-
How quantum computing can enhance biomarker discovery
Authors:
Frederik F. Flöther,
Daniel Blankenberg,
Maria Demidik,
Karl Jansen,
Raga Krishnakumar,
Rajiv Krishnakumar,
Nouamane Laanait,
Laxmi Parida,
Carl Saab,
Filippo Utro
Abstract:
Biomarkers play a central role in medicine's gradual progress towards proactive, personalized precision diagnostics and interventions. However, finding biomarkers that provide very early indicators of a change in health status, for example for multi-factorial diseases, has been challenging. Discovery of such biomarkers stands to benefit significantly from advanced information processing and means…
▽ More
Biomarkers play a central role in medicine's gradual progress towards proactive, personalized precision diagnostics and interventions. However, finding biomarkers that provide very early indicators of a change in health status, for example for multi-factorial diseases, has been challenging. Discovery of such biomarkers stands to benefit significantly from advanced information processing and means to detect complex correlations, which quantum computing offers. In this perspective paper, quantum algorithms, particularly in machine learning, are mapped to key applications in biomarker discovery. The opportunities and challenges associated with the algorithms and applications are discussed. The analysis is structured according to different data types - multi-dimensional, time series, and erroneous data - and covers key data modalities in healthcare - electronic health records (EHRs), omics, and medical images. An outlook is provided concerning open research challenges.
△ Less
Submitted 9 February, 2025; v1 submitted 15 November, 2024;
originally announced November 2024.
-
An entanglement-based volumetric benchmark for near-term quantum hardware
Authors:
Kathleen E. Hamilton,
Nouamane Laanait,
Akhil Francis,
Sophia E. Economou,
George S. Barron,
Kübra Yeter-Aydeniz,
Titus Morris,
Harrison Cooley,
Muhun Kang,
Alexander F. Kemper,
Raphael Pooser
Abstract:
We introduce a volumetric benchmark for near-term quantum platforms based on the generation and verification of genuine entanglement across n-qubits using graph states and direct stabilizer measurements. Our benchmark evaluates the robustness of multipartite and bipartite n-qubit entanglement with respect to many sources of hardware noise: qubit decoherence, CNOT and swap gate noise, and readout e…
▽ More
We introduce a volumetric benchmark for near-term quantum platforms based on the generation and verification of genuine entanglement across n-qubits using graph states and direct stabilizer measurements. Our benchmark evaluates the robustness of multipartite and bipartite n-qubit entanglement with respect to many sources of hardware noise: qubit decoherence, CNOT and swap gate noise, and readout error. We demonstrate our benchmark on multiple superconducting qubit platforms available from IBM (ibmq_belem, ibmq_toronto, ibmq_guadalupe and ibmq_jakarta). Subsets of $n<10$ qubits are used for graph state preparation and stabilizer measurement. Evaluation of genuine and biseparable entanglement witnesses we report observations of $5$ qubit genuine entanglement, but robust multipartite entanglement is difficult to generate for $n>4$ qubits and identify two-qubit gate noise as strongly correlated with the quality of genuine multipartite entanglement.
△ Less
Submitted 1 September, 2022;
originally announced September 2022.
-
Scalable Balanced Training of Conditional Generative Adversarial Neural Networks on Image Data
Authors:
Massimiliano Lupo Pasini,
Vittorio Gabbi,
Junqi Yin,
Simona Perotto,
Nouamane Laanait
Abstract:
We propose a distributed approach to train deep convolutional generative adversarial neural network (DC-CGANs) models. Our method reduces the imbalance between generator and discriminator by partitioning the training data according to data labels, and enhances scalability by performing a parallel training where multiple generators are concurrently trained, each one of them focusing on a single dat…
▽ More
We propose a distributed approach to train deep convolutional generative adversarial neural network (DC-CGANs) models. Our method reduces the imbalance between generator and discriminator by partitioning the training data according to data labels, and enhances scalability by performing a parallel training where multiple generators are concurrently trained, each one of them focusing on a single data label. Performance is assessed in terms of inception score and image quality on MNIST, CIFAR10, CIFAR100, and ImageNet1k datasets, showing a significant improvement in comparison to state-of-the-art techniques to training DC-CGANs. Weak scaling is attained on all the four datasets using up to 1,000 processes and 2,000 NVIDIA V100 GPUs on the OLCF supercomputer Summit.
△ Less
Submitted 20 February, 2021;
originally announced February 2021.
