TACCSTER 2023 Proceedings

Permanent URI for this collectionhttps://hdl.handle.net/2152/122160


Recent Submissions

Now showing 1 - 20 of 30
  • Item
    A Parallelized Viscous Vorticity Numerical Model for Predicting Propeller Performance in Turbulent Flows
    (2023-09-28) You, Rui; Kinnas, Spyros A.
    VISVE, a computational tool developed with the support of the Office of Naval Research (ONR), is specifically designed for analyzing marine propeller performance in turbulent flows. To facilitate the computational demands of the simulations, all the development work and tests were conducted on TACC-affiliated supercomputers. The performance of the code was analyzed through profiling, identifying the most time-consuming components. To optimize the computational efficiency, the code was parallelized using a hybrid OpenMP and MPI strategy. Scalability tests were conducted on the ICX nodes, Stampede2, with a focus on a 3-D propeller operating within turbulent flows. This study contributes to the development of a robust numerical model, which can be effectively utilized for analyzing propeller performance in complex turbulent flow conditions. The parallelization strategy employed in VISVE significantly enhances its computational capabilities, making it suitable for high-performance computing platforms at TACC.
  • Item
    Validating Alzheimer’s Disease Atrophy from Heatmap-based Explainable AI Methods with a Large Meta-Analysis of Neuroimaging Studies
    (2023-09-28) Wang, Di
    Deep learning has shown great potential in Alzheimer's disease (AD) research, but its complexity makes interpretation and explanation challenging. To address this, heatmap-based explainable artificial intelligence (XAI) methods have emerged as a popular tool for visually interpreting deep learning models. However, when no ground truth is available, evaluating and validating the quality of these heatmaps becomes difficult. Our study aimed to address this challenge by quantifying the overlap between heatmaps generated by deep neural networks for Alzheimer's disease classification and a ground truth map obtained from a large meta-analysis. Using T1-weighted MRI scans from the ADNI dataset, we trained 3D CNN classifiers and employed three state-of-the-art XAI heatmap methods: Layer-wise Relevance Propagation (LRP), Integrated Gradients (IG), and Guided grad-CAM (GGC). The heatmaps produced by these methods were compared to a binary brain map derived from a meta-analysis of voxel-based morphometry studies conducted on other T1 MRI scans. Remarkably, all three heatmap methods effectively captured brain regions that showed significant overlap with the meta-analysis map, with the IG method demonstrating the most promising results. Moreover, the performance of the three heatmap methods outperformed that of linear Support Vector Machine (SVM) models, indicating that using the latest heatmap techniques to analyze deep nonlinear models can generate more meaningful brain maps compared to linear and shallow models. In conclusion, our research highlights the potential of heatmap visualization techniques in comprehending the effects of Alzheimer's disease on brain regions. By enhancing interpretability and explicability, these methods contribute to the advancement of AD research and hold promise for potential applications in other neurological conditions in the future.
  • Item
    Investigating Vibrio Cholerae ToxT-Ligand Interactions Using GROMACS
    (2023-09-28) Castellanos, Hugo Villar; Touhami, Ahmed; Hanke, Andreas
    Vibrio cholerae is a bacterium responsible for the potentially fatal disease known as Cholera, and the ToxT protein is responsible for the transcription of most virulence genes in the bacterium. The purpose of this preliminary study is to investigate the binding affinity of a ligand that attaches to the DNA binding domain in ToxT using the GROMACS molecular dynamics package. Based on a literature review, a drug-like ligand (ZINC database id 14749003) was selected for this study. Docking of the ligand to ToxT using Molegro Virtual Docker (MVD) showed two prevalent binding sites (residues Arg214 and Arg187); the ligand binds to the protein via two hydrogen bonds with each site, with binding energies of -1.35 kcal/mol and -0.8 kcal/mol for Arg214, -1.6 kcal/mol and -1.46 kcal/mol for Arg187. Then analysis of a 1-ns molecular dynamics run (on Lonestar6) of the docked ligand-protein complex showed that the attachment of the ligand to the protein was relatively stable, with a maximum hydrogen bond occupancy rate of 31.7% and with the top donor-acceptor pairs involving residues Arg214 and Arg187, which validates the result from MVD. The conformation of the complex was also stable with a radius of gyration fluctuating around a value of 1.9 nm.
