-
A piezoelectric ski-jump laser beam scanning chip-to-free space photonic link
Authors:
Matt Saha,
Y. Henry Wen,
Andrew S. Greenspon,
Matthew Zimmermann,
Kevin J. Palm,
Alex Witte,
Mark Dong,
Andrew J. Leenheer,
Genevieve Clark,
Gerald Gilbert,
Matt Eichenfield,
Dirk Englund
Abstract:
A seamless interface between integrated photonic processors and targets in free-space enables wide-ranging advancements in telescopy, free-space communication, optical ranging, materials processing, biomedical imaging, near eye display, machine optical intelligence and quantum control. An optimal solution allows for 2D scanning from anywhere on a photonic chip over a large number of diffraction li…
▽ More
A seamless interface between integrated photonic processors and targets in free-space enables wide-ranging advancements in telescopy, free-space communication, optical ranging, materials processing, biomedical imaging, near eye display, machine optical intelligence and quantum control. An optimal solution allows for 2D scanning from anywhere on a photonic chip over a large number of diffraction limited spots in the far field. Leading approaches rely on scanners where the numerical aperture and actuator size are linked, resulting in a trade off between resolution, speed and footprint, whereas scanning fibers have been limited to bulk optical and mechanical components. Here, we introduce a CMOS fabricated photonic "ski-jump" composed of a broadband, single mode silicon nitride waveguide monolithically integrated atop a piezo-actuated cantilever. The ski-jump passively curl 90 degrees out-of-plane via mechanical meta-stress engineering in a footprint of less than 0.1 mm squared and emit submicron diffraction-limited optical modes with piezoelectric steering. They also exhibit kHz-rate longitudinal and lateral mechanical resonances with displacement ranges exceeding 400 micron and 180 micron, respectively, and quality factors Q>10,000 under vacuum. These resonances enable 2D beam scanning at footprint-adjusted spot-rates of 68.6 Megaspot/s-mm squared surpassing state-of-the-art MEMS mirrors by more than 50. Using these devices, we demonstrate arbitrary 2D image projection and the repeatable initialization and readout of single photons from silicon vacancies in diamond waveguides. Based on current device performance, we identify pathways for achieving >1 Giga-spots in a square cm area to provide a seamless, scalable optical pipeline between integrated photonic processors and the free-space world.
△ Less
Submitted 25 June, 2024;
originally announced June 2024.
-
Uncertainty estimates for semantic segmentation: providing enhanced reliability for automated motor claims handling
Authors:
Jan Küchler,
Daniel Kröll,
Sebastian Schoenen,
Andreas Witte
Abstract:
Deep neural network models for image segmentation can be a powerful tool for the automation of motor claims handling processes in the insurance industry. A crucial aspect is the reliability of the model outputs when facing adverse conditions, such as low quality photos taken by claimants to document damages. We explore the use of a meta-classification model to empirically assess the precision of s…
▽ More
Deep neural network models for image segmentation can be a powerful tool for the automation of motor claims handling processes in the insurance industry. A crucial aspect is the reliability of the model outputs when facing adverse conditions, such as low quality photos taken by claimants to document damages. We explore the use of a meta-classification model to empirically assess the precision of segments predicted by a model trained for the semantic segmentation of car body parts. Different sets of features correlated with the quality of a segment are compared, and an AUROC score of 0.915 is achieved for distinguishing between high- and low-quality segments. By removing low-quality segments, the average mIoU of the segmentation output is improved by 16 percentage points and the number of wrongly predicted segments is reduced by 77%.
△ Less
Submitted 17 May, 2024; v1 submitted 17 January, 2024;
originally announced January 2024.
-
Jets of foliations and $b^k$-algebroids
Authors:
Francis Bischoff,
Álvaro del Pino,
Aldo Witte
Abstract:
In this article, we introduce and study singular foliations of $b^k$-type. These singular foliations formalize the properties of vector fields that are tangent to order $k$ along a submanifold $W \subset M$. Our first result is a classification of these foliations, relating them to geometric structures defined in a formal neighborhood of the submanifold, such as jets of distributions that are invo…
▽ More
In this article, we introduce and study singular foliations of $b^k$-type. These singular foliations formalize the properties of vector fields that are tangent to order $k$ along a submanifold $W \subset M$. Our first result is a classification of these foliations, relating them to geometric structures defined in a formal neighborhood of the submanifold, such as jets of distributions that are involutive up to order $k-1$.
When $W$ is a hypersurface, singular foliations of $b^k$-type are Lie algebroids. In this particular case, they are generalizations of the $b^k$-tangent bundles introduced by Scott. Indeed, they are always locally isomorphic to $b^k$-tangent bundles, but globally such an isomorphism is obstructed by a holonomy invariant. Our second main result is a Riemann-Hilbert-style classification of singular foliations of $b^k$-type in terms of holonomy representations.