-
Application of variational policy gradient to atomic-scale materials synthesis
Authors:
Siyan Liu,
Nikolay Borodinov,
Lukas Vlcek,
Dan Lu,
Nouamane Laanait,
Rama K. Vasudevan
Abstract:
Atomic-scale materials synthesis via layer deposition techniques present a unique opportunity to control material structures and yield systems that display unique functional properties that cannot be stabilized using traditional bulk synthetic routes. However, the deposition process itself presents a large, multidimensional space that is traditionally optimized via intuition and trial and error, s…
▽ More
Atomic-scale materials synthesis via layer deposition techniques present a unique opportunity to control material structures and yield systems that display unique functional properties that cannot be stabilized using traditional bulk synthetic routes. However, the deposition process itself presents a large, multidimensional space that is traditionally optimized via intuition and trial and error, slowing down progress. Here, we present an application of deep reinforcement learning to a simulated materials synthesis problem, utilizing the Stein variational policy gradient (SVPG) approach to train multiple agents to optimize a stochastic policy to yield desired functional properties. Our contributions are (1) A fully open source simulation environment for layered materials synthesis problems, utilizing a kinetic Monte-Carlo engine and implemented in the OpenAI Gym framework, (2) Extension of the Stein variational policy gradient approach to deal with both image and tabular input, and (3) Developing a parallel (synchronous) implementation of SVPG using Horovod, distributing multiple agents across GPUs and individual simulation environments on CPUs. We demonstrate the utility of this approach in optimizing for a material surface characteristic, surface roughness, and explore the strategies used by the agents as compared with a traditional actor-critic (A2C) baseline. Further, we find that SVPG stabilizes the training process over traditional A2C. Such trained agents can be useful to a variety of atomic-scale deposition techniques, including pulsed laser deposition and molecular beam epitaxy, if the implementation challenges are addressed.
△ Less
Submitted 28 June, 2020;
originally announced June 2020.
-
Exascale Deep Learning for Scientific Inverse Problems
Authors:
Nouamane Laanait,
Joshua Romero,
Junqi Yin,
M. Todd Young,
Sean Treichler,
Vitalii Starchenko,
Albina Borisevich,
Alex Sergeev,
Michael Matheson
Abstract:
We introduce novel communication strategies in synchronous distributed Deep Learning consisting of decentralized gradient reduction orchestration and computational graph-aware grouping of gradient tensors. These new techniques produce an optimal overlap between computation and communication and result in near-linear scaling (0.93) of distributed training up to 27,600 NVIDIA V100 GPUs on the Summit…
▽ More
We introduce novel communication strategies in synchronous distributed Deep Learning consisting of decentralized gradient reduction orchestration and computational graph-aware grouping of gradient tensors. These new techniques produce an optimal overlap between computation and communication and result in near-linear scaling (0.93) of distributed training up to 27,600 NVIDIA V100 GPUs on the Summit Supercomputer. We demonstrate our gradient reduction techniques in the context of training a Fully Convolutional Neural Network to approximate the solution of a longstanding scientific inverse problem in materials imaging. The efficient distributed training on a dataset size of 0.5 PB, produces a model capable of an atomically-accurate reconstruction of materials, and in the process reaching a peak performance of 2.15(4) EFLOPS$_{16}$.
△ Less
Submitted 24 September, 2019;
originally announced September 2019.
-
Competing Phases in Epitaxial Vanadium Dioxide at Nanoscale
Authors:
Yogesh Sharma,
Martin V. Holt,
Nouamane Laanait,
Xiang Gao,
Ilia Ivanov,
Liam Collins,
Changhee Sohn,
Zhaoliang Liao,
Elizabeth Skoropata,
Sergei V. Kalinin,
Nina Balke,
Gyula Eres,
Thomas Z. Ward,
Ho Nyung Lee
Abstract:
Phase competition in correlated oxides offers tantalizing opportunities as many intriguing physical phenomena occur near the phase transitions. Owing to a sharp metal-insulator transition (MIT) near room temperature, correlated vanadium dioxide (VO2) exhibits a strong competition between insulating and metallic phases that is important for practical applications. However, the phase boundary underg…
▽ More
Phase competition in correlated oxides offers tantalizing opportunities as many intriguing physical phenomena occur near the phase transitions. Owing to a sharp metal-insulator transition (MIT) near room temperature, correlated vanadium dioxide (VO2) exhibits a strong competition between insulating and metallic phases that is important for practical applications. However, the phase boundary undergoes strong modification when strain is involved, yielding complex phase transitions. Here, we report the emergence of the nanoscale M2 phase domains in VO2 epitaxial films under anisotropic strain relaxation. The phase states of the films are imaged by multi-length-scale probes, detecting the structural and electrical properties in individual local domains. Competing evolution of the M1 and M2 phases indicates a critical role of lattice-strain on both the stability of the M2 Mott phase and the energetics of the MIT in VO2 films. This study demonstrates how strain engineering can be utilized to design phase states, which allow deliberate control of MIT behavior at the nanoscale in epitaxial VO2 films.
