SEARCH · Engineering Papers
Results for “COMPUTER NETWORKS”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Total absorption spectroscopy of two isomers in 70 Cu influencing nucleosynthesis signatures
Isomers have long been known to be important for astrophysical nucleosynthesis processes, yet they are often neglected in network calculations due to computational limitations or lack of data. "Astromers" are astrophysically metastable nuclear states that can greatly impact nucleosynthesis pathways. In this work we show that astromers further impact the time-dependent electromagnetic signal during heavy element nucleosynthesis. In an experiment performed at the National Superconducting Cyclotron Laboratory, three 𝛽-decaying states of 70 Cu (6 − ground state, and two isomeric states: 3 − and 1 + ) were produced. 𝛽-feeding values were extracted from experimental spectra and compared to shell-model and QRPA+PVC calculations. Here, average 𝛾-ray energies from the 𝛽-decay events were incorporated into simulations of heavy element nucleosynthesis and were found to exhibit different energy release profiles over time, which may impact, in aggregate, time-dependent observable signals.
Scalable Static and Dynamic Community Detection Using Grappolo
Graph clustering, popularly known as community detection, is a fundamental kernel for several applications of relevance to the Defense Advanced Research Projects Agency’s (DARPA) Hierarchical Identify Verify Exploit (HIVE) Pro- gram. Clusters or communities represent natural divisions within a network that are densely connected within a cluster and sparsely connected to the rest of the network. The need to compute clustering on large scale data necessitates the development of efficient algorithms that can exploit modern architectures that are fundamentally parallel in nature. How- ever, due to their irregular and inherently sequential nature, many of the current algorithms for community detection are challenging to parallelize. In response to the HIVE Graph Challenge, we present several parallelization heuristics for fast community detection using the Louvain method as the serial template. We implement all the heuristics in a software library called Grappolo. Using the inputs from the HIVE Challenge, we demonstrate superior performance and high quality solutions based on four parallelization heuristics. We use Grappolo on static graphs as the first step towards community detection on streaming graphs.
Understanding and Estimating Error Propagation in Neural Networks for Scientific Data Analysis
Neural networks are increasingly integrated into scientific discovery, where input data reduction and model quantization play a key role in accelerating inference. However, understanding and mitigating the impact of these techniques on output error is critical for ensuring reliable results, particularly in tasks demanding high numerical precision. This paper introduces a comprehensive framework for optimizing neural network inference in scientific computing by combining data reduction and weight quantization while maintaining error-controlled outcomes. We develop theoretical analyses to bound error propagation under these reductions and propose a framework that balances computational performance with error constraints. Evaluation on real-world learning-based combustion simulations and satellite image classification demonstrates that our derived error bounds accurately predict observed errors while enabling significant computational speedup under our framework. This work highlights the potential for further leveraging advancements in modern lossy compression algorithms and hardware accelerators that support lower-precision formats.
Communication-Avoiding and Memory-Constrained Sparse Matrix-Matrix Multiplication at Extreme Scale
Sparse matrix-matrix multiplication (SpGEMM) is a widely used kernel in various graph, scientific computing and machine learning algorithms. In this paper, we consider SpGEMMs performed on hundreds of thousands of processors generating trillions of nonzeros in the output matrix. Distributed SpGEMM at this extreme scale faces two key challenges: (1) high communication cost and (2) inadequate memory to generate the output. Furthermore, we address these challenges with an integrated communication-avoiding and memory-constrained SpGEMM algorithm that scales to 262,144 cores (more than 1 million hardware threads) and can multiply sparse matrices of any size as long as inputs and a fraction of output fit in the aggregated memory. As we go from 16,384 cores to 262,144 cores on a Cray XC40 supercomputer, the new SpGEMM algorithm runs 10x faster when multiplying large-scale protein-similarity matrices.
