Creating robots to help older adults
How can robot assists Older Adults in achieving and maintaining their sense of meaning and purpose in life? A team led by Profs. Selma Sabanovic and David Crandall is developing a robot that leads natural language conversations with people to help understand what is meaningful to them, and then provides suggestions for activities that will enhance OAs’ own capabilities. The project, funded by Toyota Research Institute, brings a participatory approach and a cross-cultural dimension to the design of robots to support aging.
Learn more
Developing trustworthy AI
As AI becomes increasingly common, people will need to know when they can trust an AI system. This project, a $15 million, 3-year collaboration with Purdue, IU, IUPUI, and Naval Surface Warfare Center Crane Division, is investigating how to develop AI technology that can be trusted, and how do we make sure that people develop the appropriate level of trust in AI. The project is also developing the next generation of AI talent by involving several dozen students at all levels -- undergrad, masters, PhD -- in hands-on research. The IU project is led by Prof. David Crandall and 11 other Co-PIs in Computer Science, Informatics, Engineering, Psychology, and Statistics.
Learn more
Supporting musicians with AI symphonies
Few musicians ever have the chance to solo with a live orchestra. But Prof. Chris Raphael's music accompaniment system, the Informatics Philharmonic, gives them the next best thing: an AI-based system that creates the orchestra accompaniment in real time by listening, following, and learning in real-time. The system listens through a hidden Markov model that tracks the player's progress through a musical score, predicts future evolution based on what has been heard, and adapt to the particular player on each different piece. The system is used in music conservatories in the US and China.
Learn more about the Informatics Philharmonic
Making speech recognition work better -- for everyone
Many modern devices from "smart" hearing aids to virtual assistants like Alexa rely on understanding what people are saying, but there's a big problem: they understand some people much better than others. Prof. Minje Kim's work on speech enhancement, funded by his NSF CAREER award, is working to change that by developing personalized speech enhancement models. His techniques learn to adapt themselves to a particular person's voice and environment, while still respecting their privacy. His technique also requires much less computation, which makes it work on low-power devices.
Learn more about Prof. Kim's Personalized Speech Enhancement
Responding to emergencies with robots
Could robots respond to emergencies, such as rescuing people from burning homes or remote places monitor dangerous environments, without putting human first responders at risk? Prof. Lantao Liu's Vehicle Autonomy and Intelligence Lab is developing robot systems that can navigate through difficult environments -- land, sea, or air. His lab's work includes algorithms that quickly map an unknown environment, and that can plan safe routes through an area that is littered with dangerous obstacles.
Learn more about the Vehicle Autonomy and Intelligence Lab
Finding cyberbullies automatically
Cyberbullying and abusive postings on social media have been found to be detrimental to people's mental health -- especially for children and teenagers. Prof. Sandra Kuebler, an expert in Computational Linguistics, is studying how to detect these abusive posts automatically. This is a challenging problem both because abuse can take on many forms, and because it is difficult to collect representative datasets for training abuse detection algorithms. Her work is investigating how to collect datasets and train these algorithms in a fair, unbiased way.
Learn about Prof. Kuebler's work