Our ‘Unlearning’ Research for GenAI Highlighted in News of Cockrell School of Engineering

Excited to see our collaborative research with JPMorgan Chase on an AI unlearning algorithm spotlighted in news of Cockrell School of Engineering! Check out the article at OpenReview.

Key Points

  • Unlearning for AI: We have created an algorithm that allows generative AI models to “unlearn” information. This is especially important for trustworthy and reliable GenAI models.
  • Why it’s significant: This is a novel direction in AI research, which focuses on teaching AI systems to forget. This unlearning approach helps ensure AI doesn’t store potentially harmful or unnecessary data.
  • How it works: The algorithm teaches the AI model to forget unwanted details, allowing for alterations to images while still maintaining the image’s core identity.
  • Potential applications: This technology could be used to:
    • Remove sensitive information from images before sharing them.
    • Protect privacy
    • Combat the spread of misinformation or deepfakes
    • Avoid copyright infringement

Our ‘Unlearning’ Research for GenAI Featured in The Daily Texan Post

Excited to see our collaborative research with JPMorgan Chase on an AI unlearning algorithm spotlighted in The Daily Texan! Check out the article at OpenReview.

 

Key Points

  • Unlearning for AI: We have created an algorithm that allows generative AI models to “unlearn” information. This is especially important for trustworthy and reliable GenAI models.
  • Why it’s significant: This is a novel direction in AI research, which focuses on teaching AI systems to forget. This unlearning approach helps ensure AI doesn’t store potentially harmful or unnecessary data.
  • How it works: The algorithm teaches the AI model to forget unwanted details, allowing for alterations to images while still maintaining the image’s core identity.
  • Potential applications: This technology could be used to:
    • Remove sensitive information from images before sharing them.
    • Protect privacy
    • Combat the spread of misinformation or deepfakes
    • Avoid copyright infringement
Edge AI 1600 x 1100

Edge AI

EdgeAI refers to the ability to run various AI applications directly on edge devices, hence minimizing or even eliminating the need to rely on the cloud. Given its huge potential to enable new opportunities for various IoT applications (e.g., image classification, object detection, autonomous driving, language processing, etc.), edge computing/IoT is currently one of the hottest research areas. Our research is primarily focused on developing new energy-aware machine learning techniques and hardware prototypes that leverage the network and the system characteristics to enable edge/IoT computing.

Read more
networks image

Networks

Networks are all around us. As such, network science is crucial for our understanding of many applications of high societal relevance (e.g., social and technological networks, epidemics, biological networks). Our research focuses on developing new machine learning methods to discover complex interactions and collective behaviors that determine how various types of events and behaviors in social networks are generated and propagated. In particular, we are interested in developing new approaches for social sensing that are relevant to the immediate concerns around pandemic detection and mitigation.

Read more
systems image

Systems

Futurists often speak of society’s inevitable technological “singularity,” a point in the near future where computers will become ubiquitous units, seamlessly integrated in everyday objects. This trend is already being foreshadowed by manycore processing via the network-on-chip approach, a novel paradigm which implements on-chip networks that enable platforms with extreme parallel capabilities. Our group seeks to develop new machine learning, optimization, and resource management techniques which can enable such a fundamental shift for energy-efficient, cost-effective, large-scale distributed computational platforms for both embedded and high-performance applications.

Read more