Jonas Mueller jonasmueller á csail.mit.edu Chief Scientist @ Cleanlab Hiring AI researchers and interns! Please email me.

---

I am Chief Scientist and co-founder at Cleanlab, building a reliability layer for AI. Previously, I was a senior scientist at Amazon Web Services, devising AutoML & Deep Learning Services used by thousands of enterprises. Before that, I completed my Ph.D. at the MIT Computer Science & Artificial Intelligence Lab. I've also co-developed some of the most popular software for AutoML (autogluon) and for Data-Centric AI (cleanlab).

---

Select Publications
[Full list at Google Scholar]


Quantifying uncertainty in answers from any language model and enhancing their trustworthiness
J Chen, J Mueller
Proceedings of the Association for Computational Linguistics (ACL), 2024
[Featured in MIT Technology Review, MarkTechPost, Unite.ai, Nvidia Blog, LlamaIndex] [Blog Post] [Code]


Task-Agnostic Continual Reinforcement Learning: Gaining Insights and Overcoming Challenges
M Caccia, J Mueller, T Kim, L Charlin, R Fakoor
Conference on Lifelong Learning Agents (CoLLAs), 2023


Deep Learning for the Partially Linear Cox Model
Q Zhong, J Mueller, J Wang
Annals of Statistics, 2022


CROWDLAB: Supervised learning to infer consensus labels and quality scores for data with multiple annotators
HW Goh, U Tkachenko, J Mueller
NeurIPS Human in the Loop Learning Workshop, 2022
[Code to run method] [Code to reproduce results] [Blog Post] [Lecture]


ResNeSt: Split-Attention Networks
H Zhang, C Wu, Z Zhang, Y Zhu, Z Zhang, H Lin, Y Sun, T He, J Mueller, R Manmatha, M Li, A Smola
IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2022
[Featured in Synced, Analytics India Magazine, Pytorch Image Models, Top Kaggle Solutions] [Code]


Pervasive Label Errors in Test Sets Destabilize Machine Learning Benchmarks
C Northcutt, A Athalye, J Mueller
Advances in Neural Information Processing Systems (NeurIPS), 2021
[Featured as Oral Presentation and in Wired, VentureBeat, MIT Technology Review, CNN News18] [Webpage] [Code] [Blog]


Benchmarking Multimodal AutoML for Tabular Data with Text Fields
X Shi^*, J Mueller^*, N Erickson, M Li, A Smola
Advances in Neural Information Processing Systems (NeurIPS), 2021
[Featured as Contributed Talk at ICML Workshop on Automated Machine Learning] [Code] [Benchmark Datasets]


Overinterpretation reveals image classification model pathologies
B Carter, S Jain, J Mueller, D Gifford
Advances in Neural Information Processing Systems (NeurIPS), 2021
[Featured in TechXplore, VentureBeat, World Economic Forum, NY Tech Media] [Lecture Notes] [Code]


Continuous Doubly Constrained Batch Reinforcement Learning
R Fakoor, J Mueller, K Asadi, P Chaudhari, A Smola
Advances in Neural Information Processing Systems (NeurIPS), 2021
[Code]


Fast, Accurate, and Simple Models for Tabular Data via Augmented Distillation
R Fakoor^*, J Mueller^*, P Chaudhari, A Smola
Advances in Neural Information Processing Systems (NeurIPS), 2020
[Lecture Notes] [Code]


AutoGluon-Tabular: Robust and Accurate AutoML for Structured Data
N Erickson^*, J Mueller^*, A Shirkov, H Zhang, P Larroy, M Li, A Smola
ICML Workshop on Automated Machine Learning, 2020
[Featured in VentureBeat, InfoWorld, SiliconAngle, Synced, SD Times, Forbes, Top Kaggle Solutions] [KDD Tutorial] [Code]


Antibody Complementarity Determining Region Design Using High-Capacity Machine Learning
G Liu^*, H Zeng^*, J Mueller, B Carter, Z Wang, J Schilz, G Horny, M Birnbaum, S Ewert, D Gifford
Bioinformatics, 2020
[Featured in NYU Center for Data Science, Protein & Antibody Engineering Summit Keynote]


Low-Rank Bandit Methods for High-Dimensional Dynamic Pricing
J Mueller, V Syrgkanis, M Taddy
Advances in Neural Information Processing Systems (NeurIPS), 2019
[Code]


What made you do this? Understanding black-box decisions with sufficient input subsets
B Carter^*, J Mueller^*, S Jain, D Gifford
Artificial Intelligence and Statistics (AISTATS), 2019
[Featured as Contributed Talk at NIPS Workshop on Interpretability and Robustness] [Slides] [Lecture Notes] [Code]


A peninsular structure coordinates asynchronous differentiation with morphogenesis to generate pancreatic islets
N Sharon^*, R Chawla^*, J Mueller, J Vanderhooft, J Whitehorn, B Rosenthal, M Gurtler, R Estanboulieh, D Shvartsman, D Gifford, C Trapnell, D Melton
Cell, 2019
[Part of effort leading to a cure for Type 1 Diabetes] [Featured as a Research Highlight in Nature Reviews Endocrinology] [Video]


Modeling Persistent Trends in Distributions
J Mueller, T Jaakkola, D Gifford
Journal of the American Statistical Association (JASA), 2018
[Journal Version] [Code]


Sequence to Better Sequence: Continuous Revision of Combinatorial Structures
J Mueller, D Gifford, T Jaakkola
International Conference on Machine Learning (ICML), 2017
[Slides] [Code]


Learning Optimal Interventions
J Mueller, D Reshef, G Du, T Jaakkola
Artificial Intelligence and Statistics (AISTATS), 2017
[Featured as Contributed Talk at NIPS Workshop on Bayesian Optimization] [Slides] [Code]


Siamese Recurrent Architectures for Learning Sentence Similarity
J Mueller, A Thyagarajan
Proceedings of the 30th AAAI Conference on Artificial Intelligence (AAAI), 2016
[Featured in Most Influential AAAI Papers] [Lecture Notes] [Code]


Principal Differences Analysis: Interpretable Characterization of Differences between Distributions
J Mueller and T Jaakkola
Advances in Neural Information Processing Systems (NeurIPS), 2015
[Featured in IEEE Signal Processing Society, Foundations of Geometric Methods in Data Analysis] [Slides] [Code]


General Triallelic Frequency Spectrum Under Demographic Models with Variable Population Size
P Jenkins, J Mueller, Y Song
Genetics, 2014
[Featured as a Research Highlight in Nature Reviews Genetics]


Flexible models for understanding and optimizing complex populations
J Mueller
Ph.D. Thesis, MIT Department of Electrical Engineering and Computer Science, 2018

---

^* Equal Contribution