In this paper, we study singular foliations of $b^k$-type from several different perspectives. In particular: (1) We study the problem of extending a $k$-th-order foliation to a $(k+1)$-th order foliation and prove that this is obstructed by a characteristic class. (2) When $W$ is a hypersurface, we give a detailed study of algebroid differential forms and extend Scott's calculation of the cohomology. (3) We study algebroid symplectic forms in terms of the geometric structures induced on $W$. In particular, we find that there is a close relationship between the above obstruction class for extensions and the symplectic variation of the symplectic foliation induced on $W$.
△ Less
Submitted 28 November, 2023;
originally announced November 2023.
-
Sustainable Concrete via Bayesian Optimization
Authors:
Sebastian Ament,
Andrew Witte,
Nishant Garg,
Julius Kusuma
Abstract:
Eight percent of global carbon dioxide emissions can be attributed to the production of cement, the main component of concrete, which is also the dominant source of CO2 emissions in the construction of data centers. The discovery of lower-carbon concrete formulae is therefore of high significance for sustainability. However, experimenting with new concrete formulae is time consuming and labor inte…
▽ More
Eight percent of global carbon dioxide emissions can be attributed to the production of cement, the main component of concrete, which is also the dominant source of CO2 emissions in the construction of data centers. The discovery of lower-carbon concrete formulae is therefore of high significance for sustainability. However, experimenting with new concrete formulae is time consuming and labor intensive, as one usually has to wait to record the concrete's 28-day compressive strength, a quantity whose measurement can by its definition not be accelerated. This provides an opportunity for experimental design methodology like Bayesian Optimization (BO) to accelerate the search for strong and sustainable concrete formulae. Herein, we 1) propose modeling steps that make concrete strength amenable to be predicted accurately by a Gaussian process model with relatively few measurements, 2) formulate the search for sustainable concrete as a multi-objective optimization problem, and 3) leverage the proposed model to carry out multi-objective BO with real-world strength measurements of the algorithmically proposed mixes. Our experimental results show improved trade-offs between the mixtures' global warming potential (GWP) and their associated compressive strengths, compared to mixes based on current industry practices. Our methods are open-sourced at github.com/facebookresearch/SustainableConcrete.
△ Less
Submitted 20 November, 2023; v1 submitted 27 October, 2023;
originally announced October 2023.
-
ResBuilder: Automated Learning of Depth with Residual Structures
Authors:
Julian Burghoff,
Matthias Rottmann,
Jill von Conta,
Sebastian Schoenen,
Andreas Witte,
Hanno Gottschalk
Abstract:
In this work, we develop a neural architecture search algorithm, termed Resbuilder, that develops ResNet architectures from scratch that achieve high accuracy at moderate computational cost. It can also be used to modify existing architectures and has the capability to remove and insert ResNet blocks, in this way searching for suitable architectures in the space of ResNet architectures. In our exp…
▽ More
In this work, we develop a neural architecture search algorithm, termed Resbuilder, that develops ResNet architectures from scratch that achieve high accuracy at moderate computational cost. It can also be used to modify existing architectures and has the capability to remove and insert ResNet blocks, in this way searching for suitable architectures in the space of ResNet architectures. In our experiments on different image classification datasets, Resbuilder achieves close to state-of-the-art performance while saving computational cost compared to off-the-shelf ResNets. Noteworthy, we once tune the parameters on CIFAR10 which yields a suitable default choice for all other datasets. We demonstrate that this property generalizes even to industrial applications by applying our method with default parameters on a proprietary fraud detection dataset.
△ Less
Submitted 16 August, 2023;
originally announced August 2023.
