arXiv subjects

Meng Zhang

Publications and source records attributed to Meng Zhang.

At least 127 records · Page 7Linked to original sources

arXiv2024

A Method for the Runtime Validation of AI-based Environment Perception in Automated Driving System

Environment perception is a fundamental part of the dynamic driving task executed by Autonomous Driving Systems (ADS). Artificial Intelligence (AI)-based approaches have prevailed over classical techniques for realizing the environment perception. Current safety-relevant standards for automotive systems, International Organization for Standardization (ISO) 26262 and ISO 21448, assume the existence of comprehensive requirements specifications. These specifications serve as the basis on which the functionality of an automotive system can be rigorously tested and checked for compliance with safety regulations. However, AI-based perception systems do not have complete requirements specification. Instead, large datasets are used to train AI-based perception systems. This paper presents a function monitor for the functional runtime monitoring of a two-folded AI-based environment perception for ADS, based respectively on camera and LiDAR sensors. To evaluate the applicability of the function monitor, we conduct a qualitative scenario-based evaluation in a controlled laboratory environment using a model car. The evaluation results then are discussed to provide insights into the monitor's performance and its suitability for real-world applications.

cs.RO↗

arXiv2024

Neural Garment Dynamic Super-Resolution

Achieving efficient, high-fidelity, high-resolution garment simulation is challenging due to its computational demands. Conversely, low-resolution garment simulation is more accessible and ideal for low-budget devices like smartphones. In this paper, we introduce a lightweight, learning-based method for garment dynamic super-resolution, designed to efficiently enhance high-resolution, high-frequency details in low-resolution garment simulations. Starting with low-resolution garment simulation and underlying body motion, we utilize a mesh-graph-net to compute super-resolution features based on coarse garment dynamics and garment-body interactions. These features are then used by a hyper-net to construct an implicit function of detailed wrinkle residuals for each coarse mesh triangle. Considering the influence of coarse garment shapes on detailed wrinkle performance, we correct the coarse garment shape and predict detailed wrinkle residuals using these implicit functions. Finally, we generate detailed high-resolution garment geometry by applying the detailed wrinkle residuals to the corrected coarse garment. Our method enables roll-out prediction by iteratively using its predictions as input for subsequent frames, producing fine-grained wrinkle details to enhance the low-resolution simulation. Despite training on a small dataset, our network robustly generalizes to different body shapes, motions, and garment types not present in the training data. We demonstrate significant improvements over state-of-the-art alternatives, particularly in enhancing the quality of high-frequency, fine-grained wrinkle details.

cs.CV↗

arXiv2024

SimPhony: A Device-Circuit-Architecture Cross-Layer Modeling and Simulation Framework for Heterogeneous Electronic-Photonic AI System

Electronic-photonic integrated circuits (EPICs) offer transformative potential for next-generation high-performance AI but require interdisciplinary advances across devices, circuits, architecture, and design automation. The complexity of hybrid systems makes it challenging even for domain experts to understand distinct behaviors and interactions across design stack. The lack of a flexible, accurate, fast, and easy-to-use EPIC AI system simulation framework significantly limits the exploration of hardware innovations and system evaluations on common benchmarks. To address this gap, we propose SimPhony, a cross-layer modeling and simulation framework for heterogeneous electronic-photonic AI systems. SimPhony offers a platform that enables (1) generic, extensible hardware topology representation that supports heterogeneous multi-core architectures with diverse photonic tensor core designs; (2) optics-specific dataflow modeling with unique multi-dimensional parallelism and reuse beyond spatial/temporal dimensions; (3) data-aware energy modeling with realistic device responses, layout-aware area estimation, link budget analysis, and bandwidth-adaptive memory modeling; and (4) seamless integration with model training framework for hardware/software co-simulation. By providing a unified, versatile, and high-fidelity simulation platform, SimPhony enables researchers to innovate and evaluate EPIC AI hardware across multiple domains, facilitating the next leap in emerging AI hardware. We open-source our codes at https://github.com/ScopeX-ASU/SimPhony

physics.optics↗

arXiv2024

BOSON$^{-1}$: Understanding and Enabling Physically-Robust Photonic Inverse Design with Adaptive Variation-Aware Subspace Optimization