  • Item
    Multi-GPU FFT Matvec for Inverse Problems Involving Shift-Invariant Systems
    (2023-09-28) Venkat, Sreeram; Fernando, Milinda; Henneking, Stefan; Ghattas, Omar
    Hessian-based algorithms for the solution to inverse problems typically require many actions of the Hessian matrix on a vector (matvecs). A direct approach is often computationally intractable for problems with high-dimensional parameter fields or expensive-to-evaluate forward models. For systems that exhibit shift-invariance (e.g. autonomous systems) structure in their discretized form, the discretized linear parameter-to-observable (p2o) maps are block Toeplitz matrices. Moreover, considering causality for time-invariant systems the p2o map and its adjoint are lower- and upper-triangular block Toeplitz, respectively. By exploiting this structure, Hessian matrices for these types of systems can be compactly represented and Hessian matvecs can be efficiently computed through scalable multi-GPU FFT matvecs. Compact representation follows directly from the definition of Block Toeplitz matrices. Fast matrix-vector multiplication is achieved by embedding the block Toeplitz matrix within a block circulant matrix which is diagonalized by the Discrete Fourier Transform. The matrix-vector product then becomes an element-wise vector operation in Fourier space. Furthermore, the action of the adjoint p2o map corresponds to simply applying the complex conjugate in Fourier space, eliminating the need to separately store the Fourier-transformed forward and adjoint map. Exploiting the triangular block Toeplitz structure in this way yields memory savings proportional to the number of time steps Nt and a computational speedup of O(Nt/ log(Nt)). In the context of explicit methods that are suitable for GPU-based computation, the number of time steps is typically very large due to the CFL condition, making the savings of the algorithm substantial. We develop a multi-GPU FFT matvec code for Hessians corresponding to block Toeplitz p2o matrices utilizing the cuFFT and NCCL libraries. Our implementation achieves 75-90% of the maximum memory bandwidth on NVIDIA A100 80GB GPUs for all custom GPU kernels — which correspond to memory-bound operations. We also show strong and weak multi-GPU scaling on the Frontera RTX nodes with up to 81 GPUs.
  • Item
    Graph Theory Modeling on the BrainMap Community Portal: Network topology reveals a central role for the medial frontal gyrus in mesial temporal lobe epilepsy
    (2023-09-28) Towne, Jonathan M.; Eslami, Vahid; Cavazos, José E.; Fox, Peter T.
    "Mesial temporal lobe epilepsy(MTLE) is a disorder of neural networks, often amenable to surgical treatment. Yet, resection of the seizure-onset zone can be non-curative. Seizure recurrence is attributable to distributed network-mediation of ictogenesis(seizure-onset). To address reasons for surgical failure, it is crucial to understand higher-level network properties in MTLE pathology. Graph theory models(GTM) interrogate network properties by quantifying features of network topology. Imaging studies have leveraged graph theory to detect compensatory network changes and predict seizure-onset laterality in MTLE, portending GTM utility in biomarker development. GTM was recently adapted for coordinate-based meta-analysis(CBMA) of voxel-based morphometry(VBM) and physiology(VBP) studies, instantiating a multi-variate extension of activation likelihood estimation(mass-univariate-CBMA). We applied meta-analytic-GTM(M-GTM) to VBM/VBP-studies of MTLE, to infer organizational properties of MTLE network pathology. To quantify topological organization of MTLE network pathology, BrainMap applications (M-GTM/Mango) were used to derive a co-alteration GTM from 74 experiments. M-GTM was used to model coordinates of MTLE pathology as spatial probability distributions, defining nodes at peaks in the joint distribution of pathology and computing edges as co-alterations in individual studies. Clusters of network pathology (modules) were detected via spectral-partition and interpreted in Mango via Regional Behavioral/Diseases Analyses of the BrainMap database(21,435-Task-Activation/4,398-VBM-experiments;n=107,167/115,627-subjects). MATLAB/Cytoscape were used to compute topology metrics and nodal influence on MTLE network topology. Two distributed network-modules were