-
Synchronous micromechanically resonant programmable photonic circuits
Authors:
Mark Dong,
Julia M. Boyle,
Kevin J. Palm,
Matthew Zimmermann,
Alex Witte,
Andrew J. Leenheer,
Daniel Dominguez,
Gerald Gilbert,
Matt Eichenfield,
Dirk Englund
Abstract:
Programmable photonic integrated circuits (PICs) are emerging as powerful tools for the precise manipulation of light, with applications in quantum information processing, optical range finding, and artificial intelligence. The leading architecture for programmable PICs is the mesh of Mach-Zehnder interferometers (MZIs) embedded with reconfigurable optical phase shifters. Low-power implementations…
▽ More
Programmable photonic integrated circuits (PICs) are emerging as powerful tools for the precise manipulation of light, with applications in quantum information processing, optical range finding, and artificial intelligence. The leading architecture for programmable PICs is the mesh of Mach-Zehnder interferometers (MZIs) embedded with reconfigurable optical phase shifters. Low-power implementations of these PICs involve micromechanical structures driven capacitively or piezoelectrically but are limited in modulation bandwidth by mechanical resonances and high operating voltages. However, circuits designed to operate exclusively at these mechanical resonances would reduce the necessary driving voltage from resonantly enhanced modulation as well as maintaining high actuation speeds. Here we introduce a synchronous, micromechanically resonant design architecture for programmable PICs, which exploits micromechanical eigenmodes for modulation enhancement. This approach combines high-frequency mechanical resonances and optically broadband phase shifters to increase the modulation response on the order of the mechanical quality factor $Q_m$, thereby reducing the PIC's power consumption, voltage-loss product, and footprint. The architecture is useful for broadly applicable circuits such as optical phased arrays, $1$ x $N$, and $N$ x $N$ photonic switches. We report a proof-of-principle programmable 1 x 8 switch with piezoelectric phase shifters at specifically targeted mechanical eigenfrequencies, showing a full switching cycle of all eight channels spaced by approximately 11 ns and operating at >3x average modulation enhancement across all on-chip modulators. By further leveraging micromechanical devices with high $Q_m$, which can exceed 1 million, our design architecture should enable a new class of low-voltage and high-speed programmable PICs.
△ Less
Submitted 6 June, 2023;
originally announced June 2023.
-
Learned multiphysics inversion with differentiable programming and machine learning
Authors:
Mathias Louboutin,
Ziyi Yin,
Rafael Orozco,
Thomas J. Grady II,
Ali Siahkoohi,
Gabrio Rizzuti,
Philipp A. Witte,
Olav Møyner,
Gerard J. Gorman,
Felix J. Herrmann
Abstract:
We present the Seismic Laboratory for Imaging and Modeling/Monitoring (SLIM) open-source software framework for computational geophysics and, more generally, inverse problems involving the wave-equation (e.g., seismic and medical ultrasound), regularization with learned priors, and learned neural surrogates for multiphase flow simulations. By integrating multiple layers of abstraction, our softwar…
▽ More
We present the Seismic Laboratory for Imaging and Modeling/Monitoring (SLIM) open-source software framework for computational geophysics and, more generally, inverse problems involving the wave-equation (e.g., seismic and medical ultrasound), regularization with learned priors, and learned neural surrogates for multiphase flow simulations. By integrating multiple layers of abstraction, our software is designed to be both readable and scalable. This allows researchers to easily formulate their problems in an abstract fashion while exploiting the latest developments in high-performance computing. We illustrate and demonstrate our design principles and their benefits by means of building a scalable prototype for permeability inversion from time-lapse crosswell seismic data, which aside from coupling of wave physics and multiphase flow, involves machine learning.
△ Less
Submitted 11 April, 2023;
originally announced April 2023.
-
Non-principal T-duality, generalized complex geometry and blow-ups
Authors:
Gil R. Cavalcanti,
Aldo Witte
Abstract:
We extend the notion of T-duality to manifolds endowed with non-principal torus actions. The singularities of the torus action are controlled by a certain Lie algebroid, called the elliptic tangent bundle. Using this Lie algebroid, we explain how certain invariant generalized complex structures can be transported via T-duality. Along the way, we use the elliptic tangent bundle to define connection…
▽ More
We extend the notion of T-duality to manifolds endowed with non-principal torus actions. The singularities of the torus action are controlled by a certain Lie algebroid, called the elliptic tangent bundle. Using this Lie algebroid, we explain how certain invariant generalized complex structures can be transported via T-duality. Along the way, we use the elliptic tangent bundle to define connections for these torus action, and give new insight to the classification of such actions by Haefliger-Salem.
△ Less
Submitted 24 May, 2023; v1 submitted 30 November, 2022;
originally announced November 2022.
-
Regularisation of Lie algebroids and Applications
Authors:
Álvaro del Pino,
Aldo Witte
Abstract:
We describe a procedure, called regularisation, that allows us to study geometric structures on Lie algebroids via foliated geometric structures on a manifold of higher dimension. This procedure applies to various classes of Lie algebroids; namely, those whose singularities are of b^k, complex-log, or elliptic type, possibly with self-crossings.
One of our main applications is a proof of the Wei…
▽ More
We describe a procedure, called regularisation, that allows us to study geometric structures on Lie algebroids via foliated geometric structures on a manifold of higher dimension. This procedure applies to various classes of Lie algebroids; namely, those whose singularities are of b^k, complex-log, or elliptic type, possibly with self-crossings.
