model governance framework to mitigate potential sources of model risk and external threats. This presentation will discuss how to develop, install and monitor a model governance framework to safeguard company modeling assets and intellectual capital. Track: Advanced Practitioner Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Brian Chen Presenter(s): Randal Olson This session will cover what automated machine learning (AutoML) is, why is it important and how can it be integrated into prediction workflows. Track: Manager/Supervisor Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Technical Skills & Analytical Problem Solving Moderator(s): Marshall Lagani Presenter(s): Ben Johnson This session will demonstrate how sufficiency in the R programming language also allows the R programmer to build a web application using R shiny. This session will focus how to present the results of predictive analytics (PA) in an easy-to-use and interactive way. Track: All Session Coordinator(s) Anders Larson, FSA, MAAA Facilitator(s) Presentation(s): View Presentation Competency: Leadership Moderator(s): Bradley Joseph Lipic Presenter(s): Jordan Durlester Now that many organizations have successfully proved the concept of data science as a value generator, the demand curve has hockey-sticked. Scaling data science linearly is not a practical approach, and it must be decided which projects should be worked on immediately, be delayed and be shelved. In this session, the process of developing a data science prioritization framework will be covered; from getting buy-in to implementation best practices. Track: All Session Coordinator(s) Rosmery Cruz Facilitator(s) |
|
5:00 p.m. – 6:00 p.m.
|