-
Semi-supervised Learning of Pushforwards For Domain Translation & Adaptation
Authors:
Nishant Panda,
Natalie Klein,
Dominic Yang,
Patrick Gasda,
Diane Oyen
Abstract:
Given two probability densities on related data spaces, we seek a map pushing one density to the other while satisfying application-dependent constraints. For maps to have utility in a broad application space (including domain translation, domain adaptation, and generative modeling), the map must be available to apply on out-of-sample data points and should correspond to a probabilistic model over…
▽ More
Given two probability densities on related data spaces, we seek a map pushing one density to the other while satisfying application-dependent constraints. For maps to have utility in a broad application space (including domain translation, domain adaptation, and generative modeling), the map must be available to apply on out-of-sample data points and should correspond to a probabilistic model over the two spaces. Unfortunately, existing approaches, which are primarily based on optimal transport, do not address these needs. In this paper, we introduce a novel pushforward map learning algorithm that utilizes normalizing flows to parameterize the map. We first re-formulate the classical optimal transport problem to be map-focused and propose a learning algorithm to select from all possible maps under the constraint that the map minimizes a probability distance and application-specific regularizers; thus, our method can be seen as solving a modified optimal transport problem. Once the map is learned, it can be used to map samples from a source domain to a target domain. In addition, because the map is parameterized as a composition of normalizing flows, it models the empirical distributions over the two data spaces and allows both sampling and likelihood evaluation for both data sets. We compare our method (parOT) to related optimal transport approaches in the context of domain adaptation and domain translation on benchmark data sets. Finally, to illustrate the impact of our work on applied problems, we apply parOT to a real scientific application: spectral calibration for high-dimensional measurements from two vastly different environments
△ Less
Submitted 17 April, 2023;
originally announced April 2023.
-
Generative structured normalizing flow Gaussian processes applied to spectroscopic data
Authors:
Natalie Klein,
Nishant Panda,
Patrick Gasda,
Diane Oyen
Abstract:
In this work, we propose a novel generative model for mapping inputs to structured, high-dimensional outputs using structured conditional normalizing flows and Gaussian process regression. The model is motivated by the need to characterize uncertainty in the input/output relationship when making inferences on new data. In particular, in the physical sciences, limited training data may not adequate…
▽ More
In this work, we propose a novel generative model for mapping inputs to structured, high-dimensional outputs using structured conditional normalizing flows and Gaussian process regression. The model is motivated by the need to characterize uncertainty in the input/output relationship when making inferences on new data. In particular, in the physical sciences, limited training data may not adequately characterize future observed data; it is critical that models adequately indicate uncertainty, particularly when they may be asked to extrapolate. In our proposed model, structured conditional normalizing flows provide parsimonious latent representations that relate to the inputs through a Gaussian process, providing exact likelihood calculations and uncertainty that naturally increases away from the training data inputs. We demonstrate the methodology on laser-induced breakdown spectroscopy data from the ChemCam instrument onboard the Mars rover Curiosity. ChemCam was designed to recover the chemical composition of rock and soil samples by measuring the spectral properties of plasma atomic emissions induced by a laser pulse. We show that our model can generate realistic spectra conditional on a given chemical composition and that we can use the model to perform uncertainty quantification of chemical compositions for new observed spectra. Based on our results, we anticipate that our proposed modeling approach may be useful in other scientific domains with high-dimensional, complex structure where it is important to quantify predictive uncertainty.
△ Less
Submitted 14 December, 2022;
originally announced December 2022.