One of our main applications is a proof of the Weinstein conjecture for overtwisted b^k-contact structures. This was proven by Miranda-Oms using a certain technical hypothesis. Our approach avoids this assumption by reducing the proof to the foliated setting. As a by-product, we also prove the Weinstein conjecture for other Lie algebroids.
Along the way we also introduce tangent distributions, i.e. subbundles of Lie algebroids, as interesting objects of study and present a number of constructions for them.
△ Less
Submitted 27 November, 2022;
originally announced November 2022.
-
SciAI4Industry -- Solving PDEs for industry-scale problems with deep learning
Authors:
Philipp A. Witte,
Russell J. Hewett,
Kumar Saurabh,
AmirHossein Sojoodi,
Ranveer Chandra
Abstract:
Solving partial differential equations with deep learning makes it possible to reduce simulation times by multiple orders of magnitude and unlock scientific methods that typically rely on large numbers of sequential simulations, such as optimization and uncertainty quantification. Two of the largest challenges of adopting scientific AI for industrial problem settings is that training datasets must…
▽ More
Solving partial differential equations with deep learning makes it possible to reduce simulation times by multiple orders of magnitude and unlock scientific methods that typically rely on large numbers of sequential simulations, such as optimization and uncertainty quantification. Two of the largest challenges of adopting scientific AI for industrial problem settings is that training datasets must be simulated in advance and that neural networks for solving large-scale PDEs exceed the memory capabilities of current GPUs. We introduce a distributed programming API in the Julia language for simulating training data in parallel on the cloud and without requiring users to manage the underlying HPC infrastructure. In addition, we show that model-parallel deep learning based on domain decomposition allows us to scale neural networks for solving PDEs to commercial-scale problem settings and achieve above 90% parallel efficiency. Combining our cloud API for training data generation and model-parallel deep learning, we train large-scale neural networks for solving the 3D Navier-Stokes equation and simulating 3D CO2 flow in porous media. For the CO2 example, we simulate a training dataset based on a commercial carbon capture and storage (CCS) project and train a neural network for CO2 flow simulation on a 3D grid with over 2 million cells that is 5 orders of magnitudes faster than a conventional numerical simulator and 3,200 times cheaper.
△ Less
Submitted 23 November, 2022;
originally announced November 2022.
-
Programmable photonic integrated meshes for modular generation of optical entanglement links
Authors:
Mark Dong,
Matthew Zimmermann,
David Heim,
Hyeongrak Choi,
Genevieve Clark,
Andrew J. Leenheer,
Kevin J. Palm,
Alex Witte,
Daniel Dominguez,
Gerald Gilbert,
Matt Eichenfield,
Dirk Englund
Abstract:
Large-scale generation of quantum entanglement between individually controllable qubits is at the core of quantum computing, communications, and sensing. Modular architectures of remotely-connected quantum technologies have been proposed for a variety of physical qubits, with demonstrations reported in atomic and all-photonic systems. However, an open challenge in these architectures lies in const…
▽ More
Large-scale generation of quantum entanglement between individually controllable qubits is at the core of quantum computing, communications, and sensing. Modular architectures of remotely-connected quantum technologies have been proposed for a variety of physical qubits, with demonstrations reported in atomic and all-photonic systems. However, an open challenge in these architectures lies in constructing high-speed and high-fidelity reconfigurable photonic networks for optically-heralded entanglement among target qubits. Here we introduce a programmable photonic integrated circuit (PIC), realized in a piezo-actuated silicon nitride (SiN)-in-oxide CMOS-compatible process, that implements an N x N Mach-Zehnder mesh (MZM) capable of high-speed execution of linear optical transformations. The visible-spectrum photonic integrated mesh is programmed to generate optical connectivity on up to N = 8 inputs for a range of optically-heralded entanglement protocols. In particular, we experimentally demonstrated optical connections between 16 independent pairwise mode couplings through the MZM, with optical transformation fidelities averaging 0.991 +/- 0.0063. The PIC's reconfigurable optical connectivity suffices for the production of 8-qubit resource states as building blocks of larger topological cluster states for quantum computing. Our programmable PIC platform enables the fast and scalable optical switching technology necessary for network-based quantum information processors.
△ Less
Submitted 29 August, 2022;
originally announced August 2022.
