 |
Joshua P. Gardner |
I am a research scientist on the foundation modeling team at Apple. Our team develops the fundamental research insights, models, and methods that drive Apple Foundation Models and Apple Intelligence.
I received a PhD in computer science from the University of Washington's Paul G. Allen School of Computer Science & Engineering, where I was fortunate to be advised by Ludwig Schmidt. I hold an M.S. in Applied Statistics and an M.S. in Information Science from the University of Michigan. I also hold a B.A. with Highest Honors in Philosophy from the University of Michigan.
My research focuses on empirical machine learning: designing controlled experiments to understand artificial intelligence/machine learning models, and using the insights from this experimentation to develop improved models. My recent research applies this lens to the training and fine-tuning of large "foundation" models. I am particularly interested in the impact of data on machine learning models, and how we can develop models that understand new kinds of data or reason across modalities. I have studied a diverse set of problems under this general theme, including those related to tabular data, large language models, multimodal learning, music and audio, and federated and collaborative learning.
Selected Publications
For a full list of publications see my research page or Google Scholar profile.
-
Large Scale Transfer Learning for Tabular Data via Language Modeling
Josh Gardner, Juan C. Perdomo, Ludwig Schmidt.
Neural Information Processing Systems (NeurIPS) 2024.
[arxiv] [code] [model + data]
-
DataComp-LM: In search of the next generation of training sets for language models
Neural Information Processing Systems (NeurIPS) 2024 (Datasets & Benchmarks Track).
Jeffrey Li, Alex Fang, Georgios Smyrnis, Maor Ivgi, [...] Josh Gardner, [...], Achal Dave, Ludwig Schmidt, Vaishaal Shankar. (59 total authors)
[arxiv] [web]
-
LLark: A Multimodal Instruction-Following Language Model for Music
Josh Gardner, Simon Durand, Daniel Stoller, Rachel Bittner.
International Conference on Machine Learning (ICML) 2024.
[arxiv] [code] [web] [blog]
-
Benchmarking Distribution Shift in Tabular Data with TableShift
Josh Gardner, Zoran Popović, Ludwig Schmidt.
Neural Information Processing Systems (NeurIPS) 2023 (Datasets & Benchmarks Track).
[arxiv] [code] [web]
-
Cross-Institutional Transfer Learning for Educational Models: Implications for Model Performance, Fairness, and Equity
Josh Gardner, Renzhe Yu, Quan Nguyen, Christopher Brooks, Rene Kizilcec.
ACM Conference on Fairness, Accountability, and Transparency (ACM FAccT) 2023.
[pdf] [arxiv] [code]
-
Subgroup Robustness Grows on Trees: An Empirical Baseline Study
Josh Gardner, Zoran Popović, Ludwig Schmidt.
Neural Information Processing Systems (NeurIPS) 2022.
[arxiv] [code]
-
OpenFlamingo: An Open-Source Framework for Training Vision-Language Models with In-Context Learning
Anas Awadalla, Irena Gao, Josh Gardner, Jack Hessel, Yusuf Hanafy, Wanrong Zhu, Kalyani Marathe, Yonatan Bitton, Samir Gadre, Jenia Jitsev, Simon Kornblith, Pang Wei Koh, Gabriel Ilharco, Mitchell Wortsman, Ludwig Schmidt.
[arxiv] [blog] [code]
-
MT3: Multi-Task Multitrack Music Transcription
Josh Gardner, Ian Simon, Ethan Manilow, Curtis Hawthorne, Jesse Engel.
International Conference on Learning Representations (ICLR) 2022.
Spotlight Presentation (top 6.7% of submissions)
[arxiv] [web] [blog] [code]
-
Evaluating the Fairness of Predictive Student Models Through Slicing Analysis
Josh Gardner, Christopher Brooks, and Ryan Baker.
International Conference on Learning Analytics and Knowledge (LAK) 2019.
Best Paper Award
[pdf]