IRisk Lab Fall 2024 Projects

Cyber risks have been posing increasing concerns to both the public and private sectors, drawing significant attention to the research on these risks. The study of cyber risks inherently involves multiple disciplines and is thus approached from various angles, including actuarial science, information science, computer science, insurance economics, legislation and policymaking. While research from each discipline provides valuable insights, there is a lack of a comprehensive review that outlines the connections among studies in these different fields. The purpose of this project is to conduct a comprehensive review of existing literature on cyber risks, identify the interconnections among studies from different approaches, and provide insights and inspirations for future synergistic research across various disciplines.

Students are expected to collect and organize literature on cyber risk studies and present the findings to the supervisor on a regular basis. Proficiency in scientific reading and presentation is required. A background in actuarial science, statistics, computer science, or information technology is preferred.

Supervisor: Wei Wei
Graduate Supervisor: Jiajie Yang

Machine learning techniques have revolutionized data analysis across various industries, including actuarial science. This project aims to explore and apply machine learning paradigms to actuarial problems using Python and pre-built packages. By leveraging these tools, students will gain practical experience in employing advanced computational methods to solve real-world actuarial challenges.

Students will begin with comprehensive tutorials and workshops on Python-based machine learning libraries. Each topic will span approximately 3-4 weeks, consisting of theoretical study, practical implementation, and project-specific applications.

• Policyholder Classification (Decision Trees, k-means)
  • Objective: Develop models to predict which individuals are likely to purchase term life
    insurance and their expected coverage amounts.
  • Example Dataset: Utilize TermLife.csv to build predictive models based on demographic and
    financial data.
• Optical Character Recognition (OCR, Tesseract)
  • Objective: Convert documents into readable and searchable text.
  • Example Application: Extracting data from claims documents to facilitate easier analysis and processing of insurance claims.
• Sentiment Analysis (TextBlob)
  • Objective: Analyze text to determine attitudes or sentiments.
  • Example Application: Assessing the sentiment of ESG (Environmental, Social, and Governance) investments or analyzing the sentiment within claim documents.
• Image Classification (Convolutional Neural Networks, planet.org)
  • Objective: Apply edge detection techniques on satellite images to monitor changes in coastlines.
  • Example Application: Supporting flood insurance and catastrophe (CAT) bonds by identifying coastal erosion trends.

This project offers students a unique opportunity to bridge theoretical knowledge with practical application in actuarial science using advanced machine learning techniques. Students will gain an appreciation for machine learning and how it can be used to solve problems old and new.

Supervisor: Eric Icaza

The significant impact that extreme weather events have on the financial, insurance, and energy sectors is no longer news, prompting companies to prioritize physical and transition risks along with macroeconomic risk when managing their risk exposure. As climate change intensifies, it is important that companies develop their risk management tools, incorporating climate risk and existing scenario models for their analysis. However, there is no specific model that is widely adopted across sectors to perform scenario analysis related to climate change. With that in mind, this research aims to fill that gap by investigating and developing a model that systematically integrates expert knowledge, research, and data.

This research is divided into two key projects, and a student is expected to work on
one of the projects:

• Impact of Climate Risk on Catastrophe Bonds: This project involves studying how climate risk affects catastrophe bonds. We will conduct a comprehensive literature review, examining existing research on the impact of climate change on these bonds. Additionally, we will explore the various methodologies and approaches widely adopted in the financial markets for climate change scenario analysis.
• Pricing Weather Derivatives: This project focuses on investigating and implementing different pricing models for weather derivatives. The initial step involves a literature review of papers discussing the dynamics of temperature using stochastic models. We will then explore models that effectively capture extreme climate events while maintaining analytical tractability. By pursuing these projects, we aim to investigate and develop models that can aid in managing the multifaceted risks associated with climate change.

Supervisor: Tolulope Fadina

As data-driven decision-making becomes increasingly important in all fields, it is crucial to have a comprehensive understanding of the datasets at our disposal to facilitate research, analysis, and innovation across disciplines. The primary objective of this project is to conduct an extensive data discovery process within UIUC to identify and catalog various datasets that exist across different units, research centers, and government entities. By consolidating these datasets into a centralized repository, we can provide researchers, students, and faculty members with a unified platform to access a wide range of data for their projects and initiatives. We anticipate that this project will require dedicated personnel, access to relevant systems, perform data visualization, creating a relational database. In previous semesters, we prepared an environment to host the database at NCSA and discovered hundreds of datasets. This semester, we will migrate these datasets into the database, allowing students to gain experience with database operations and data cleaning.

Supervisors: Zhiyu (Frank) Quan, Eli O’Donohue
Graduate Supervisor: Jiayi Guo

Actuarial science and data science skills extend beyond traditional areas like insurance. In this project, we explore various data-driven models to enhance pharmaceutical company operations and guide business decision-making. For instance, we use risk control skills to create a sales order approval system, predictive modeling to forecast future sales, and optimization skills to automate logistics and dispatch for delivery trucks.