Ionizing radiation effects in SONOS-based neuromorphic inference accelerators
Here, we evaluate the sensitivity of neuromorphic inference accelerators based on Silicon-Oxide-Nitride-Oxide-Silicon (SONOS) charge trap memory arrays to total ionizing dose (TID) effects. Data retention statistics were collected for 16 Mbit of 40 nm SONOS digital memory exposed to ionizing radiation from a Co-60 source, showing good retention of the bits up to the maximum dose of 500 krad(Si). Using this data, we formulate a rate-equation-based model for the TID response of trapped charge carriers in the ONO stack, and predict the effect of TID on intermediate device states between ‘program’ and ‘erase’. This model is then used to simulate arrays of low-power, analog SONOS devices that store 8-bit neural network weights and support in situ matrix-vector multiplication. We evaluate the accuracy of the irradiated SONOS-based inference accelerator on two image recognition tasks – CIFAR-10 and the challenging ImageNet dataset – using state-of-the-art convolutional neural networks, such as ResNet-50. We find that across the datasets and neural networks evaluated, the accelerator tolerates a maximum TID between 10 krad(Si) and 100 krad(Si), with deeper networks being more susceptible to accuracy losses due to TID.
Distributed Finite-Time Termination for Consensus Algorithm in Switching Topologies
Here, in this article, we present a finite-time stopping criterion for consensus algorithms in networks with dynamic communication topology. Prior state of the art has established convergence to the consensus value; however, the asymptotic convergence of these algorithms poses a challenge in practical settings where the response from agents is required in finite time. To this end, we propose a maximum-minimum protocol that propagates the global maximum and minimum values of agent states (while running the consensus algorithm) in the network. This article focuses on establishing that the global maximum and minimum values are strictly monotonic even for a dynamic topology, and they can be used to distributively ascertain the closeness to convergence in finite time. We rigorously show that each node can have access to the global maximum and minimum by running the proposed maximum-minimum protocol to realize a finite-time stopping criterion for the otherwise asymptotic consensus algorithm. The practical utility of the algorithm is illustrated through experiments where each agent is instantiated by a NodeJS socket.io server.
Vibrational spectroscopy analysis of silica and silicate glass networks
Vibrational spectroscopy has been widely used to investigate various structural aspects of the glass network, and there are a plethora of papers reporting subtle but consistent changes in infrared and Raman spectral features of glass upon alterations of glass compositions, thermal histories, mechanical stresses, or surface treatments. However, interpretations of such spectral features are still obscured due to the lack of well-established physical principles accurately describing vibrational modes of the non-crystalline glass network. Due to the non-equilibrium nature of the glass network, three-dimensionally connected without any long-range orders, vibrational spectral features of glass cannot be interpreted using the analogy with those of isolated molecular moieties or crystalline counterparts. Here, this feature article explains why such comparisons are outdated and describes the recent advances made from theoretical calculations of vibrational spectral features of amorphous networks or comparisons of computational results with experimental data. For the interpretation of vibrational spectral features of silica and silicate glasses, the following empirical relationships are suggested: (i) the intensity-weighted peak position of the Si-O-Si stretch mode negatively correlates with the weighted average of the Si-O bond length distribution, and (ii) the broad band of the Si-O-Si bending mode negatively correlates with the Si-O-Si bond angle distribution. Selected examples of vibrational spectroscopic imaging of surface defects are discussed to deliberate the implication of these findings in the structure-property relationship of silica and silicate glass materials. Unanswered questions and continuing research challenges are identified.
Quantum entanglement at ambient conditions in a macroscopic solid-state spin ensemble
Entanglement is a key resource for quantum computers, quantum-communication networks, and high-precision sensors. Macroscopic spin ensembles have been historically important in the development of quantum algorithms for these prospective technologies and remain strong candidates for implementing them today. This strength derives from their long-lived quantum coherence, strong signal, and ability to couple collectively to external degrees of freedom. Nonetheless, preparing ensembles of genuinely entangled spin states has required high magnetic fields and cryogenic temperatures or photochemical reactions. We demonstrate that entanglement can be realized in solid-state spin ensembles at ambient conditions. We use hybrid registers comprising of electron-nuclear spin pairs that are localized at color-center defects in a commercial SiC wafer. We optically initialize 10 3 identical registers in a 40-μm 3 volume (with 0.95$_{-0.07}^{+0.05}$ fidelity) and deterministically prepare them into the maximally entangled Bell states (with 0.88 ± 0.07 fidelity). To verify entanglement, we develop a register-specific quantum-state tomography protocol. The entanglement of a macroscopic solid-state spin ensemble at ambient conditions represents an important step toward practical quantum technology.