Nanophotonic device design aims to optimize photonic structures to meet specific requirements across various applications. Inverse design has unlocked non-intuitive, high-dimensional design spaces, enabling the discovery of high-performance devices beyond heuristic or analytic methods. The adjoint method, which calculates gradients for all variables using just two simulations, enables efficient navigation of this complex space. However, many inverse-designed structures, while numerically plausible, are difficult to fabricate and sensitive to variations, limiting their practical use. The discrete nature with numerous local-optimal structures also pose significant optimization challenges, often causing gradient-based methods to converge on suboptimal designs. In this work, we formulate inverse design as a fabrication-restricted, discrete, probabilistic optimization problem and introduce BOSON-1, an end-to-end, variation-aware subspace optimization framework to address the challenges of manufacturability, robustness, and optimizability. To overcome optimization difficulty, we propose dense target-enhanced gradient flows to mitigate misleading local optima and introduce a conditional subspace optimization strategy to create high-dimensional tunnels to escape local optima. Furthermore, we significantly reduce the runtime associated with optimizing across exponential variation samples through an adaptive sampling-based robust optimization, ensuring both efficiency and variation robustness. On three representative photonic device benchmarks, our proposed inverse design methodology BOSON^-1 delivers fabricable structures and achieves the best convergence and performance under realistic variations, outperforming prior arts with 74.3% post-fabrication performance. We open-source our codes at https://github.com/ScopeX-ASU/BOSON.

physics.optics↗

arXiv2024

Absence of altermagnetic spin splitting character in rutile oxide RuO$_2$

Rutile RuO$_2$ has been posited as a potential $d$-wave altermagnetism candidate, with a predicted significant spin splitting up to 1.4 eV. Despite accumulating theoretical predictions and transport measurements, direct spectroscopic observation of spin splitting has remained elusive. Here, we employ spin- and angle-resolved photoemission spectroscopy to investigate the band structures and spin polarization of thin-film and single-crystal RuO$_2$. Contrary to expectations of altermagnetism, our analysis indicates that RuO$_2$'s electronic structure aligns with those predicted under non-magnetic conditions, exhibiting no evidence of the hypothesized spin splitting. Additionally, we observe significant in-plane spin polarization of the low-lying bulk bands, which is antisymmetric about the high-symmetry plane and contrary to the $d$-wave spin texture due to time-reversal symmetry breaking in altermagnetism. These findings definitively challenge the altermagnetic order previously proposed for rutile RuO$_2$, prompting a reevaluation of its magnetic properties.

cond-mat.mtrl-sci↗

arXiv2024

ADEPT-Z: Zero-Shot Automated Circuit Topology Search for Pareto-Optimal Photonic Tensor Cores

Photonic tensor cores (PTCs) are essential building blocks for optical artificial intelligence (AI) accelerators based on programmable photonic integrated circuits. Most PTC designs today are manually constructed, with low design efficiency and unsatisfying solution quality. This makes it challenging to meet various hardware specifications and keep up with rapidly evolving AI applications. Prior work has explored gradient-based methods to learn a good PTC structure differentiably. However, it suffers from slow training speed and optimization difficulty when handling multiple non-differentiable objectives and constraints. Therefore, in this work, we propose a more flexible and efficient zero-shot multi-objective evolutionary topology search framework ADEPT-Z that explores Pareto-optimal PTC designs with advanced devices in a larger search space. Multiple objectives can be co-optimized while honoring complicated hardware constraints. With only <3 hours of search, we can obtain tens of diverse Pareto-optimal solutions, 100x faster than the prior gradient-based method, outperforming prior manual designs with 2x higher accuracy weighted area-energy efficiency. The code of ADEPT-Z is available at https://github.com/ScopeX-ASU/ADEPT-Z.