identified in the MTLE co-alteration network, connected only via three most-influential nodes: hippocampus/MDN-thalamus/medial frontal gyrus. Module-1 regions(M1:mesial-temporal/deep-nuclear/frontal/precentral/postcentral/cingulate/inf-parietal structures) were associated with emotion-cognition(Z=3.7) and weakly with social-cognition/explicit-memory(Z=2.5/2.3). Module-2(M2:cerebellar/occipital/temporoparietal/precentral/inf.frontal/med.frontal gyri) regional associations included language/speech/semantic-cognition(Z=4.1/3.0/2.3). M1 matched known VBM-patterns in Alzheimer’s(Z=4.1); both M1/M2 matched those of structural epilepsy pathology(Z=3.4/3.6). A 2-node module(M3:parahippocampus/amygdala) and disconnected pair(M4:mid.frontal/cingulate gyri) were noted. Discrete co-alteration networks exist in MTLE. The medial frontal gyrus likely mediates interactions and evolution of limbic-M1 and verbal-M2 symptoms in MTLE. Pathology modules and intermodular connections represent potential targets for disease monitoring/therapeutic modulation. This study was funded by R01MH074457, T32GM113896, T32TR004545, F31NS131025, and the American Epilepsy Society."
  • Item
    Distinguishing Gravitational Wave Parameter Markers: Eccentricity vs Precession
    (2023-09-28) Tibrewal, Snehal; Iglesias, Hector; Lange, Jacob; Ferguson, Deborah; Shoemaker, Deirdre
    When two massive accelerating objects collide, they generate ripples in the space-time continuum, referred to as gravitational waves (GWs). These GW signals can be detected by the LIGO detectors and provide us with information on the merging binary black holes (BBH). Our work aims to study the distinguishability between eccentricity and spin precession - two important parameters of a GW signal. There has been evidence of ambiguity when attempting a recovery of these parameters due to similarities in their GW markers. Since real data is limited and our knowledge is mostly biased by the accuracy of search pipelines, we are using injections that are signals from simulated BBH mergers. These simulated signals are obtained using numerical relativity (NR) and include both eccentricity and spin precession, something which previous studies haven't been able to mimic due to limitations of current day waveform models. TACC plays a central role in our ability to run these NR simulations. Without supercomputing resources like TACC, the computational complexity of solving Einstein's Field Equations can take months, or even years. We utilize our in-house NR code Maya-Waves to run simulations for various configurations. With these simulations, we are able to control the explorable parameter space as well as test the accuracy of our recovery pipelines. Additionally, these NR simulations will be highly advantageous to the future development of more accurate waveform models. Our preliminary results point towards possible issues with distinguishability, especially in certain parameter spaces. However we continue to run experiments to be able to comment more clearly on these bounds.
  • Item
    What Happens When SAV Fleets Compete: A Fare-Based Analysis
    (2023-09-28) Sambasivam, Balasubramanian; Gurumurthy, Krishna Murthy; Kockelman, Kara M.
    In cities across the world, especially under dense settings, transportation network companies (TNCs) account for a non-trivial mode share. With the introduction of shared autonomous vehicles (SAVs), ridehailing will become more common. When multiple TNCs or SAV fleets compete for customers, fares, wait times, and other performance metrics are affected. This study simulates competition between two SAV operators for the Bloomington, Illinois region with a profit-maximization objective. Fare strategies such as a time-of-day (TOD) factor, zone- based surge pricing (ZSP), and a combination of the two are simulated. Customers are assumed to choose operators based on low fares, after accounting for bias in operator preference. Results from a large-scale agent-based simulator called POLARIS suggest that the implementation of the TOD and ZSP simultaneously appears to be advantageous for fleet operators, as it helps increase their profit. While dynamic ridesharing is beneficial for passengers by reducing their fares, it leads to losses for SAV operators. Fleet size and, consequently, coverage for request assignment impacts profit.