-
Model-Parallel Fourier Neural Operators as Learned Surrogates for Large-Scale Parametric PDEs
Authors:
Thomas J. Grady II,
Rishi Khan,
Mathias Louboutin,
Ziyi Yin,
Philipp A. Witte,
Ranveer Chandra,
Russell J. Hewett,
Felix J. Herrmann
Abstract:
Fourier neural operators (FNOs) are a recently introduced neural network architecture for learning solution operators of partial differential equations (PDEs), which have been shown to perform significantly better than comparable deep learning approaches. Once trained, FNOs can achieve speed-ups of multiple orders of magnitude over conventional numerical PDE solvers. However, due to the high dimen…
▽ More
Fourier neural operators (FNOs) are a recently introduced neural network architecture for learning solution operators of partial differential equations (PDEs), which have been shown to perform significantly better than comparable deep learning approaches. Once trained, FNOs can achieve speed-ups of multiple orders of magnitude over conventional numerical PDE solvers. However, due to the high dimensionality of their input data and network weights, FNOs have so far only been applied to two-dimensional or small three-dimensional problems. To remove this limited problem-size barrier, we propose a model-parallel version of FNOs based on domain-decomposition of both the input data and network weights. We demonstrate that our model-parallel FNO is able to predict time-varying PDE solutions of over 2.6 billion variables on Perlmutter using up to 512 A100 GPUs and show an example of training a distributed FNO on the Azure cloud for simulating multiphase CO$_2$ dynamics in the Earth's subsurface.
△ Less
Submitted 1 February, 2023; v1 submitted 3 April, 2022;
originally announced April 2022.
-
Accelerating innovation with software abstractions for scalable computational geophysics
Authors:
Mathias Louboutin,
Philipp A. Witte,
Ali Siahkoohi,
Gabrio Rizzuti,
Ziyi Yin,
Rafael Orozco,
Felix J. Herrmann
Abstract:
We present the SLIM (https://github.com/slimgroup) open-source software framework for computational geophysics, and more generally, inverse problems based on the wave-equation (e.g., medical ultrasound). We developed a software environment aimed at scalable research and development by designing multiple layers of abstractions. This environment allows the researchers to easily formulate their probl…
▽ More
We present the SLIM (https://github.com/slimgroup) open-source software framework for computational geophysics, and more generally, inverse problems based on the wave-equation (e.g., medical ultrasound). We developed a software environment aimed at scalable research and development by designing multiple layers of abstractions. This environment allows the researchers to easily formulate their problem in an abstract fashion, while still being able to exploit the latest developments in high-performance computing. We illustrate and demonstrate the benefits of our software design on many geophysical applications, including seismic inversion and physics-informed machine learning for geophysics (e.g., loop unrolled imaging, uncertainty quantification), all while facilitating the integration of external software.
△ Less
Submitted 28 March, 2022;
originally announced March 2022.
-
Piezo-optomechanical cantilever modulators for VLSI visible photonics
Authors:
Mark Dong,
David Heim,
Alex Witte,
Genevieve Clark,
Andrew J. Leenheer,
Daniel Dominguez,
Matthew Zimmermann,
Y. Henry Wen,
Gerald Gilbert,
Dirk Englund,
Matt Eichenfield
Abstract:
Visible-wavelength very large-scale integration (VLSI) photonic circuits have potential to play important roles in quantum information and sensing technologies. The realization of scalable, high-speed, and low-loss photonic mesh circuits depends on reliable and well-engineered visible photonic components. Here we report a low-voltage optical phase shifter based on piezo-actuated mechanical cantile…
▽ More
Visible-wavelength very large-scale integration (VLSI) photonic circuits have potential to play important roles in quantum information and sensing technologies. The realization of scalable, high-speed, and low-loss photonic mesh circuits depends on reliable and well-engineered visible photonic components. Here we report a low-voltage optical phase shifter based on piezo-actuated mechanical cantilevers, fabricated on a CMOS compatible, 200 mm wafer-based visible photonics platform. We show linear phase and amplitude modulation with 6 V$_π$-cm in differential operation, -1.5 dB to -2 dB insertion loss, and up to 40 dB contrast in the 700 nm - 780 nm range. By adjusting selected cantilever parameters, we demonstrate a low-displacement and a high-displacement device, both exhibiting a nearly flat frequency response from DC to a peak mechanical resonance at 23 MHz and 6.8 MHz respectively, which through resonant enhancement of Q~40, further decreases the operating voltage down to 0.15 V$_π$-cm.
△ Less
Submitted 28 January, 2022;
originally announced January 2022.
-
The cohomology of the elliptic tangent bundle
Authors:
Aldo Witte
Abstract:
In this note we compute the cohomology of the elliptic tangent bundle, a Lie algebroid used to describe singular symplectic forms arising from generalized complex geometry.
In this note we compute the cohomology of the elliptic tangent bundle, a Lie algebroid used to describe singular symplectic forms arising from generalized complex geometry.
△ Less
Submitted 10 April, 2021;
originally announced April 2021.