cs.ET↗

arXiv2024

Scalable tensor network algorithm for thermal quantum many-body systems in two dimension

Simulating strongly-correlated quantum many-body systems at finite temperatures is a significant challenge in computational physics. In this work, we present a scalable finite-temperature tensor network algorithm for two-dimensional quantum many-body systems. We employ the (fermionic) projected entangled pair state (PEPS) to represent the vectorization of the quantum thermal state and utilize a stochastic reconfiguration method to cool down the quantum states from infinite temperature. We validate our method by benchmarking it against the 2D antiferromagnetic Heisenberg model, the $J_1$-$J_2$ model, and the Fermi-Hubbard model, comparing physical properties such as internal energy, specific heat, and magnetic susceptibility with results obtained from stochastic series expansion (SSE), exact diagonalization, and determinant quantum Monte Carlo (DQMC).

cond-mat.str-el↗

arXiv2024

Ground-roll Separation From Land Seismic Records Based on Convolutional Neural Network

Ground-roll wave is a common coherent noise in land field seismic data. This Rayleigh-type surface wave usually has low frequency, low apparent velocity, and high amplitude, therefore obscures the reflection events of seismic shot gathers. Commonly used techniques focus on the differences of ground-roll and reflection in transformed domain such as $f-k$ domain, wavelet domain, or curvelet domain. These approaches use a series of fixed atoms or bases to transform the data in time-space domain into transformed domain to separate different waveforms, thus tend to suffer from the complexity for a delicate design of the parameters of the transform domain filter. To deal with these problems, a novel way is proposed to separate ground-roll from reflections using convolutional neural network (CNN) model based method to learn to extract the features of ground-roll and reflections automatically based on training data. In the proposed method, low-pass filtered seismic data which is contaminated by ground-roll wave is used as input of CNN, and then outputs both ground-roll component and low-frequency part of reflection component simultaneously. Discriminative loss is applied together with similarity loss in the training process to enhance the similarity to their train labels as well as the difference between the two outputs. Experiments are conducted on both synthetic and real data, showing that CNN based method can separate ground roll from reflections effectively, and has generalization ability to a certain extent.

cs.CV↗

arXiv2024

Efficient Active Flow Control Strategy for Confined Square Cylinder Wake Using Deep Learning-Based Surrogate Model and Reinforcement Learning

This study presents a deep learning model-based reinforcement learning (DL-MBRL) approach for active control of two-dimensional (2D) wake flow past a square cylinder using antiphase jets. The DL-MBRL framework alternates between interacting with a deep learning surrogate model (DL-SM) and computational fluid dynamics (CFD) simulations to suppress wake vortex shedding, significantly reducing computational costs. The DL-SM, which combines a Transformer and a multiscale enhanced super-resolution generative adversarial network (MS-ESRGAN), effectively models complex flow dynamics, efficiently emulating the CFD environment. Trained on 2D direct numerical simulation (DNS) data, the Transformer and MS-ESRGAN demonstrated excellent agreement with DNS results, validating the DL-SM's accuracy. Error analysis suggests replacing the DL-SM with CFD every five interactions to maintain reliability. While DL-MBRL showed less robust convergence than model-free reinforcement learning (MFRL) during training, it reduced training time by 49.2%, from 41.87 hours to 20.62 hours. Both MFRL and DL-MBRL achieved a 98% reduction in shedding energy and a 95% reduction in the standard deviation of the lift coefficient (C_L). However, MFRL exhibited a nonzero mean lift coefficient due to insufficient exploration, whereas DL-MBRL improved exploration by leveraging the randomness of the DL-SM, resolving the nonzero mean C_L issue. This study demonstrates that DL-MBRL is not only comparably effective but also superior to MFRL in flow stabilization, with significantly reduced training time, highlighting the potential of combining deep reinforcement learning with DL-SM for enhanced active flow control.

physics.flu-dyn↗

arXiv2024

Flow Reconstruction Using Spatially Restricted Domains Based on Enhanced Super-Resolution Generative Adversarial Networks