  • Item
    Medical Data Augmentation via ChatGPT: A Case Study on Medication Identification and Medication Event Classification
    (2023-09-28) Sarker, Shouvon; Li, Xiangfang; Dong, Xishuang; Qian, Lijun
    The identification of key factors such as medications, diseases, and relationships within electronic health records and clinical notes has a wide range of applications in the clinical field. In the N2C2 2022 competitions, various tasks were presented to promote the identification of key factors in electronic health records (EHRs) using the Contextualized Medication Event Dataset (CMED). Pretrained large language models (LLMs) demonstrated exceptional performance in these tasks. This study aims to explore the utilization of LLMs, specifically ChatGPT, for data augmentation to overcome the limited availability of annotated data for identifying the key factors in EHRs. Additionally, different pre-trained BERT models, initially trained on extensive datasets like Wikipedia and MIMIC, were employed to develop models for identifying these key variables in EHRs through fine-tuning on augmented datasets. The experimental results of two EHR analysis tasks, namely medication identification and medication event classification, indicate that data augmentation based on ChatGPT proves beneficial in improving performance for both medication identification and medication event classification.
  • Item
    Probabilistic Assessment of Coastal Bridge Vulnerability to Wave Loading During Hurricanes
    (2023-09-28) Pervaiz, Fahad; Hummel, Michelle
    Bridges located in coastal areas may be subjected to extreme wave and surge loading during coastal storm events, potentially leading to damage or even collapse. Understanding the loading and response of bridges during storms is critical to ensuring the resilience and reliability of transportation infrastructure in coastal regions. This study develops and applies a probabilistic bridge vulnerability framework that allows for coupled simulation of time-varying wave loading using OpenFOAM, dynamic structural response using OpenSees, and uncertainty quantification using Dakota. The framework is applied to evaluate bridge failure potential under wave and surge loading based on conditions observed during Hurricane Ike. Simulations were run on the Stampede2 system at TACC. Results indicate that waves with high peak periods and long wavelengths are the primary contributors to bridge instability. During Hurricane Ike, extreme wave loading exceeds the resistance of the typical bridge structure over 70% of the time, leading to sliding, uplifting, and overturning. Vertical stability of the bridge is most sensitive to uncertainties in the concrete density, while horizontal stability is most sensitive to uncertainties in the lateral stiffness of the bearings. Results from this research can inform the development of improved design standards for coastal bridges and/or the selection of appropriate countermeasures to improve bridge stability. The proposed computational framework can be applied across transportation systems to quantify failure probabilities and prioritize maintenance and retrofitting efforts for coastal bridges.
  • Item
    Coupled Magnetic Domain Wall Oscillations for Neuromorphic Lateral Excitation Behavior in Periodic Ferromagnetic Nanowire Arrays
    (2023-09-28) Park, Jiwoo; Rogers, Vivian; Incorvia, Jean Anne
    "The Domain Wall (DW) racetrack is a ferromagnetic nanowire that encodes data in the spatial position of magnetization boundary—a magnetic DW. Many emerging spintronic nanodevices utilize DW racetracks for post-Von-Neumann computing applications, as nonvolatile analog information encoded in DW position can be modified and later decoded by magnetic tunnel junctions (MTJ). We explore the physics of DW oscillations for use in neuromorphic computing, an emerging computer architecture inspired by biological spiking neural systems. In neurons, lateral excitation and resonant firing are behaviors where neurons excite their neighbors without direct connections or spike when driven at a certain frequency. Observing these behaviors in DW racetracks is of great value to us in terms of power-efficiency and computing capability, by natively encoding neuro-mimetic behaviors in the hardware. To study this, we construct a mathematical model where we treat DW racetracks as damped harmonic oscillators with magnetostatic interactions. Then, a Green’s function analysis is applied to the Hamiltonian of the system, obtaining the transfer function of racetracks responding to an external magnetic field or current. By utilizing the transfer function of a racetrack with the convolution theorem, we estimate any DW position displacement resulting from signal applied to a neighboring racetrack. To test the model, we simulated CoFeB racetracks in the Mumax3 micromagnetic software on TACC’s Lonestar6. As a result, coupled oscillations were observed, implying the possibility of lateral excitation. We predict magnetically co-aligned racetracks would allow neuro-realistic behavior, where a bias current tilt the DW pinning potential so that oscillations allow the DW to de-pin and fire easier; magnetically anti-aligned racetracks would exhibit DW oscillations due to the intrinsic resonance of the magnetic stray field DW interactions. The results show the frequency-dependent behaviors of magnetically coupled DW-MTJ racetracks and their possible application in nanoelectronic filters, antennas, and neuromorphic computing."