-
Unusual heat transport of the Kitaev material Na$_2$Co$_2$TeO$_6$: putative quantum spin liquid and low-energy spin excitations
Authors:
Xiaochen Hong,
Matthias Gillig,
Richard Hentrich,
Weiliang Yao,
Vilmos Kocsis,
Arthur R. Witte,
Tino Schreiner,
Danny Baumann,
Nicolás Pérez,
Anja U. B. Wolter,
Yuan Li,
Bernd Büchner,
Christian Hess
Abstract:
We studied the field dependent thermal conductivity ($κ$) of Na$_2$Co$_2$TeO$_6$, a compound considered as the manifestation of the Kitaev model based on the high-spin $d^7$ Co$^{2+}$ ions. We found that in-plane magnetic fields beyond a critical value $B_c \approx$~10 T are able to drastically enhance $κ$ at low temperatures, resulting in a double-peak structure of $κ(T)$ that closely resembles t…
▽ More
We studied the field dependent thermal conductivity ($κ$) of Na$_2$Co$_2$TeO$_6$, a compound considered as the manifestation of the Kitaev model based on the high-spin $d^7$ Co$^{2+}$ ions. We found that in-plane magnetic fields beyond a critical value $B_c \approx$~10 T are able to drastically enhance $κ$ at low temperatures, resulting in a double-peak structure of $κ(T)$ that closely resembles the behavior of $α$-RuCl$_3$. This result suggests that heat transport in Na$_2$Co$_2$TeO$_6$ is primarily phononic, and it is strongly affected by scattering from magnetic excitations that are highly tunable by external fields. Interestingly, for magnetic fields $B // a$ (i.e., along the zigzag direction of the Co-Co bonds), there is an extended field range which separates the long-range magnetic order for $B\leq B_c\approx10$ T and the partially spin-polarized gapped high-field phase for $B\gtrsim 12$ T. The low-energy phonon scattering is particularly strong in this field range, consistent with the notion that the system becomes a quantum spin liquid with prominent spin fluctuations down to energies of no more than 2 meV.
△ Less
Submitted 28 January, 2021;
originally announced January 2021.
-
Preconditioned training of normalizing flows for variational inference in inverse problems
Authors:
Ali Siahkoohi,
Gabrio Rizzuti,
Mathias Louboutin,
Philipp A. Witte,
Felix J. Herrmann
Abstract:
Obtaining samples from the posterior distribution of inverse problems with expensive forward operators is challenging especially when the unknowns involve the strongly heterogeneous Earth. To meet these challenges, we propose a preconditioning scheme involving a conditional normalizing flow (NF) capable of sampling from a low-fidelity posterior distribution directly. This conditional NF is used to…
▽ More
Obtaining samples from the posterior distribution of inverse problems with expensive forward operators is challenging especially when the unknowns involve the strongly heterogeneous Earth. To meet these challenges, we propose a preconditioning scheme involving a conditional normalizing flow (NF) capable of sampling from a low-fidelity posterior distribution directly. This conditional NF is used to speed up the training of the high-fidelity objective involving minimization of the Kullback-Leibler divergence between the predicted and the desired high-fidelity posterior density for indirect measurements at hand. To minimize costs associated with the forward operator, we initialize the high-fidelity NF with the weights of the pretrained low-fidelity NF, which is trained beforehand on available model and data pairs. Our numerical experiments, including a 2D toy and a seismic compressed sensing example, demonstrate that thanks to the preconditioning considerable speed-ups are achievable compared to training NFs from scratch.
△ Less
Submitted 11 January, 2021;
originally announced January 2021.
-
Fibrations in semi-toric and generalized complex geometry
Authors:
Gil R. Cavalcanti,
Ralph L. Klaasse,
Aldo Witte
Abstract:
This paper studies the interplay between self-crossing boundary Lefschetz fibrations and generalized complex structures. We show that these fibrations arise from the moment maps in semi-toric geometry and use them to construct self-crossing stable generalized complex four-manifolds using Gompf--Thurston methods for Lie algebroids. These results bring forth further structure on several previously k…
▽ More
This paper studies the interplay between self-crossing boundary Lefschetz fibrations and generalized complex structures. We show that these fibrations arise from the moment maps in semi-toric geometry and use them to construct self-crossing stable generalized complex four-manifolds using Gompf--Thurston methods for Lie algebroids. These results bring forth further structure on several previously known examples of generalized complex manifolds. We moreover show that these fibrations are compatible with taking connected sums, and use this to prove a singularity trade result between two types of singularities occurring in these fibrations.
△ Less
Submitted 12 January, 2021; v1 submitted 24 December, 2020;
originally announced December 2020.
