Vector Institute Hosts Machine Learning Theory Workshop at University of Waterloo
The Vector Institute for Artificial Intelligence recently convened a Machine Learning Theory Workshop at the University of Waterloo, bringing together leading researchers to discuss foundational aspects of machine learning. Organized by faculty members Shai Ben-David and Ruth Urner, the event featured presentations on mathematical underpinnings of AI, including learnability characterizations and gradient-based optimization. Key speakers included Canada CIFAR AI Chairs Shai Ben-David and Murat Erdogdu, who presented findings on statistical learning dimensions and neural network feature learning respectively. The workshop facilitated interactive discussions and poster sessions, allowing graduate students and established theorists to share insights on complex AI issues. Topics ranged from the limitations of VC dimensions in general statistical learning to the separation between kernel methods and two-layer neural networks. This gathering underscores the critical role of theoretical research in driving paradigm shifts in artificial intelligence problem-solving methodologies, highlighting the collaborative efforts within the Vector community to advance the field's fundamental understanding.
Wire timeline
Vector Institute Hosts Machine Learning Theory Workshop at University of Waterloo
The Vector Institute for Artificial Intelligence recently convened a Machine Learning Theory Workshop at the University of Waterloo, bringing together leading researchers to discuss foundational aspects of machine learning. Organized by faculty members Shai Ben-David and Ruth Urner, the event featured presentations on mathematical underpinnings of AI, including learnability characterizations and gradient-based optimization. Key speakers included Canada CIFAR AI Chairs Shai Ben-David and Murat Erdogdu, who presented findings on statistical learning dimensions and neural network feature learning respectively. The workshop facilitated interactive discussions and poster sessions, allowing graduate students and established theorists to share insights on complex AI issues. Topics ranged from the limitations of VC dimensions in general statistical learning to the separation between kernel methods and two-layer neural networks. This gathering underscores the critical role of theoretical research in driving paradigm shifts in artificial intelligence problem-solving methodologies, highlighting the collaborative efforts within the Vector community to advance the field's fundamental understanding.
Vector Institute for Artificial Intelligence