Gleipnir: toward practical error analysis for Quantum programs
Practical error analysis is essential for the design, optimization, and evaluation of Noisy Intermediate-Scale Quantum(NISQ) computing. However, bounding errors in quantum programs is a grand challenge, because the effects of quantum errors depend on exponentially large quantum states. In this work, we present Gleipnir, a novel methodology toward practically computing verified error bounds in quantum programs. Gleipnir introduces the (ρ,δ)-diamond norm, an error metric constrained by a quantum predicate consisting of the approximate state ρ and its distance δ to the ideal state ρ. This predicate (ρ,δ) can be computed adaptively using tensor networks based on the Matrix Product States. Gleipnir features a lightweight logic for reasoning about error bounds in noisy quantum programs, based on the (ρ,δ)-diamond norm metric. Furthermore, our experimental results show that Gleipnir is able to efficiently generate tight error bounds for real-world quantum programs with 10 to 100 qubits, and can be used to evaluate the error mitigation performance of quantum compiler transformations.
cuTS: Scaling Subgraph Isomorphism on Distributed Multi-GPUSystems Using Trie Based Data Structure
Subgraph isomorphism is a pattern-matching algorithm widely used in many domains such as chem-informatics, bioinformatics, databases, and social network analysis. It is computationally expensive and is a proven NP-hard problem. The massive parallelism offered by the GPU hardware is well suited for solving the subgraph isomorphism. However, current GPU implementations are far from the achievable performance. Moreover, the enormous memory requirement of current approaches limits the problem size that can be handled. This work analyzes the fundamental challenges associated with processing the subgraph isomorphism on GPUs and develops an efficient GPU hardware-aware implementation. We also develop a new GPU-friendly trie-based data structure to drastically reduce the intermediate storage space requirement. Hence, our approach runs larger benchmarks than the competitors. We also develop the first distributed sub-graph isomorphism algorithm for GPUs. Our experimental evaluation section demonstrates the efficacy of our approach by comparing the execution time and number of cases that we can handle against the state-of-the-art GPU implementations.
Master State Threat Identifier (masti)
The MSE cyber sensor offers a novel approach for scalable network monitoring using heterogeneous computing. Packet data is collected, analyzed, and stored by a stand alone cyber sensor. Machine learning algorithms are run to discover patterns in the communication without the need for deep packet inspection. Tuning of the detection algorithms happens automatically, without need for an engineer to manually create detection rules.
Targeted Energy Management System Toolset
TEMST integrates advanced sensor systems with advanced computational intelligence, e.g. neural networks and fuzzy arithmetic, to enable effective energy management while minimizing negative human comfort consequences and promoting improved occupant satisfaction. Preliminary work suggests that energy cost reductions above ten percent are easily achievable. In addition to optimizing existing Energy Management Systems (EMS), the proposed toolset will be both resilient and secure in order to proactively maintain a safe level of operational normalcy in response to anomalies and malicious actions.
Design of an ultra-low mode volume piezo-optomechanical quantum transducer
Coherent transduction of quantum states from the microwave to the optical domain can play a key role in quantum networking and distributed quantum computing. We present the design of a piezo-optomechanical device formed in a hybrid lithium niobate on silicon platform, that is suitable for microwave-to-optical quantum transduction. Our design is based on acoustic hybridization of an ultra-low mode volume piezoacoustic cavity with an optomechanical crystal cavity. The strong piezoelectric nature of lithium niobate allows us to mediate transduction via an acoustic mode which only minimally interacts with the lithium niobate, and is predominantly silicon-like, with very low electrical and acoustic loss. We estimate that this transducer can realize an intrinsic conversion efficiency of up to 35% with <0.5 added noise quanta when resonantly coupled to a superconducting transmon qubit and operated in pulsed mode at 10 kHz repetition rate. The performance improvement gained in such hybrid lithium niobate-silicon transducers make them suitable for heralded entanglement of qubits between superconducting quantum processors connected by optical fiber links.