This study aims to reconstruct the complete flow field from spatially restricted domain data by utilizing an Enhanced Super-Resolution Generative Adversarial Network (ESRGAN) model. The difficulty in flow field reconstruction lies in accurately capturing and reconstructing large amounts of data under nonlinear, multi-scale, and complex flow while ensuring physical consistency and high computational efficiency. The ESRGAN model has a strong information mapping capability, capturing fluctuating features from local flow fields of varying geometries and sizes. The model effectiveness in reconstructing the whole domain flow field is validated by comparing instantaneous velocity fields, flow statistical properties, and probability density distributions. Using laminar bluff body flow from Direct Numerical Simulation (DNS) as a priori case, the model successfully reconstructs the complete flow field from three non-overlapping limited regions, with flow statistical properties perfectly matching the original data. Validation of the power spectrum density (PSD) for the reconstruction results also proves that the model could conform to the temporal behavior of the real complete flow field. Additionally, tests using DNS turbulent channel flow with a friction Reynolds number ($Re_τ= 180$) demonstrate the model ability to reconstruct turbulent fields, though the quality of results depends on the number of flow features in the local regions. Finally, the model is applied to reconstruct turbulence flow fields from Particle Image Velocimetry (PIV) experimental measurements, using limited data from the near-wake region to reconstruct a larger field of view. The turbulence statistics closely match the experimental data, indicating that the model can serve as a reliable data-driven method to overcome PIV field-of-view limitations while saving computational costs.

physics.flu-dyn↗

arXiv2024

Self-Supervised Learning for Effective Denoising of Flow Fields

In this study, we proposed an efficient approach based on a deep learning (DL) denoising autoencoder (DAE) model for denoising noisy flow fields. The DAE operates on a self-learning principle and does not require clean data as training labels. Furthermore, investigations into the denoising mechanism of the DAE revealed that its bottleneck structure with a compact latent space enhances denoising efficacy. Meanwhile, we also developed a deep multiscale DAE for denoising turbulent flow fields. Furthermore, we used conventional noise filters to denoise the flow fields and performed a comparative analysis with the results from the DL method. The effectiveness of the proposed DL models was evaluated using direct numerical simulation data of laminar flow around a square cylinder and turbulent channel flow data at various Reynolds numbers. For every case, synthetic noise was augmented in the data. A separate experiment used particle-image velocimetry data of laminar flow around a square cylinder containing real noise to test DAE denoising performance. Instantaneous contours and flow statistical results were used to verify the alignment between the denoised data and ground truth. The findings confirmed that the proposed method could effectively denoise noisy flow data, including turbulent flow scenarios. Furthermore, the proposed method exhibited excellent generalization, efficiently denoising noise with various types and intensities.

physics.flu-dyn↗

arXiv2024

Garment Animation NeRF with Color Editing

Generating high-fidelity garment animations through traditional workflows, from modeling to rendering, is both tedious and expensive. These workflows often require repetitive steps in response to updates in character motion, rendering viewpoint changes, or appearance edits. Although recent neural rendering offers an efficient solution for computationally intensive processes, it struggles with rendering complex garment animations containing fine wrinkle details and realistic garment-and-body occlusions, while maintaining structural consistency across frames and dense view rendering. In this paper, we propose a novel approach to directly synthesize garment animations from body motion sequences without the need for an explicit garment proxy. Our approach infers garment dynamic features from body motion, providing a preliminary overview of garment structure. Simultaneously, we capture detailed features from synthesized reference images of the garment's front and back, generated by a pre-trained image model. These features are then used to construct a neural radiance field that renders the garment animation video. Additionally, our technique enables garment recoloring by decomposing its visual elements. We demonstrate the generalizability of our method across unseen body motions and camera views, ensuring detailed structural consistency. Furthermore, we showcase its applicability to color editing on both real and synthetic garment data. Compared to existing neural rendering techniques, our method exhibits qualitative and quantitative improvements in garment dynamics and wrinkle detail modeling. Code is available at \url{https://github.com/wrk226/GarmentAnimationNeRF}.