-
Properties of the Nili Fossae Olivine-clay-carbonate lithology: orbital and in situ at Séítah
Authors:
Adrian J. Brown,
Linda Kah,
Lucia Mandon,
Roger Wiens,
Patrick Pinet,
Elise Clavé,
Stéphane Le Mouélic,
Arya Udry,
Patrick J. Gasda,
Clément Royer,
Keyron Hickman-Lewis11,
Agnes Cousin,
Justin I. Simon,
Jade Comellas14,
Edward Cloutis,
Thierry Fouchet,
Alberto G. Fairén,
Stephanie Connell,
David Flannery,
Briony Horgan,
Lisa Mayhew,
Allan Treiman,
Jorge I. Núñez,
Brittan Wogsland,
Karim Benzerara
, et al. (9 additional authors not shown)
Abstract:
We examine the observed properties of the Nili Fossae olivine-clay-carbonate lithology from orbital data and in situ by the Mars 2020 rover at the Séítah unit in Jezero crater, including: 1) composition (Liu, 2022) 2) grain size (Tice, 2022) 3) inferred viscosity (calculated based on geochemistry collected by SuperCam (Wiens, 2022)). Based on the low viscosity and distribution of the unit we postu…
▽ More
We examine the observed properties of the Nili Fossae olivine-clay-carbonate lithology from orbital data and in situ by the Mars 2020 rover at the Séítah unit in Jezero crater, including: 1) composition (Liu, 2022) 2) grain size (Tice, 2022) 3) inferred viscosity (calculated based on geochemistry collected by SuperCam (Wiens, 2022)). Based on the low viscosity and distribution of the unit we postulate a flood lava origin for the olivine-clay-carbonate at Séítah. We include a new CRISM map of the clay 2.38 μm band and use in situ data to show that the clay in the olivine cumulate in the Séítah formation is consistent with talc or serpentine from Mars 2020 SuperCam LIBS and VISIR and MastCam-Z observations. We discuss two intertwining aspects of the history of the lithology: 1) the emplacement and properties of the cumulate layer within a lava lake, based on terrestrial analogs in the Pilbara, Western Australia, and using previously published models of flood lavas and lava lakes, and 2) the limited extent of post emplacement alteration, including clay and carbonate alteration (Clave, 2022; Mandon, 2022).
△ Less
Submitted 27 June, 2022;
originally announced June 2022.
-
Iron Mobility During Diagenesis at Vera Rubin Ridge, Gale Crater, Mars
Authors:
J l'Haridon,
N Mangold,
A Fraeman,
J Johnson,
A Cousin,
W Rapin,
G David,
E Dehouck,
V Sun,
J Frydenvang,
O Gasnault,
P Gasda,
N Lanza,
O Forni,
P. -Y Meslin,
S Schwenzer,
J Bridges,
B Horgan,
C House,
S Maurice,
R Wiens
Abstract:
The Curiosity rover investigated a topographic structure known as Vera Rubin ridge, associated with a hematite signature in orbital spectra. There, Curiosity encountered mudstones interpreted as lacustrine deposits, in continuity with the 300 m-thick underlying sedimentary rocks of the Murray formation at the base of Mount Sharp. While the presence of hematite ($α$-Fe2O3) was confirmed insitu by b…
▽ More
The Curiosity rover investigated a topographic structure known as Vera Rubin ridge, associated with a hematite signature in orbital spectra. There, Curiosity encountered mudstones interpreted as lacustrine deposits, in continuity with the 300 m-thick underlying sedimentary rocks of the Murray formation at the base of Mount Sharp. While the presence of hematite ($α$-Fe2O3) was confirmed insitu by both Mastcam and ChemCam spectral observations and by the CheMin instrument, neither ChemCam nor APXS observed any significant increase in FeO$_T$ (total iron oxide) abundances compared to the Murray formation. Instead, Curiosity discovered dark-toned diagenetic features displaying anomalously high FeO$_T$ abundances, commonly observed in association with light-toned Ca-sulfate veins but also as crystal pseudomorphs in the host rock. These iron-rich diagenetic features are predominantly observed in "grey" outcrops on the upper part of the ridge, which lack the telltale ferric signature of other Vera Rubin ridge outcrops. Their composition is consistent with anhydrous Fe-oxide, as the enrichment in iron is not associated with enrichment in any other elements, nor with detections of volatiles. The lack of ferric absorption features in the ChemCam reflectance spectra and the hexagonal crystalline structure associated with dark-toned crystals points toward coarse "grey" hematite. In addition, the host rock adjacent to these features appears bleached and show low-FeO$_T$ content as well as depletion in Mn, indicating mobilization of these redox-sensitive elements during diagenesis. Thus, groundwater fluid circulations could account for the remobilization of iron and recrystallization as crystalline hematite during diagenesis as well as color variations observed in the Vera Rubin ridge outcrops.
△ Less
Submitted 11 January, 2021;
originally announced January 2021.
-
Nuclear Spectroscopy for the Exploration of Mars and Beyond
Authors:
Katherine E. Mesick,
Patrick J. Gasda,
Travis S. J. Gabriel,
Craig Hardgrove,
William C. Feldman
Abstract:
Nuclear spectroscopy is the only instrumentation that provides bulk geochemical constraints at depth (up to one meter in the surface). These instruments identify and quantify water and other key elements relevant to planetary exploration, including assessing planetary processes, context in the search for life, and in-situ resource utilization.
Nuclear spectroscopy is the only instrumentation that provides bulk geochemical constraints at depth (up to one meter in the surface). These instruments identify and quantify water and other key elements relevant to planetary exploration, including assessing planetary processes, context in the search for life, and in-situ resource utilization.
△ Less
Submitted 16 July, 2020;
originally announced July 2020.
