Gift-funded Research Projects

2024 Robotics

Call closed as of May 2024

The Science Hub seeks proposals to advance the state of the art in Robotic Manipulation and associated AI, especially addressing the challenging scenarios of industrial logistics robots. Example topics of interest include:

• Robotics Foundation Models — Architectures and training methods to create machine learning models that understand the physics of objects in contact, based on many examples of observing robot actions.
• Fine, Compliant Manipulation — Handling of items in dense clutter, including a wide mixture of object types and the need for multi-touch interactions to retrieve or store objects.
• Simulation — Techniques for high-fidelity simulation of deformable and articulated objects (e.g., books, bagged clothing, etc.), including those that learn from real-world interactions.
• Video AI for Robotics — Machine learning for understanding a robotic action’s outcome/effects or for monitoring the likely success of a robotic behavior, based on video.
• Mobile Manipulation — Methods for planning and control of mobile manipulators, such as required for quickly transferring objects across a room or loading/unloading densely packed boxes

All UW researchers are welcome to apply. The application includes basic demographic questions, and requests an uploaded copy of your CV and a 2 to 4-page research proposal (references are not included in the 4-page limit). The proposal should include a short description of the budget request of up to $100k for one-year. Awards are structured as unrestricted gifts to the principal investigator. Each gift funded proposal will be subject to a 5% assessment per UW Administrative Policy (APS) 36.2. No third parties should be involved in the projects. Per default, project results will be made available to the general public through publication and open source

2023 Robotics

Call closed as of February 2023

• The Science Hub seeks proposals to advance the state of the art and enable new robotics and AI solutions, in the following topic areas:
  • Robotic Manipulation — for example: picking and packing objects from/to cluttered, confined spaces; robust grasping for a wide variety of many object types, including fragile, deformable, and deconstructable (e.g., lidded shoe boxes); modeling and prediction for fast compliant control
  • Computer Vision — for example: robust 3D scene understanding with low-cost sensors; tracking and predicting the 3D poses of dynamic actors; object damage detection from images
  • Machine Learning — for example: continual learning; sample-efficient model (re)training; applications of ML to robotic planning and control 
  • Modeling, Simulation, and Verification — for example: modeling and simulation of deformable objects in contact; closing sim-to-real gaps in synthetic imagery; design of experiments and verification of autonomous systems
  • Autonomous Mobility — for example: safe and fluent navigation in complex dynamic environments; multi-agent planning and coordination (10s–100s of robots); localization and mapping

2022 Robotics

Call Closed as of April 2022

• The Science Hub seeks proposals to advance the state of the art and enable new robotics and AI solutions, in the following topic areas:
• • Robotic Manipulation
  • Autonomous Mobility 
  • Computer Vision
  • Machine Learning
  • Modeling, Simulation, and Verification
  • Human-Robot Interaction

Previous Awardees

Call closed as of November 2022

This call solicits gift-funded research proposals with innovative approaches to designing, building, or operating large-scale database services or distributed cloud services with a strong data management component. Proposals should seek to advance the state-of-the-art in database management and distributed systems in the following topic areas:

1. Intelligent database services in which machine learning techniques are applied to optimize the configuration or operation of the system, including efficient resource management.
2. Global database services where data is distributed and replicated in multiple regions of the world while the system ensures consistency, performance, and data sovereignty.
3. Correct database services whose design and operation includes formal methods, model checking, fuzzing, and other techniques for ensuring that the system performs according to a specification.
4. Private and secure database services that incorporate homomorphic encryption, secure multi-party protocols, differential privacy or other techniques to protect data at rest while permitting complex queries.
5. Real-time database services for predictive analytics where frequently ingested data feeds into continuous queries.

Learn More

Healthcare 2024

Call closed as of May 2024

The Science Hub seeks proposals to advance the state of the art in healthcare and associated AI in the following topic areas:

• AI in Clinical and Primary Care Settings
• Trustworthy and Responsible AI for Healthcare
• AI for Autonomous Research in Health (knowledge development for prevention, diagnosis, and treatment)
• Automation and Optimization of Prescription Transfers (across pharmacies) (ML/Reinforcement Learning/Gen-AI for message generation)
• Prescription Recommendation and Next-Period Prediction Based on Multimodal Personalized Medical Data and Static Information (e.g., demographics)
• Prescription Fraud (and anomaly detection) Based on Multimodal Personalized Medical Data and Static Information (e.g., demographics)

All UW researchers are welcome to apply. The application includes basic demographic questions, and requests an uploaded copy of your CV and a 2 to 4-page research proposal (references are not included in the 4-page limit). The proposal should include a short description of the budget request of up to $100k for one-year. Awards are structured as unrestricted gifts to the principal investigator. Each gift funded proposal will be subject to a 5% assessment per UW Administrative Policy (APS) 36.2. No third parties should be involved in the projects. Per default, project results will be made available to the general public through publication and open source.