AI for Scientists: Accelerating Discovery Through Knowledge, Data, and Learning

Citation

Sun, Jennifer J. (2024) AI for Scientists: Accelerating Discovery Through Knowledge, Data, and Learning. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/d6y8-4590. https://resolver.caltech.edu/CaltechTHESIS:11162023-054627670

Abstract

With rapidly growing amounts of experimental data, machine learning is increasingly crucial for automating scientific data analysis. However, many real-world workflows demand expert-in-the-loop attention and require models that not only interface with data, but also with experts and domain knowledge. My research develops full stack solutions that enable scientists to scalably extract insights from diverse and messy experimental data with minimal supervision. My approaches learn from both data and expert knowledge, while exploiting the right level of domain knowledge for generalization. This thesis presents progress towards developing automated scientist-in-the-loop solutions, including methods that automatically discover meaningful structure from data such as self-supervised keypoints from videos of diverse behaving organisms. We will then discuss methods that use these interpretable structures to inject domain knowledge into the learning process, such as guiding representation learning using symbolic programs of behavioral features computed from keypoints. This work is the result of close collaborations with domain experts, such as behavioral neuroscientists, in order to identify bottlenecks and integrate these methods in real-world workflows. My aim is to enable AI that collaborates with scientists to accelerate the scientific process.

Item Type:

Thesis (Dissertation (Ph.D.))

Subject Keywords:

machine learning; computer vision; AI for science; neurosymbolic models

Degree Grantor:

California Institute of Technology

Division:

Engineering and Applied Science

Major Option:

Computing and Mathematical Sciences

Awards:

Ben P.C. Chou Doctoral Prize in IST

Thesis Availability:

Not set

Research Advisor(s):

Perona, Pietro (co-advisor)
Yue, Yisong (co-advisor)

Thesis Committee:

Bouman, Katherine L. (chair)
Perona, Pietro
Yue, Yisong
Chaudhuri, Swarat
Kennedy, Ann

Defense Date:

15 September 2023

Funders:

Funding Agency	Grant Number
Natural Sciences and Engineering Research Council of Canada	UNSPECIFIED
Amazon AI4Science Fellowship	UNSPECIFIED
Kortchak Scholarship	UNSPECIFIED

Record Number:

CaltechTHESIS:11162023-054627670

Persistent URL:

https://resolver.caltech.edu/CaltechTHESIS:11162023-054627670

DOI:

10.7907/d6y8-4590

Related URLs:

URL	URL Type	Description
https://arxiv.org/pdf/2112.05121.pdf	arXiv	Article for Chapter 4
https://arxiv.org/pdf/2212.07401.pdf	arXiv	Article for Chapter 4
https://arxiv.org/pdf/1912.01001.pdf	arXiv	Article for Chapter 5
https://arxiv.org/pdf/2010.13321.pdf	arXiv	Article for Chapter 5
https://arxiv.org/pdf/2207.10553.pdf	arXiv	Article for Chapter 6
https://arxiv.org/pdf/2107.13132.pdf	arXiv	Article for Chapter 7
https://arxiv.org/pdf/2011.13917.pdf	arXiv	Article for Chapter 9
https://arxiv.org/pdf/2007.12101.pdf	arXiv	Article for Chapter 10
https://arxiv.org/pdf/2106.06114.pdf	arXiv	Article for Chapter 10
https://arxiv.org/pdf/2111.15186.pdf	arXiv	Article for Chapter 11
https://arxiv.org/pdf/2210.05050.pdf	arXiv	Article for Chapter 12

ORCID:

Author	ORCID
Sun, Jennifer J.	0000-0002-0906-6589

Default Usage Policy:

No commercial reproduction, distribution, display or performance rights in this work are provided.

ID Code:

16247

Collection:

CaltechTHESIS

Deposited By:

Jennifer Sun

Deposited On:

02 Dec 2023 00:58

Last Modified:

06 Dec 2023 17:06

Full text not available from this repository.

Repository Staff Only: item control page