Dr. Alice Smith

Developing theoretical foundations for deep learning algorithms, with focus on generalization bounds and optimization landscapes

• Non-convex optimization in neural networks
• Generalization bounds for large language models

Ensuring AI systems behave safely and in accordance with human values

• Mechanistic interpretability in transformers
• Reward modeling for human feedback

Building systems that can understand and generate human language

• Improving reasoning in large language models
• Multimodal understanding and generation