Amortized Inverse Modeling Across Tasks from Incomplete, Unpaired Data

Conlain Kelly; Juli Mueller

Amortized Inverse Modeling Across Tasks from Incomplete, Unpaired Data

Conlain Kelly
, Juli Mueller

Computational Science

National Laboratory of the Rockies

Research output: Contribution to conference › Paper

Abstract

This whitepaper describes challenges in inverse modeling for large-scale, DOE-relevant applications. In particular, we outline difficulties associated with incomplete or unpaired observations of an underlying process of interest, and propose future research directions to address the challenges.

Original language	American English
Number of pages	2
State	Published - 2025
Event	ASCR Inverse Problems Workshop - Rockville, Maryland Duration: 10 Jun 2025 → 12 Jun 2025

Conference

Conference	ASCR Inverse Problems Workshop
City	Rockville, Maryland
Period	10/06/25 → 12/06/25

NLR Publication Number

NLR/CP-2C00-97686

Keywords

generative AI
inverse problems
machine learning

Access to Document

https://www.orau.gov/support_files/2025InverseMethods/KellyC.pdf

Cite this

@conference{05df2fb9d8754856a8ccb66dc0facf75,

title = "Amortized Inverse Modeling Across Tasks from Incomplete, Unpaired Data",

abstract = "This whitepaper describes challenges in inverse modeling for large-scale, DOE-relevant applications. In particular, we outline difficulties associated with incomplete or unpaired observations of an underlying process of interest, and propose future research directions to address the challenges.",

keywords = "generative AI, inverse problems, machine learning",

author = "Conlain Kelly and Juli Mueller",

year = "2025",

language = "American English",

note = "ASCR Inverse Problems Workshop ; Conference date: 10-06-2025 Through 12-06-2025",

}

TY - CONF

T1 - Amortized Inverse Modeling Across Tasks from Incomplete, Unpaired Data

AU - Kelly, Conlain

AU - Mueller, Juli

PY - 2025

Y1 - 2025

N2 - This whitepaper describes challenges in inverse modeling for large-scale, DOE-relevant applications. In particular, we outline difficulties associated with incomplete or unpaired observations of an underlying process of interest, and propose future research directions to address the challenges.

AB - This whitepaper describes challenges in inverse modeling for large-scale, DOE-relevant applications. In particular, we outline difficulties associated with incomplete or unpaired observations of an underlying process of interest, and propose future research directions to address the challenges.

KW - generative AI

KW - inverse problems

KW - machine learning

M3 - Paper

T2 - ASCR Inverse Problems Workshop

Y2 - 10 June 2025 through 12 June 2025

ER -

Amortized Inverse Modeling Across Tasks from Incomplete, Unpaired Data

Abstract

Conference

NLR Publication Number

Keywords

Access to Document

Fingerprint

Cite this