  • Item
    Potential Roles of Histone Deacetylase 1 and 2 in Regulation of Tubulin Acetylation and Phagocytosis
    (2023-09-29) Owusu, Christian Adu
    Although histone deacetylases (HDACs) were discovered originally to regulate gene expression through the regulation of the chromatin structure by removal of acetylation of the core histone proteins, a structural component of chromatin, recent studies showed that HDACs also regulate the acetylation of non-histone proteins and play crucial roles in regulating various cellular processes, including gene expression, cellular differentiation, cancer development and immune responses. Recently, we showed that HDAC2 plays a critical role in enhanced inflammatory cytokine IL-1β production macrophages by regulation of NLRP3 inflammasome. In this study, we investigated the involvement of HDAC1 and HDAC2 in macrophage phagocytosis of Mycobacterium tuberculosis, a devastating human intracellular microbial pathogen. Our preliminary data showed that inhibition of HDAC1 and HDAC2 affects phagocytosis by regulation of tubulin acetylation, a process known to be regulated by HDAC6. To understand these unexpected roles of HDAC1 and HDAC2 in regulation of tubulin acetylation, we propose to study the structural basis of chemical inhibitors of HDAC1 and HDAC2 and their potential interaction with HDAC6. In a similar approach, we will also study the potential interaction of the HDAC6 inhibitors with HDAC1 and HDAC2 using molecular docking and other relevant in silico analysis using Texas Supercomputer. The functional significance of the findings from these studies will be confirmed with CRISPR-Cas9 mediated gene editing approach to definitively determine the roles of HDAC1 and HDAC2 in regulation of tubulin acetylation, a critical posttranslational modification of cellular tubulin with significance in cellular functions including phagocytosis. The findings from these studies will help us understand the novel functions of HDACs and guide design novel small molecule-based therapeutics for the better control of tuberculosis infection.
  • Item
    Role of Perisynaptic Astroglia during Plasticity in a Synaptic Cluster
    (2023-09-28) Nam, Andrea J.; Kuwajima, Masaaki; Mendenhall, John M.; Parker, Patrick H.; Hubbard, Dusten D.; Hanka, Dakota C.; Abraham, Wickliffe C.; Harris, Kristen M.
