東華大學圖書館 |

Merging Ultrafast Gas-Phase Diffraction Experiment, Theory, and Machine Learning for a New Look at Molecular Dynamics /

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Merging Ultrafast Gas-Phase Diffraction Experiment, Theory, and Machine Learning for a New Look at Molecular Dynamics // Kareem Hegazy.
作者:	Hegazy, Kareem,
面頁冊數:	1 electronic resource (204 pages)
附註:	Source: Dissertations Abstracts International, Volume: 85-06, Section: B.
Contained By:	Dissertations Abstracts International85-06B.
標題:	Anisotropy. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30726786
ISBN:	9798381018585

Merging Ultrafast Gas-Phase Diffraction Experiment, Theory, and Machine Learning for a New Look at Molecular Dynamics /
Hegazy, Kareem,

Merging Ultrafast Gas-Phase Diffraction Experiment, Theory, and Machine Learning for a New Look at Molecular Dynamics /Kareem Hegazy. - 1 electronic resource (204 pages)

Source: Dissertations Abstracts International, Volume: 85-06, Section: B.

Chemical Physics is a broad field of study that bridges Physics and Chemistry to ultimately understand molecular dynamics from the first principles of quantum mechanics. Studying molecular dynamics is complicated by two properties, most molecules - including those with < 10 atoms - are considered a many-body problem and Chemistry is a time-dependent process. Each problem can be formidable on its own, and when combined into a single study they amplify their respective difficulties. Although we have long understood the molecular Hamiltonians that govern molecular dynamics, calculating such dynamics with a priorisimulation techniques and retrieving them from measurement still remains a difficult task. This is primarily due to the high-dimensional space of the molecular system, caused by the number of atoms and the consideration of both its nuclear and electronic components. Worse, this high-dimensional picture is constantly changing in time in such ways that areas of interest in this high-dimensional space can alternate in a discontinuous fashion. Such difficulties that arise from high-dimensional problems are known as the curse of dimensionality.In molecular dynamic simulations, one often evaluates trajectories within the coupled electronic and nuclear potential energy landscape as a function of time. With enough trajectories, these simulations are thought to sufficiently sample the space of all likely dynamics. However, this is not always true and in some instances the true dynamics may not be accessible to current numerical methods.In experiment, the signal from our many-body system is generally summed together into a single image or signal. Disentangling such measurements for more individual responses is generally intractable due to the lack of some orthogonal basis set to project upon and too limited of information to distinguish these individual contributions. Consequently, one often relies on the aforementioned simulations to provide the time-dependent molecular dynamics and its subsequent experimental signature.Over the past half-century, improvements in both experimental and simulation techniques have pushed this field forward. We are now able to study dynamics of larger (> 15 atoms) molecules with a temporal resolution on the timescale of nuclear motion (~100 fs) and spatial resolution less than typical bond distances (< 1 A). Today, the resurgence of machine learning has heralded a renaissance in statistics and large datasets. In this work, we take inspiration from previous and current statistical approaches to pioneer a general method that addresses the high-dimensional nature of this problem and effectively inverts experimental measurement for the physics of interest.To address this inverse problem, we first combine theory, experiment, and statistics to reinterpret the measured results. Using insights from previous works that access the molecular frame through ensemble anisotropy, we derive a detailed and fully quantum mechanical expression for the molecular frame structure in terms of ultrafast gas-phase diffraction patterns. This derivation provides a template for other experiments where such an expression is not currently available. We then effectively invert this expression for the probability of molecular structures |Ψ(R, t)| 2 using Bayesian Inference, a statistical method that uses Bayes' rule to update some theory given new data. Bayesian Inference uses the statistical nature of our measurement to effectively invert the experiment through a probabilistic relation that expresses the physical parameters of interest in terms of our measurement.