-
Faster Uncertainty Quantification for Inverse Problems with Conditional Normalizing Flows
Authors:
Ali Siahkoohi,
Gabrio Rizzuti,
Philipp A. Witte,
Felix J. Herrmann
Abstract:
In inverse problems, we often have access to data consisting of paired samples $(x,y)\sim p_{X,Y}(x,y)$ where $y$ are partial observations of a physical system, and $x$ represents the unknowns of the problem. Under these circumstances, we can employ supervised training to learn a solution $x$ and its uncertainty from the observations $y$. We refer to this problem as the "supervised" case. However,…
▽ More
In inverse problems, we often have access to data consisting of paired samples $(x,y)\sim p_{X,Y}(x,y)$ where $y$ are partial observations of a physical system, and $x$ represents the unknowns of the problem. Under these circumstances, we can employ supervised training to learn a solution $x$ and its uncertainty from the observations $y$. We refer to this problem as the "supervised" case. However, the data $y\sim p_{Y}(y)$ collected at one point could be distributed differently than observations $y'\sim p_{Y}'(y')$, relevant for a current set of problems. In the context of Bayesian inference, we propose a two-step scheme, which makes use of normalizing flows and joint data to train a conditional generator $q_θ(x|y)$ to approximate the target posterior density $p_{X|Y}(x|y)$. Additionally, this preliminary phase provides a density function $q_θ(x|y)$, which can be recast as a prior for the "unsupervised" problem, e.g.~when only the observations $y'\sim p_{Y}'(y')$, a likelihood model $y'|x$, and a prior on $x'$ are known. We then train another invertible generator with output density $q'_φ(x|y')$ specifically for $y'$, allowing us to sample from the posterior $p_{X|Y}'(x|y')$. We present some synthetic results that demonstrate considerable training speedup when reusing the pretrained network $q_θ(x|y')$ as a warm start or preconditioning for approximating $p_{X|Y}'(x|y')$, instead of learning from scratch. This training modality can be interpreted as an instance of transfer learning. This result is particularly relevant for large-scale inverse problems that employ expensive numerical simulations.
△ Less
Submitted 15 July, 2020;
originally announced July 2020.
-
Parameterizing uncertainty by deep invertible networks, an application to reservoir characterization
Authors:
Gabrio Rizzuti,
Ali Siahkoohi,
Philipp A. Witte,
Felix J. Herrmann
Abstract:
Uncertainty quantification for full-waveform inversion provides a probabilistic characterization of the ill-conditioning of the problem, comprising the sensitivity of the solution with respect to the starting model and data noise. This analysis allows to assess the confidence in the candidate solution and how it is reflected in the tasks that are typically performed after imaging (e.g., stratigrap…
▽ More
Uncertainty quantification for full-waveform inversion provides a probabilistic characterization of the ill-conditioning of the problem, comprising the sensitivity of the solution with respect to the starting model and data noise. This analysis allows to assess the confidence in the candidate solution and how it is reflected in the tasks that are typically performed after imaging (e.g., stratigraphic segmentation following reservoir characterization). Classically, uncertainty comes in the form of a probability distribution formulated from Bayesian principles, from which we seek to obtain samples. A popular solution involves Monte Carlo sampling. Here, we propose instead an approach characterized by training a deep network that "pushes forward" Gaussian random inputs into the model space (representing, for example, density or velocity) as if they were sampled from the actual posterior distribution. Such network is designed to solve a variational optimization problem based on the Kullback-Leibler divergence between the posterior and the network output distributions. This work is fundamentally rooted in recent developments for invertible networks. Special invertible architectures, besides being computational advantageous with respect to traditional networks, do also enable analytic computation of the output density function. Therefore, after training, these networks can be readily used as a new prior for a related inversion problem. This stands in stark contrast with Monte-Carlo methods, which only produce samples. We validate these ideas with an application to angle-versus-ray parameter analysis for reservoir characterization.
△ Less
Submitted 16 April, 2020;
originally announced April 2020.
-
Self-crossing stable generalized complex structures
Authors:
Gil R. Cavalcanti,
Ralph L. Klaasse,
Aldo Witte
Abstract:
We extend the notion of (smooth) stable generalized complex structures to allow for an anticanonical section with normal self-crossing singularities. This weakening not only allows for a number of natural examples in higher dimensions but also sheds some light into the smooth case in dimension four. We show that in four dimensions there is a natural connected sum operation for these structures as…
▽ More
We extend the notion of (smooth) stable generalized complex structures to allow for an anticanonical section with normal self-crossing singularities. This weakening not only allows for a number of natural examples in higher dimensions but also sheds some light into the smooth case in dimension four. We show that in four dimensions there is a natural connected sum operation for these structures as well as a smoothing operation which changes a self-crossing stable generalized complex structure into a smooth stable generalized complex structure on the same manifold. This allows us to construct large families of stable generalized complex manifolds.
