Research
My research interests lie in studying situations in which machine learning algorithms fail, including problems such as, uncertainty quantification,
anomaly detection, distribution shifts, and out-of-distribution detection.
|
|
Semi-supervised sequence modeling for improved behavioral segmentation
bioRXiv, 2021
Advised by Prof. Liam Paninski and
Matthew Whitway
@ the Zuckerman Institute
bioRXiv
/
code
We introduce a semi-supervised sequence model for behavioural segmentation that takes advantage of weak- and self-supervision.
This approach displays superior performance when including a large number of unlabeled frames in the regime of sparse hand labels,
and including a small number of hand labeled frames in an unsupervised setting.
|
|
Community implications for gun violence prevention during co-occurring pandemics
PubMed, 2022
Advised by Prof. Shih-Fu Chang
and Prof. Chris Thomas
@ the DVMMLab
PubMed
We develop a natural language processing (NLP) and computer vision (CV) data science system that leverages Twitter data
with multimodal machine learning to automatically identify tweets related to expression of loss and aggression.
|
My Research Notes (On-Going)
|
|
Quantile Risk Control: A Flexible Framework for Bounding the Probability of High-Loss Predictions
J. C. Snell, T. P. Zollo, Z. Deng, T. Pitassi, and R. Zemel, “Quantile Risk Control: A Flexible Framework for Bounding the Probability of High-Loss Predictions.” arXiv, Dec. 27, 2022. Accessed: Dec. 17, 2023. [Online].
Available: http://arxiv.org/abs/2212.13629
Learn then Test: Calibrating Predictive Algorithms to Achieve Risk Control
A. N. Angelopoulos, S. Bates, E. J. Candès, M. I. Jordan, and L. Lei, “Learn then Test: Calibrating Predictive Algorithms to Achieve Risk Control.” arXiv, Sep. 29, 2022. Accessed: Dec. 18, 2023. [Online]. Available:
http://arxiv.org/abs/2110.01052
Distribution-free, Risk-controlling Prediction Sets
S. Bates, A. Angelopoulos, L. Lei, J. Malik, and M. Jordan, “Distribution-free, Risk-controlling Prediction Sets,” J. ACM, vol. 68, no. 6, pp. 1–34, Dec. 2021, doi: 10.1145/3478535.
Conformal Risk Control
A. N. Angelopoulos, S. Bates, A. Fisch, L. Lei, and T. Schuster, “Conformal Risk Control.” arXiv, Apr. 29, 2023. Accessed: Dec. 14, 2023. [Online]. Available:
http://arxiv.org/abs/2208.02814
The Nonexistence of Certain Statistical Procedures in Nonparametric Problems
R. R. Bahadur, L. J. Savage, "The Nonexistence of Certain Statistical Procedures in Nonparametric Problems." The Annals of Mathematical Statistics, 27 (4)
1115-1122 December, 1956. https://doi.org/10.1214/aoms/1177728077
Conformal Prediction
A. N. Angelopoulos and S. Bates, “A Gentle Introduction to Conformal Prediction and Distribution-Free Uncertainty Quantification.” arXiv, Dec. 07, 2022. Accessed: Dec. 14, 2023. [Online]. Available:
http://arxiv.org/abs/2107.07511
|
|
CME 102: Ordinary Differential Equations
Fall 2023
Analytical and numerical methods for solving ordinary differential equations arising in engineering applications.
|
|
MATH 1102: Calculus II
Fall 2021
Methods of integration, applications of the integral, Taylors theorem, infinite series.
|
Sometimes I like to draw and/or paint!
I also dabble in composing. Click on the title below to listen!
| |
|
Blackbird
2018
During senior year of high school, I was attacked by a blackbird while walking to school.
It was a memorable experience, which I decided to depict in my composition.
In the beginning, the calls of the bird (strings playing pizz.) were subtle nuisances.
As I walked along the road, the calls became more distinct (winds and percussion joining in).
Out of nowhere, the bird came swooping onto me and started biting my hair (trumpet and trombone solo),
to which I responded by brushing it away (violins). At that point, I was extremely traumatized
(section B, eerie and gloomy character), but I continued walking (viola – col legno).
Just as I began to relax, the bird came again (section C)! I hurried and ran away (clarinet’s and bassoon’s fast figures)
but the bird chased after me (violins)... And the story continues.
|
|
|