- Pauching Yap, PhD candidate at University College London (UCL) Centre for Artificial Intelligence, has been awarded a grant by research and development company, GoodAI, for her project that aims to insert Badger principles into an existing state-of-the art continual meta-learning framework.
- She will use Bayesian Online Meta-Learning (BOML) framework to support continual learning while avoiding catastrophic forgetting.
- She aims to extend the framework to support dynamic architectures for a wide range of tasks and to handle sequential task learning on different neural network topologies.
GoodAI has awarded a research grant to Pauching Yap, PhD candidate at UCL Centre for Artificial Intelligence, who aims to create AI that can continually acquire knowledge in different domains as well as utilize past experiences to quickly adapt to new unseen tasks.
Both continual learning, the ability to continually learn new things while avoiding catastrophic forgetting, and gradual learning, the ability to acquire new skills and solve new unseen tasks using previously acquired skills, are key problems in GoodAI’s Badger architecture. They are both problems of keen interest in the AI research world.
Pauching will take a novel approach to these problems, using the method of Bayesian Online Meta-Learning (BOML), which addresses the catastrophic forgetting problem on sequential tasks. The project will aim to extend the BOML framework using badger architecture principles to enable generalization to neural network topologies not seen during training, thus allowing adaptation to tasks that vary in input size, or enable solving tasks from various domains. Furthermore, the BOML framework will be adjusted to fit into GoodAI’s Badger architecture by encouraging communication between experts for a quick task adaptation.
The grant is part of the ongoing GoodAI Grants initiative, which has awarded over $600,000 so far. The initiative is supporting research groups across the world that are solving problems related to GoodAI’s Badger architecture. The vision is that each research grant, along with the work of the GoodAI research team, will contribute in some way to basic AI research, and all together fill in some of the gaps in the roadmap to advanced, increasingly human-like AI.
Pauching said: “This is a great opportunity to take my research further whilst collaborating with the research team at Good AI. ”
Marek Rosa, CEO, and CTO of GoodAI said: “The research team at GoodAI has been following Pauching’s work with interest. We believe she is taking a novel approach to solving the issues of continual learning and gradual learning and are excited to see where her research leads.”
Collaborate with us
For the latest from our blog sign up for our newsletter.