△ Less
Submitted 13 August, 2019;
originally announced August 2019.
-
USID and Pycroscopy -- Open frameworks for storing and analyzing spectroscopic and imaging data
Authors:
Suhas Somnath,
Chris R. Smith,
Nouamane Laanait,
Rama K. Vasudevan,
Anton Ievlev,
Alex Belianinov,
Andrew R. Lupini,
Mallikarjun Shankar,
Sergei V. Kalinin,
Stephen Jesse
Abstract:
Materials science is undergoing profound changes due to advances in characterization instrumentation that have resulted in an explosion of data in terms of volume, velocity, variety and complexity. Harnessing these data for scientific research requires an evolution of the associated computing and data infrastructure, bridging scientific instrumentation with super- and cloud- computing. Here, we de…
▽ More
Materials science is undergoing profound changes due to advances in characterization instrumentation that have resulted in an explosion of data in terms of volume, velocity, variety and complexity. Harnessing these data for scientific research requires an evolution of the associated computing and data infrastructure, bridging scientific instrumentation with super- and cloud- computing. Here, we describe Universal Spectroscopy and Imaging Data (USID), a data model capable of representing data from most common instruments, modalities, dimensionalities, and sizes. We pair this schema with the hierarchical data file format (HDF5) to maximize compatibility, exchangeability, traceability, and reproducibility. We discuss a family of community-driven, open-source, and free python software packages for storing, processing and visualizing data. The first is pyUSID which provides the tools to read and write USID HDF5 files in addition to a scalable framework for parallelizing data analysis. The second is Pycroscopy, which provides algorithms for scientific analysis of nanoscale imaging and spectroscopy modalities and is built on top of pyUSID and USID. The instrument-agnostic nature of USID facilitates the development of analysis code independent of instrumentation and task in Pycroscopy which in turn can bring scientific communities together and break down barriers in the age of open-science. The interested reader is encouraged to be a part of this ongoing community-driven effort to collectively accelerate materials research and discovery through the realms of big data.
△ Less
Submitted 27 March, 2019; v1 submitted 22 March, 2019;
originally announced March 2019.
-
Reconstruction of 3-D Atomic Distortions from Electron Microscopy with Deep Learning
Authors:
Nouamane Laanait,
Qian He,
Albina Y. Borisevich
Abstract:
Deep learning has demonstrated superb efficacy in processing imaging data, yet its suitability in solving challenging inverse problems in scientific imaging has not been fully explored. Of immense interest is the determination of local material properties from atomically-resolved imaging, such as electron microscopy, where such information is encoded in subtle and complex data signatures, and whos…
▽ More
Deep learning has demonstrated superb efficacy in processing imaging data, yet its suitability in solving challenging inverse problems in scientific imaging has not been fully explored. Of immense interest is the determination of local material properties from atomically-resolved imaging, such as electron microscopy, where such information is encoded in subtle and complex data signatures, and whose recovery and interpretation necessitate intensive numerical simulations subject to the requirement of near-perfect knowledge of the experimental setup. We demonstrate that an end-to-end deep learning model can successfully recover 3-dimensional atomic distortions of a variety of oxide perovskite materials from a single 2-dimensional experimental scanning transmission electron (STEM) micrograph, in the process resolving a longstanding question in the recovery of 3-D atomic distortions from STEM experiments. Our results indicate that deep learning is a promising approach to efficiently address unsolved inverse problems in scientific imaging and to underpin novel material investigations at atomic resolution.
△ Less
Submitted 18 February, 2019;
originally announced February 2019.