    While astrocytes have long been recognized for their involvement in countless vital synaptic processes, acknowledgement of their active participation at the tripartite synapse has been comparatively recent. Astrocytes occupy largely non-overlapping domains, and perisynaptic astrocytic processes (PAPs) interdigitate the dense neuropil where they influence information processing. Previous studies of astrocyte nanoarchitecture show that astrocytes structurally associate with distinct synaptic clusters. Astrocyte calcium elevations that modulate neuronal activity may influence coordinated activation of synapses within such clusters. However, the astrocytic role in modulating synapse clusters during plasticity remains unknown. Here we leveraged 3D reconstruction from serial section electron microscopy (3DEM) and machine learning to investigate whether PAPs enhance synaptic plasticity in a cluster. An automated pipeline was developed to explore PAP morphological changes 30 minutes and 2 hours following long-term potentiation (LTP) or concurrent long-term depression (cLTD) induction, widely accepted cellular mechanisms of learning and memory. LTP and cLTD were induced in vivo in the rat dentate gyrus middle and outer molecular layers, respectively, via delta-burst stimulation of the medial perforant pathway. The contralateral hemispheres received medial path baseline stimulation to serve as within-subject controls. Moreover, serial section electron microscopy was used to capture high resolution images of the dentate gyrus neuropil. Preliminary analyses revealed that over 70% of dentate gyrus synapses exhibited PAP association at the axon-spine interface (ASI). The ASI length with astrocytic coverage scaled linearly with synapse size, thus maintaining proportional availability of astrocytic resources. In addition, the relationship between astrocytic contact length and individual synapse size was expanded in parallel with synapse growth during LTP. Synapses in closer proximity to each other were more similar in size during both control and LTP conditions. Thus, these early results suggest astrocytic processes respond to or coordinate changes in synapse size in synaptic clusters.
  • Item
    Ortho Effect or Ortho-fluoro Effect: A Comparison of the C—CN Bond Activation of Substituted Benzonitriles with [Ni(dmpe)] Fragment using DFT Calculations at TACC
    (2023-09-28) Meza, Jessica; Chowdhury, Jahid Inam; Atesin, Abdurrahman; Jones, William D.; Ateşin, Tülay
    Carbon—carbon (C—C) bond activation by homogeneous transition-metal catalysts is an active area of research due to its wide range of applications in industry and complex organic molecule synthesis. Despite its significance, the activation of thermodynamically stable and kinetically inert C—C s-bonds under mild homogeneous conditions still remains a challenge. The activation is primarily limited to systems in which either relief of strain or aromatization serves as a driving force. An exception to this is the activation of unstrained C—CN bonds of nitriles. In this study, we will report a comparison of the C—CN bond activation of methyl and trifluoromethyl substituted benzonitriles to the fluorinated benzonitriles with the [Ni(dmpe)] fragment by using DFT calculations at TACC. The energetics of the resulting nickel complexes and the C—CN bond activation products in a polar solvent (tetrahydrofuran) and a nonpolar solvent (toluene) will be presented. Reactions with aryl groups containing two ortho substituents will be compared to those with a single ortho substituent. This research will expediate the development of this powerful reaction and lead to more efficient synthesis of complex organic molecules.
  • Item
    Numerical Stability of Mixed-Precision Training on GPU Systems
    (2023-09-28) Mehta, Chintan; Colmenares-Diaz, Eduardo
    The accelerating growth in size of modern AI models calls for more computing power. This trend has motivated top industry players, including Nvidia and Intel, to prioritize hardware optimization, with the goal of accelerating training for state-of-the-art AI models. More recently, these companies have proposed using mixed-precision computation as a potential optimization method. This method applies high precision to sensitive model areas and lower precision to lesser sensitive ones. As mixed-precision methods gain more popularity, studying their numerical stability becomes increasingly important. Our research aims to contribute to understanding the numerical stability of mixed-precision methods when applied to a popular optimization algorithm. We chose to use the mini-batch gradient descent (GD) algorithm in this study, due to its high suitability for parallelization. Leveraging TACC's Lonestar6 GPU nodes, we were able to train a 3-layer image classification model on the popular MNIST image dataset. For model training, we used double precision (FP64) as the baseline configuration. The mixed-precision experiments used half precision (FP16) for all calculations except the gradient accumulation, which was done in double precision. Through empirical analysis of our mixed-precision algorithm, we have obtained some promising results. We were able to significantly accelerate model training using mixed-precision computation, with the accuracy comparable to the standard double precision computation. In addition, experimental results strongly favor the use of larger, over-parameterized models for low precision training. Using double precision (FP64) to accumulate gradients was helpful in maintaining numerical stability for the gradient values. At the same time, half precision (FP16) effectively reduced memory usage during forward and backward propagation. In conclusion, the experimental findings from this study provide a strong motivation to further examine the numerical stability of mixed precision techniques, in order to accelerate the training of scalable AI architectures.