English

ISBN: 9798381018585Subjects--Topical Terms:

596747
Anisotropy.

Merging Ultrafast Gas-Phase Diffraction Experiment, Theory, and Machine Learning for a New Look at Molecular Dynamics /
LDR:05185nmm a22004453i 4500 001 2400486
005 20250522084137.5
006 m o d
007 cr|nu||||||||
008 251215s2023 miu||||||m |||||||eng d
020 $a 9798381018585
035 $a (MiAaPQD)AAI30726786
035 $a (MiAaPQD)STANFORDqz077fp1847
035 $a AAI30726786
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Hegazy, Kareem, $e author. $3 3770503
245 1 0 $a Merging Ultrafast Gas-Phase Diffraction Experiment, Theory, and Machine Learning for a New Look at Molecular Dynamics / $c Kareem Hegazy.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2023
300 $a 1 electronic resource (204 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-06, Section: B.
500 $a Advisors: Coffee, Ryan; Dunne, Mike; Reis, David; Bucksbaum, Philip Committee members: Bent, Stacey F.
502 $b Ph.D. $c Stanford University $d 2023.
520 $a Chemical Physics is a broad field of study that bridges Physics and Chemistry to ultimately understand molecular dynamics from the first principles of quantum mechanics. Studying molecular dynamics is complicated by two properties, most molecules - including those with < 10 atoms - are considered a many-body problem and Chemistry is a time-dependent process. Each problem can be formidable on its own, and when combined into a single study they amplify their respective difficulties. Although we have long understood the molecular Hamiltonians that govern molecular dynamics, calculating such dynamics with a priorisimulation techniques and retrieving them from measurement still remains a difficult task. This is primarily due to the high-dimensional space of the molecular system, caused by the number of atoms and the consideration of both its nuclear and electronic components. Worse, this high-dimensional picture is constantly changing in time in such ways that areas of interest in this high-dimensional space can alternate in a discontinuous fashion. Such difficulties that arise from high-dimensional problems are known as the curse of dimensionality.In molecular dynamic simulations, one often evaluates trajectories within the coupled electronic and nuclear potential energy landscape as a function of time. With enough trajectories, these simulations are thought to sufficiently sample the space of all likely dynamics. However, this is not always true and in some instances the true dynamics may not be accessible to current numerical methods.In experiment, the signal from our many-body system is generally summed together into a single image or signal. Disentangling such measurements for more individual responses is generally intractable due to the lack of some orthogonal basis set to project upon and too limited of information to distinguish these individual contributions. Consequently, one often relies on the aforementioned simulations to provide the time-dependent molecular dynamics and its subsequent experimental signature.Over the past half-century, improvements in both experimental and simulation techniques have pushed this field forward. We are now able to study dynamics of larger (> 15 atoms) molecules with a temporal resolution on the timescale of nuclear motion (~100 fs) and spatial resolution less than typical bond distances (< 1 A). Today, the resurgence of machine learning has heralded a renaissance in statistics and large datasets. In this work, we take inspiration from previous and current statistical approaches to pioneer a general method that addresses the high-dimensional nature of this problem and effectively inverts experimental measurement for the physics of interest.To address this inverse problem, we first combine theory, experiment, and statistics to reinterpret the measured results. Using insights from previous works that access the molecular frame through ensemble anisotropy, we derive a detailed and fully quantum mechanical expression for the molecular frame structure in terms of ultrafast gas-phase diffraction patterns. This derivation provides a template for other experiments where such an expression is not currently available. We then effectively invert this expression for the probability of molecular structures |Ψ(R, t)| 2 using Bayesian Inference, a statistical method that uses Bayes' rule to update some theory given new data. Bayesian Inference uses the statistical nature of our measurement to effectively invert the experiment through a probabilistic relation that expresses the physical parameters of interest in terms of our measurement.
546 $a English
590 $a School code: 0212
650 4 $a Anisotropy. $3 596747
650 4 $a Physics. $3 516296
650 4 $a Electrons. $3 516581
650 4 $a Inverse problems. $3 3686950
650 4 $a Lasers. $3 535503
650 4 $a Nitrogen. $3 1314426
650 4 $a Atomic physics. $3 3173870
650 4 $a Mathematics. $3 515831
650 4 $a Optics. $3 517925
690 $a 0605
690 $a 0800
690 $a 0748
690 $a 0405
690 $a 0752
710 2 $a Stanford University. $e degree granting institution. $3 3765820
720 1 $a Coffee, Ryan $e degree supervisor.
720 1 $a Dunne, Mike $e degree supervisor.
720 1 $a Reis, David $e degree supervisor.
720 1 $a Bucksbaum, Philip $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 85-06B.
790 $a 0212
791 $a Ph.D.
792 $a 2023
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30726786