-
Mapping mesoscopic phase evolution during e-beam induced transformations via deep learning of atomically resolved images
Authors:
Rama K. Vasudevan,
Nouamane Laanait,
Erik M. Ferragut,
Kai Wang,
David B. Geohegan,
Kai Xiao,
Maxim A. Ziatdinov,
Stephen Jesse,
Ondrej E. Dyck,
Sergei V. Kalinin
Abstract:
Understanding transformations under electron beam irradiation requires mapping the structural phases and their evolution in real time. To date, this has mostly been a manual endeavor comprising of difficult frame-by-frame analysis that is simultaneously tedious and prone to error. Here, we turn towards the use of deep convolutional neural networks (DCNN) to automatically determine the Bravais latt…
▽ More
Understanding transformations under electron beam irradiation requires mapping the structural phases and their evolution in real time. To date, this has mostly been a manual endeavor comprising of difficult frame-by-frame analysis that is simultaneously tedious and prone to error. Here, we turn towards the use of deep convolutional neural networks (DCNN) to automatically determine the Bravais lattice symmetry present in atomically-resolved images. A DCNN is trained to identify the Bravais lattice class given a 2D fast Fourier transform of the input image. Monte-Carlo dropout is used for determining the prediction probability, and results are shown for both simulated and real atomically-resolved images from scanning tunneling microscopy and scanning transmission electron microscopy. A reduced representation of the final layer output allows to visualize the separation of classes in the DCNN and agrees with physical intuition. We then apply the trained network to electron beam-induced transformations in WS2, which allows tracking and determination of growth rate of voids. These results are novel in two ways: (1) It shows that DCNNs can be trained to recognize diffraction patterns, which is markedly different from the typical "real image" cases, and (2) it provides a method with in-built uncertainty quantification, allowing the real-time analysis of phases present in atomically resolved images.
△ Less
Submitted 28 February, 2018;
originally announced February 2018.
-
Correlated Disorder in van der Waals Heterostructures
Authors:
Nouamane Laanait,
Zhan Zhang,
Christian M. Schleputz,
Ying Liu,
Michael Wojcik,
Rachael L. Myers-Ward,
D. Kurt Gaskill,
Paul Fenter,
Lian Li
Abstract:
The individual building blocks of van der Waals (vdW) heterostructures host fascinating physical phenomena, ranging from ballistic electron transport in graphene to striking optical properties of MoSe2 sheets. The presence of bonded and non-bonded cohesive interactions in a vdW heterostructure, promotes diversity in their structural arrangements, which in turn profoundly modulate the properties of…
▽ More
The individual building blocks of van der Waals (vdW) heterostructures host fascinating physical phenomena, ranging from ballistic electron transport in graphene to striking optical properties of MoSe2 sheets. The presence of bonded and non-bonded cohesive interactions in a vdW heterostructure, promotes diversity in their structural arrangements, which in turn profoundly modulate the properties of their individual constituents. Here, we report on the presence of correlated structural disorder coexisting with the nearly perfect crystallographic order along the growth direction of epitaxial vdW heterostructures of Bi2Se3/graphene/SiC. Using the depth penetration of X-ray diffraction microscopy and scattering, we probed their crystal structure from atomic to mesoscopic length scales, to reveal that their structural diversity is underpinned by spatially correlated disorder states. The presence of the latter induces on a system, widely considered to behave as a collection of nearly independent 2-dimensional units, a pseudo-3-dimensional character, when subjected to epitaxial constraints and ordered substrate interactions. These findings shed new light on the nature of the vast structural landscape of vdW heterostructures and could enable new avenues in modulating their unique properties by correlated disorder.
△ Less
Submitted 12 June, 2016;
originally announced June 2016.
-
Monovalent Ion Condensation at the Electrified Liquid/Liquid Interface
Authors:
Nouamane Laanait,
Jaesung Yoon,
Binyang Hou,
Petr Vanysek,
Mati Meron,
Binhua Lin,
Guangming Luo,
Ilan Benjamin,
Mark L. Schlossman
Abstract:
X-ray reflectivity studies demonstrate the condensation of a monovalent ion at the electrified interface between electrolyte solutions of water and 1,2-dichloroethane. Predictions of the ion distributions by standard Poisson-Boltzmann (Gouy-Chapman) theory are inconsistent with these data at higher applied interfacial electric potentials. Calculations from a Poisson-Boltzmann equation that incorpo…
▽ More
X-ray reflectivity studies demonstrate the condensation of a monovalent ion at the electrified interface between electrolyte solutions of water and 1,2-dichloroethane. Predictions of the ion distributions by standard Poisson-Boltzmann (Gouy-Chapman) theory are inconsistent with these data at higher applied interfacial electric potentials. Calculations from a Poisson-Boltzmann equation that incorporates a non-monotonic ion-specific potential of mean force are in good agreement with the data.
△ Less
Submitted 23 June, 2010;
originally announced June 2010.