  • Item
    A Parametric Study of Oxygen Ion Cyclotron Harmonic Wave Excitation and Polarization by an Oxygen Ring Distribution
    (2023-09-28) Liu, Xu; Chen, Lunjin; Sun, Jicheng; Wang, Xueyi; Usanova, Maria E.
    Oxygen ion cyclotron harmonic (OCH) waves are electromagnetic emissions with frequencies near the harmonics of the oxygen ion cyclotron frequency. They are ubiquitously observed in the Earth’s magnetosphere. These waves can be excited by an energetic O+ ring distribution. Here, we perform a parametric study of OCH waves by an O+ ring distribution. We investigate the effects of ring concentration (ηho), velocity (vr), temperature (Tr), total O+ concentration (ηo), and wave normal angles (WNAs) on the wave growth rate and polarization. We find that four-wave modes are related to OCH waves. The growth rates and frequency range increase with ηho and ηo and decrease with Tr. The peak growth rate roughly follows the first peak of Jn2 (square of the Bessel function corresponding to the O+ ring) or cold plasma wave modes, which can be used to explain the vr and WNA dependences. OCH waves shift from the transverse mode to the compressional mode as vr increases. This work used TACC to perform particle-in-cell (PIC) simulation (part of Figure 1 of this work) and was published in Journal of Geophysical Research-Space Physics.
  • Item
    Accelerating Discoveries with Large Language Models
    (2023-09-29) Kumar, Krishna
  • Item
    Modeling of Tunable Spike-Timing-Dependent Plasticity (STDP) Neuromorphic Devices using Magnetic Skyrmion Manipulation Chambers
    (2023-09-28) Khodzhaev, Zulfidin; Incorvia, Jean Anne
    "Magnetic skyrmions are promising candidates for implementing synaptic plasticity in neuromorphic computing due to their dynamic and nonvolatile nature [1]. This work presents the modeling of a neuromorphic device with three magnetic skyrmion manipulation chambers, Fig. 1, to emulate spike-timing-dependent plasticity (STDP) [2]. The middle chamber stores the synaptic weight as the skyrmion count, while the side chambers generate pre- and post-synaptic spikes. By controlling the timing and magnitude of currents applied to the chambers, the final skyrmion count can be tuned to demonstrate STDP learning rules [3]. The device exhibits configurable rates of weight update, emulating location-dependent plasticity along dendrites. The modeled skyrmion device demonstrates adaptability in implementing diverse synaptic plasticity functions for efficient neuromorphic computing. [1] A. Fert, N. Reyren, and V. Cros, Nature Reviews Materials 2017 2:7 2(7), 1–15 (2017). [2] G.Q. Bi, and M.M. Poo, Journal of Neuroscience 18(24), 10464–10472 (1998). [3] R.C. Froemke, M.M. Poo, and Y. Dan, Nature 2005 434:7030 434(7030), 221–225 (2005)."
  • Item
    Subspace Methods in Era of Quantum Computing
    (2023-09-28) Jerke, Jonathan; Poirier, Bill
    "I am asking essential questions about the place of classical computing algorithms in light of the quantum supremacy claimed by GOOGLE. GOOGLE claimed they could generate the bonafide quantum states via an iteration of a matrix on a vector. I expect that someday prepared states will be used in another stage of quantum algorithms. However, is there a place for understanding as another tool besides the dominant properties of the oracle? In particular, if one writes down a subspace of active dynamical dimensions with exact quantum operators, could this understanding be faster than state preparation? I have been experimenting with cheaper solutions using a company meerkat (system76) computer. The same algorithm on HPC, which I have time on Lonestar 6, could computer bigger subspaces and more Qbits. In the figure, reporting the timing of ground state computations as a function of Qbit size shows power-law scaling with increasing Qbits. Subspaces dropped into QuTip can achieve fast computation of time dependence with Lindband and dissipation. The limiting factor is the subspace's linear dependence (quality), which has routinely been suitable for other problems. I ask the reader to consider the value of subspace methods irrespective of the quality of my production. Unlike most methods to diagonalize on the market, this method uses forward matrix-vector multiplication on Sums Of Products vector representations. The method was originally designed to work with Coloumb forces on digital lattices. Recent work recognizes the application of classical computation beside quantum computing technologies. Quantum Galaxies Corporation is a startup from Texas Tech University for classical computing based on patents for quantum computing."