Luyan is a pharmaceutical company headquartered in Xiamen, Fujian, China, with operations in health product R&D, manufacturing, and distribution. As an industry leader, it is exploring new ways of improving its services provided to its customers.

This is an exploratory project, and the tasks will be performed including but not limited to

• Data wrangling
• Risk management
• Optimization
• Feature engineering using NLP
• Supervised/unsupervised learning

Supervisor: Zhiyu Quan

This project engages students from the University of Illinois in a practical, hands-on experience with PCMI, a leader in software solutions for warranties. The project aims to explore new data product opportunities, develop a business case, and support the design and development of innovative data-driven products.

Goals

1. Identify New Data Product Opportunities: Students will investigate three potential
   areas for data product innovation:
   • Enhancing existing products by integrating analytics directly into them.
   • Developing new standalone data products for current customers.
   • Creating new standalone data products targeted at new market segments.
2. Develop a Business Case: Each team will establish a business case to justify the investment in their proposed data products. This includes market analysis, potential ROI, and strategic alignment with PCMI’s objectives.
3. P&L Model Development: Over the course of the project, students will build a profit and loss (P&L) model for the proposed data products, detailing anticipated revenue, cost, and margin scenarios.
4. Identify Data Augmentation Partners: Teams will research and propose potential partners for data augmentation to enhance the value of PCMI’s data products.
5. Requirements and Solution Design: Students will develop detailed requirements for the approved new data products and collaborate with the PCMI RDR (Research, Development, and Requirements) team on the solution design.
6. Development Backlog Creation: Following Minimum Viable Product (MVP) best practices, students will create a development backlog for a PCMI development team to execute.

Expected Outcomes

• A portfolio of new data product ideas fully explored and documented.
• Comprehensive business cases supporting further investment in new products.
• A robust P&L model for each proposed product.

Deliverables

• Final report detailing research findings, business cases, P&L models, and product
  requirements.
• Presentation to PCMI stakeholders showcasing proposed data products and strategic
  insights.
• Development backlog ready for implementation by PCMI’s development team.

Evaluation

• Quality and innovation of data product ideas.
• Feasibility and thoroughness of the business cases.
• Practicality and detail in the P&L models and development plans.
• Engagement with potential data augmentation partners and early adopters.

This project proposal offers a unique opportunity for students to apply their academic knowledge to real-world challenges, fostering innovation and strategic thinking in the development of new data products at PCMI.

Supervisor: Zhiyu Quan

This R&D project explores the development and assessment of various data-driven financial cash models to manage RGA’s global financial service products including longevity swaps, asset-intensive transactions, and pension risk transfers. Transaction-specific data runs the spectrum from voluminous to sparse. Transactions typically range from $100 million to $10+ billion. Participants will apply actuarial science and data science to develop various proxy models and assess underlying assumptions. Models are used to monitor performance, risk profiles, manage capital, and provide insights into business management. 

The iRisk team will collaborate with RGA Global Financial Services (GFS) Valuation Team Europe- Middle-East-Africa (EMEA) actuaries led by former ARMS instructor, Tim Cardinal. Fall 2024 will focus on learning the business, models, and data, developing problem definitions, and building prototypes. Spring 2025 will focus on operationalizing prototypes and learning insights into models’ strengths and weaknesses as they are applied to recent/new transactions. Without going into details, the work will be engaging, challenging, and exciting.

Students will be required to sign a non-disclosure agreement (NDA). In your application, please indicate your interest in internships and co-ops. Ideally, we would like 1-3 of the iRisk student team to intern during summer 2025 and 1-2 students to co-op in fall and/or spring 2025. RGA is one of the largest global life and health reinsurers and is #223 on the Fortune 500 with offices in dozens of countries around the world from St. Louis and Toronto to London, Dublin, Paris, and Milan to Hong Kong, Shanghai, Beijing, Tokyo, Mubai, and Dubai.

Supervisor: Xiaochen Jing, Frank Quan, Tim Cardinal

As decentralized finance (DeFi) continues to attract investors globally, ensuring the safety and reliability of smart contracts and financial protocols becomes crucial. For insurers, understanding the inherent risks associated with these technologies is essential for developing new products, pricing, and claim processing.

While smart contracts offer remarkable efficiency and reduce costs since these blockchain-based innovations can self-execute contracts without human intervention, they also introduce significant security risks and have become targets for more and more cyber-attacks. On the other hand, they also present unique research opportunities due to the high transparency of DeFi systems, which publicize details on smart contracts, financial protocols, and transactions.

This research project aims to identify potential vulnerabilities within these technologies, develop a framework for riskiness measure, and assess how the associated risks have been priced by DeFi insurance. Throughout the semester, we will review both academic literature and industry reports to understand the architecture of existing smart contracts and financial protocols, their historical vulnerabilities and exploits, and current security considerations. We will then create a comprehensive risk assessment framework that considers factors such as code vulnerabilities, business models, activity levels, and governance. This model will be applied to analyze the current DeFi insurance policies on pricing and claim processing.

Supervisor: Xiaochen Jing

Graduate Supervisor: Zhonghe Wan