△ Less
Submitted 16 April, 2020;
originally announced April 2020.
-
Serverless seismic imaging in the cloud
Authors:
Philipp A. Witte,
Mathias Louboutin,
Charles Jones,
Felix J. Herrmann
Abstract:
This abstract presents a serverless approach to seismic imaging in the cloud based on high-throughput containerized batch processing, event-driven computations and a domain-specific language compiler for solving the underlying wave equations. A 3D case study on Azure demonstrates that this approach allows reducing the operating cost of up to a factor of 6, making the cloud a viable alternative to…
▽ More
This abstract presents a serverless approach to seismic imaging in the cloud based on high-throughput containerized batch processing, event-driven computations and a domain-specific language compiler for solving the underlying wave equations. A 3D case study on Azure demonstrates that this approach allows reducing the operating cost of up to a factor of 6, making the cloud a viable alternative to on-premise HPC clusters for seismic imaging.
△ Less
Submitted 27 November, 2019;
originally announced November 2019.
-
An Event-Driven Approach to Serverless Seismic Imaging in the Cloud
Authors:
Philipp A. Witte,
Mathias Louboutin,
Henryk Modzelewski,
Charles Jones,
James Selvage,
Felix J. Herrmann
Abstract:
Adapting the cloud for high-performance computing (HPC) is a challenging task, as software for HPC applications hinges on fast network connections and is sensitive to hardware failures. Using cloud infrastructure to recreate conventional HPC clusters is therefore in many cases an infeasible solution for migrating HPC applications to the cloud. As an alternative to the generic lift and shift approa…
▽ More
Adapting the cloud for high-performance computing (HPC) is a challenging task, as software for HPC applications hinges on fast network connections and is sensitive to hardware failures. Using cloud infrastructure to recreate conventional HPC clusters is therefore in many cases an infeasible solution for migrating HPC applications to the cloud. As an alternative to the generic lift and shift approach, we consider the specific application of seismic imaging and demonstrate a serverless and event-driven approach for running large-scale instances of this problem in the cloud. Instead of permanently running compute instances, our workflow is based on a serverless architecture with high throughput batch computing and event-driven computations, in which computational resources are only running as long as they are utilized. We demonstrate that this approach is very flexible and allows for resilient and nested levels of parallelization, including domain decomposition for solving the underlying partial differential equations. While the event-driven approach introduces some overhead as computational resources are repeatedly restarted, it inherently provides resilience to instance shut-downs and allows a significant reduction of cost by avoiding idle instances, thus making the cloud a viable alternative to on-premise clusters for large-scale seismic imaging.
△ Less
Submitted 3 September, 2019;
originally announced September 2019.
-
Devito (v3.1.0): an embedded domain-specific language for finite differences and geophysical exploration
Authors:
Mathias Louboutin,
Michael Lange,
Fabio Luporini,
Navjot Kukreja,
Philipp A. Witte,
Felix J. Herrmann,
Paulius Velesko,
Gerard J. Gorman
Abstract:
We introduce Devito, a new domain-specific language for implementing high-performance finite difference partial differential equation solvers. The motivating application is exploration seismology where methods such as Full-Waveform Inversion and Reverse-Time Migration are used to invert terabytes of seismic data to create images of the earth's subsurface. Even using modern supercomputers, it can t…
▽ More
We introduce Devito, a new domain-specific language for implementing high-performance finite difference partial differential equation solvers. The motivating application is exploration seismology where methods such as Full-Waveform Inversion and Reverse-Time Migration are used to invert terabytes of seismic data to create images of the earth's subsurface. Even using modern supercomputers, it can take weeks to process a single seismic survey and create a useful subsurface image. The computational cost is dominated by the numerical solution of wave equations and their corresponding adjoints. Therefore, a great deal of effort is invested in aggressively optimizing the performance of these wave-equation propagators for different computer architectures. Additionally, the actual set of partial differential equations being solved and their numerical discretization is under constant innovation as increasingly realistic representations of the physics are developed, further ratcheting up the cost of practical solvers. By embedding a domain-specific language within Python and making heavy use of SymPy, a symbolic mathematics library, we make it possible to develop finite difference simulators quickly using a syntax that strongly resembles the mathematics. The Devito compiler reads this code and applies a wide range of analysis to generate highly optimized and parallel code. This approach can reduce the development time of a verified and optimized solver from months to days.
△ Less
Submitted 9 August, 2019; v1 submitted 6 August, 2018;
originally announced August 2018.