  • Item
    Observability of Low-Luminosity AGN in the Early Universe with JWST
    (2023-09-28) Jeon, Junehyoung; Liu, Boyuan; Bromm, Volker; Finkelstein, Steven L.
    Active galactic nuclei (AGN) in the early Universe are thought to be prominent sources of energy and ionizing photons that affected the growth of their host galaxy and their environment. However, it is still unclear how the supermassive black holes (SMBHs) that fuel these AGN grew to the observed high masses already at high redshifts. Observations of high-redshift SMBH progenitors or lower-luminosity AGN will thus help characterize the evolution of SMBHs and their impact on the surroundings. With the launch of the JWST, fainter objects at high redshifts can now be detected, including lower-luminosity AGN. We assess the observability of such low luminosity AGN, using the cosmological simulation code gizmo to provide a realistic environment for black hole growth in the early Universe. Soon after the first stars are born in the simulation run, we insert stellar-remnant black hole seeds of various initial masses, between 300 and 104 M⊙, at the center of a dark matter halo and follow their growth until z ∼ 6. Such stellar black hole seeds placed in a typical high-z environment do not significantly accrete and grow to reach masses that can be observed with the JWST under conditions of standard Bondi-Hoyle accretion, as energy input from stellar feedback and chaotic dynamics prevent efficient gas accretion onto the black holes. To be observed with the JWST, rarer but still physically feasible growth regimes, involving Eddington or super-Eddington accretion, would be required. Alternatively, AGN observability may be boosted under even rarer conditions of extreme gravitational lensing.
  • Item
    Machine Learning Hub for Tapis
    (2023-09-28) Indrakusuma, Dhanny; Freeman, Nathan; Stubbs, Joe
    "Machine learning is indispensable for extracting insights from intricate datasets, expediting data analysis, and enabling cross-disciplinary decision-making. However, the complexity of machine learning models can hinder non-technical users, necessitating for user-friendly tools. Within the Cloud and Interactive Computing group (CIC) at the Texas Advanced Computing Center (TACC), we are actively developing a Machine Learning Hub (ML Hub) API for the Tapis Framework. Comprising accessible microservices, each with an independent REST API implemented in Python's Flask and codified with OpenAPI v3 definitions, our research aims to enhance the experiences of developers, scientists, and researchers. The integration of Hugging Face's API into ML Hub provides open-source pre-trained models for state-of-the-art AI capabilities. Currently, ML Hub's Models Overview and Models Download functions offer a gateway for non-technical users to explore and download machine learning models, authenticated using a JSON Web Token (JWT) from the Tapis Authenticator API. Future developments encompass implementing the Inference Client and Training Engine, seamlessly integrating with the Tapis UI in React and Typescript. Key features of ML Hub: 1. Models Overview: A portal showcasing top Hugging Face models with filtering options. 2. Models Download: Users can obtain specific models, with options to either download a binary file of the model or a zip file containing the model's repository, cached in a version-aware manner. 3. Inference Client: Facilitating server initiation for machine learning model inference on TACC's HPC cluster, enabling rapid prototyping. 4. Training Engine: Enabling users to fine-tune models and showcase them on TACC's HPC cluster, removing technical complexities. This research contributes to the broader discourse on democratizing machine learning's potential, by providing user- friendly access to state-of-the-art models and addressing non-technical users' challenges. We hope that this project will foster innovative collaboration and user engagement, paving the way for an inclusive and impactful future in machine learning research."