Nicholas Carolan

Github / Google Scholar / LinkedIn / Twitter / Substack

I'm an engineer and researcher working on machine learning and AI safety, currently at Google. Previously, I was an Anthropic Fellow in AI Safety at the Constellation Institute. In 2024, I was part of a Google Research team that won a CVPR best paper award. I studied computer science at Amherst College.

Publications

Rich Human Feedback for Text-to-Image Generation
Youwei Liang, Junfeng He, Gang Li, Peizhao Li, Arseniy Klimovskiy, Nicholas Carolan, Jiao Sun, Jordi Pont-Tuset, Sarah Young, Feng Yang, Junjie Ke, Krishnamurthy Dj Dvijotham, Katie Collins, Yiwen Luo, Yang Li, Kai J Kohlhoff, Deepak Ramachandran, Vidhya Navalpakkam
CVPR 2024 · Best Paper Award, Oral Presentation
[ArXiv] [Blog]

Projects

Minimal Scaling Laws

Derived Chinchilla-style transformer scaling laws in a minimal setup using small models and synthetic data, all trained in a single T4 Google Colab notebook.

Mathematics for Machine Learning (COS 302)

Independent study of Princeton's COS 302 Mathematics for Machine Learning, with all problem sets and programming assignments.

These Places Do Not Exist

Fake census tracts for NeurIPS 2021 Workshop on ML for Creativity and Design.