arXiv subjects

Katie Christensen

Publications and source records attributed to Katie Christensen.

5 recordsLinked to original sources

arXiv2024

Diffusion-Based Joint Temperature and Precipitation Emulation of Earth System Models

Earth system models (ESMs) are the principal tools used in climate science to generate future climate projections under various atmospheric emissions scenarios on a global or regional scale. Generative deep learning approaches are suitable for emulating these tools due to their computational efficiency and ability, once trained, to generate realizations in a fraction of the time required by ESMs. We extend previous work that used a generative probabilistic diffusion model to emulate ESMs by targeting the joint emulation of multiple variables, temperature and precipitation, by a single diffusion model. Joint generation of multiple variables is critical to generate realistic samples of phenomena resulting from the interplay of multiple variables. The diffusion model emulator takes in the monthly mean-maps of temperature and precipitation and produces the daily values of each of these variables that exhibit statistical properties similar to those generated by ESMs. Our results show the outputs from our extended model closely resemble those from ESMs on various climate metrics including dry spells and hot streaks, and that the joint distribution of temperature and precipitation in our sample closely matches those of ESMs.

physics.ao-ph↗

arXiv2023

RoseNet: Predicting Energy Metrics of Double InDel Mutants Using Deep Learning

An amino acid insertion or deletion, or InDel, can have profound and varying functional impacts on a protein's structure. InDel mutations in the transmembrane conductor regulator protein for example give rise to cystic fibrosis. Unfortunately performing InDel mutations on physical proteins and studying their effects is a time prohibitive process. Consequently, modeling InDels computationally can supplement and inform wet lab experiments. In this work, we make use of our data sets of exhaustive double InDel mutations for three proteins which we computationally generated using a robotics inspired inverse kinematics approach available in Rosetta. We develop and train a neural network, RoseNet, on several structural and energetic metrics output by Rosetta during the mutant generation process. We explore and present how RoseNet is able to emulate the exhaustive data set using deep learning methods, and show to what extent it can predict Rosetta metrics for unseen mutant sequences with two InDels. RoseNet achieves a Pearson correlation coefficient median accuracy of 0.775 over all Rosetta scores for the largest protein. Furthermore, a sensitivity analysis is performed to determine the necessary quantity of data required to accurately emulate the structural scores for computationally generated mutants. We show that the model can be trained on minimal data (<50%) and still retain a high level of accuracy.

q-bio.BM↗

arXiv2020

Some remarks about trunks and morphisms of neural codes

We give intrinsic characterizations of neural rings and homomorphisms between them. Also we introduce the notion of a basic monomial code map and characterize monomial code maps as compositions of basic monomial code maps. Finally, we characterize monomial isomorphisms between neural codes. Our work is based on the 2015 paper by C.~Curto and N.~Youngs about neural ring homomorphisms and maps between neural codes and on the 2018 paper by R.~Amzi Jeffs about morphisms of neural rings.

math.AC↗

arXiv2018

A new characterization of principal ideal domains

In 2008 N.~Q.~Chinh and P.~H.~Nam characterized principal ideal domains as integral domains that satisfy the follo\-wing two conditions: (i) they are unique factorization domains, and (ii) all maximal ideals in them are principal. We improve their result by giving a characterization in which each of these two conditions is weakened. At the same time we improve a theorem by P.~M.~Cohn which characterizes principal ideal domains as atomic Bézout domains. We will also show that every PC domain is AP and that the notion of PC domains is incomparable with the notion of pre-Schreier domains (hence with the notions of Schreier and GCD domains as well).

math.AC↗

arXiv2018

Polarization of neural codes

The neural rings and ideals as an algebraic tool for analyzing the intrinsic structure of neural codes were introduced by C.~Curto et al. in 2013. Since then they were investigated in several papers, including the 2017 paper by Güntürkün et al., in which the notion of polarization of neural ideals was introduced. In this paper we extend their ideas by introducing the notions of polarization of motifs and neural codes. We show that the notions that we introduced have very nice properties which could allow the studying of the intrinsic structure of neural codes of length $n$ via the square free monomial ideals in $2n$ variables and interpreting the results back in the original neural code ambient space. In the last section of the paper we introduce the notions of inactive neurons, partial neural codes, and partial motifs, as well as the notions of polarization of these codes and motifs. We use these notions to give a new proof of a theorem from the paper by Güntürkün et al. that we mentioned above.

math.AC↗