Topics of Interest
American Family wants new ideas and tools in these areas.
This is an accordion element with a series of buttons that open and close related content panels.
Behavioral economics and applied cognitive science
- Analytics driven “choice architecture”
- Cognitive bias identification
- Information overload
- Customer decision making
- Computational cognitive models
Signal processing and data fusion
- Computer vision
- Internet of Things (IoT) applications
- Speech recognition
- Natural language processing (NLP) and natural language understanding (NLU)
- Aerial satellite imagery (e.g., drones, GIS)
- Temporal data analytics
- Autonomous vehicles
- Weather prediction modeling
Bias, ethics and robustness
- Bias and ethics in AI models
- Bias and ethics in human-driven processes
- Model robustness
- Adversarial AI
Robotic process automation
- Customer Service automation
- Conversational agents
- Recommender systems
Optimization and operations research
- Scalable optimization
- Optimal decision-making
Knowledge Representation
- Knowledge driven information extraction
- Data-driven knowledge construction/expansion
- Inference in knowledge graphs
- Metadata organization and management
Data management and integration
- Entity matching
- Data cleaning
- Automation of data workflows
- Advances in exact, transform, load (ETL) or exact, load transform (ELT)
- Efficient data storage and retrieval
Fraud, insurance risk, marketing and claims analytics
- Claims fraud detection
- Customer retention and valuation models
- Real-time risk prediction during catastrophic events
- Digital underwriting
- Novel approaches to pricing
Artificial Intelligence (AI) and data science are becoming a core component of the insurance industry, motivated by the increasing availability of data, including masses of unstructured data, scalable cloud computing, exponential growth in computing power, new modeling tools, and the ongoing drive for operational efficiency and cost reduction. With big data and new data sources, such as social media, Internet of Things (IoT), telematics, web logs, video and images, and clickstreams, the opportunity to apply artificial intelligence has never been greater across areas of insurance operation. American Family has partnered with UW through the American Family Insurance Data Science Institute (DSI) to offer this unique opportunity.
DSI is committed to working with researchers in all disciplines on campus and encourages anyone interested to apply. If you have questions or would like to discuss your project, please contact research@datascience.wisc.edu.
APPLY FOR FUNDING
American Family Insurance is interested in sponsoring research and applications of data science, artificial intelligence (AI), and machine learning (ML) that is directly or indirectly related to the insurance industry. Topics of interest include, but are not limited to those listed on the left.
DEADLINES
The first call for proposals is now closed.
ELIGIBILITY
You must be a UW–Madison faculty member or a UW–Madison permanent PI. CHS faculty and academic staff without permanent PI status may participate in applications as co-PIs or co-investigators.
ABOUT THE APPLICATION PROCESS
Applications should explicitly address innovation, significance, risk, and impact. Funding will be provided for 1–2 years, depending on the needs and scope of the project. The average award will be approximately $75,000, with a maximum award of $150,000.
Applications should be submitted to the American Family Data Science Institute using WISPER. Specific instructions for created the WISPER document can be found here.
Complete Applications will include:
- A Statement of Work limited to 2 pages that Includes the following:
- A Description of Project
- What are the problem(s) you are trying to solve?
- How is the problem relevant to one or more of the Topics of Interest to American Family?
- What are the innovative features, the potential impact and the significance of the research
- How will the project be transformative for the field? Why is this research important?
- A List of Deliverables: These should be tangible deliverables. Examples include: code, reports (multiple or final), models, etc. Each deliverable should be noted as standard or proprietary based on the definitions found here.
- A Description of Project
-
- A Timeline and List of Milestones for Research Related Activities: These should link back to the deliverables. Examples could include: Kick off meeting, monthly or quarterly check-ins or demo’s, final delivery, etc
- A Detailed Budget Request, indicating the individual staff to be supported and other categories of funding that will be needed. Note that normal indirect costs should be included within the budget limits. Current UW negotiated indirect cost rate is 55%. We are using a MTDC budget model.
Progress reports will be required at regular intervals throughout the period of support. The reports will provide a summary of the progress that was accomplished on the project to that point. A final report will be required within 3 months of the end of the final budget period. The final report should include citations to any published papers or conference presentations that derive from the research undertaken or the equipment purchased and any new grants that were awarded or submitted by the research team.