2D material platform for overcoming the amplitude–phase tradeoff in ring resonators
Compact and high-speed electro-optic phase modulators play a vital role in various large-scale applications including optical computing, quantum and neural networks, and optical communication links. Conventional electro-refractive phase modulators such as silicon (Si), III-V and graphene on Si suffer from a fundamental tradeoff between device length and optical loss that limits their scaling capabilities. High-finesse ring resonators have been traditionally used as compact intensity modulators, but their use for phase modulation has been limited due to the high insertion loss associated with the phase shift. Here, we show that high-finesse resonators can achieve a strong phase shift with low insertion loss by simultaneous modulation of the real and imaginary parts of the refractive index, to the same extent, i.e., Δ n Δ k ∼1. To implement this strategy, we demonstrate an active hybrid platform that combines a low-loss SiN ring resonator with 2D materials such as graphene and transition metal dichalcogenide [tungsten disulphide (WSe 2 )], which induces a strong change in the imaginary and real parts of the index. Our platform consisting of a 25 µm long Gr-Al 2 O 3 -WSe 2 capacitor embedded on a SiN ring of 50 µm radius (∼8% ring coverage) achieves a continuous phase shift of (0.46±0.05) π radians with an insertion loss (IL) of 3.18±0.20 dB and a transmission modulation (Δ T Ring ) of 1.72±0.15dB at a probe wavelength ( λ p ) of 1646.18 nm. We find that our Gr-Al 2 O 3 -WSe 2 capacitor exhibits a phase modulation efficiency ( V π 2 ⋅ L ) of 0.530±0.016V⋅cm and can support an electro-optic bandwidth of 14.9±0.1GHz. We further show that our platform can achieve a phase shift of π radians with an IL of 5 dB and a minimum Δ T of 0.046 dB. We demonstrate the broadband nature of the binary phase response, by measuring a phase shift of (1.00±0.10) π radians, with an IL of 5.20±0.31dB and a minimal Δ T Ring of 0.015±0.006dB for resonances spanning from 1564 to 1650 nm. This SiN–2D hybrid platform provides the design for compact and high-speed reconfigurable circuits with graphene and transition metal dichalcogenide (TMD) monolayers that can enable large-scale photonic systems.
Diffusion-Model-Assisted Supervised Learning of Generative Models for Density Estimation
Here, we present a supervised learning framework of training generative models for density estimation. Generative models, including generative adversarial networks (GANs), normalizing flows, and variational auto-encoders (VAEs), are usually considered as unsupervised learning models, because labeled data are usually unavailable for training. Despite the success of the generative models, there are several issues with the unsupervised training, e.g., requirement of reversible architectures, vanishing gradients, and training instability. To enable supervised learning in generative models, we utilize the score-based diffusion model to generate labeled data. Unlike existing diffusion models that train neural networks to learn the score function, we develop a training-free score estimation method. This approach uses mini-batch-based Monte Carlo estimators to directly approximate the score function at any spatial-temporal location in solving an ordinary differential equation (ODE), corresponding to the reverse-time stochastic differential equation (SDE). This approach can offer both high accuracy and substantial time savings in neural network training. Once the labeled data are generated, we can train a simple, fully connected neural network to learn the generative model in the supervised manner. Compared with existing normalizing flow models, our method does not require the use of reversible neural networks and avoids the computation of the Jacobian matrix. Compared with existing diffusion models, our method does not need to solve the reverse-time SDE to generate new samples. As a result, the sampling efficiency is significantly improved. We demonstrate the performance of our method by applying it to a set of 2D datasets as well as real data from the University of California Irvine (UCI) repository.
Fiats: Functional inference and training for surrogates
Fiats provides a platform for research on the training and deployment of neural-network surrogate models for computational science. Fiats also supports exploring, advancing, and combining functional, object-oriented, and parallel programming patterns in Fortran 2023. As such, the Fiats name has dual expansions: “Functional Inference And Training for Surrogates” or “Fortran Inference And Training for Science.” Fiats inference and training procedures are pure and therefore satisfy a language constraint imposed on procedure invocations inside Fortran’s parallel loop construct: do concurrent. Furthermore, the Fiats training procedures are built around a do concurrent parallel reduction. Several compilers can automatically parallelize do concurrent on Central Processing Units (CPUs) or Graphics Processing Units (GPUs). Fiats thus aims to achieve performance portability through standard language mechanisms.