cs.CV↗

arXiv2024

Operando probing of nanocracking in CuO-derived Cu during CO$_2$ electroreduction

Identifying and controlling active sites in electrocatalysis remains a grand challenge due to restructuring of catalysts in the complex chemical environments during operation. Inactive precatalysts can transform into active catalysts under reaction conditions, such as oxide-derived Cu (OD-Cu) for CO$_2$ electroreduction displaying improved production of multicarbon (C$_{2+}$) chemicals. Revealing the mechanism of active site origin in OD-Cu catalysts requires in situ/operando characterizations of structure, morphology, and valence state evolution with high spatial and temporal resolution. Applying newly developed electrochemical liquid cell transmission electron microscopy combined with X-ray absorption spectroscopy, our multimodal operando techniques unveil the formation pathways of OD-Cu active sites from CuO bicrystal nanowire precatalysts. Rapid reduction of CuO directly to Cu within 60 seconds generates a nanocrack network throughout the nanowire, via formation of "boundary nanocracks" along the twin boundary and "transverse nanocracks" propagating from the surface to the center of the nanowire. The nanocrack network further reconstructs, leading to a highly porous structure rich in Cu nanograins, with a boosted specific surface area and density of active sites for C$_{2+}$ products. These findings suggest a means to optimize active OD-Cu nanostructures through nanocracking by tailoring grain boundaries in CuO precatalysts. More generally, our advanced operando approach opens new opportunities for mechanistic insights to enable improved control of catalyst structure and performance.

cond-mat.mtrl-sci↗

arXiv2024

Beyond Boundaries: efficient Projected Entangled Pair States methods for periodic quantum systems

Projected Entangled Pair States (PEPS) are recognized as a potent tool for exploring two-dimensional quantum many-body systems. However, a significant challenge emerges when applying conventional PEPS methodologies to systems with periodic boundary conditions (PBC), attributed to the prohibitive computational scaling with the bond dimension. This has notably restricted the study of systems with complex boundary conditions. To address this challenge, we have developed a strategy that involves the superposition of PEPS with open boundary conditions (OBC) to treat systems with PBC. This approach significantly reduces the computational complexity of such systems while maintaining their translational invariance and the PBC. We benchmark this method against the Heisenberg model and the $J_1$-$J_2$ model, demonstrating its capability to yield highly accurate results at low computational costs, even for large system sizes. The techniques are adaptable to other boundary conditions, including cylindrical and twisted boundary conditions, and therefore significantly expands the application scope of the PEPS approach, shining new light on numerous applications.

cond-mat.str-el↗

arXiv2024

TorchGT: A Holistic System for Large-scale Graph Transformer Training

Graph Transformer is a new architecture that surpasses GNNs in graph learning. While there emerge inspiring algorithm advancements, their practical adoption is still limited, particularly on real-world graphs involving up to millions of nodes. We observe existing graph transformers fail on large-scale graphs mainly due to heavy computation, limited scalability and inferior model quality. Motivated by these observations, we propose TorchGT, the first efficient, scalable, and accurate graph transformer training system. TorchGT optimizes training at different levels. At algorithm level, by harnessing the graph sparsity, TorchGT introduces a Dual-interleaved Attention which is computation-efficient and accuracy-maintained. At runtime level, TorchGT scales training across workers with a communication-light Cluster-aware Graph Parallelism. At kernel level, an Elastic Computation Reformation further optimizes the computation by reducing memory access latency in a dynamic way. Extensive experiments demonstrate that TorchGT boosts training by up to 62.7x and supports graph sequence lengths of up to 1M.

cs.DC↗

arXiv2024

SCATTER: Algorithm-Circuit Co-Sparse Photonic Accelerator with Thermal-Tolerant, Power-Efficient In-situ Light Redistribution

Photonic computing has emerged as a promising solution for accelerating computation-intensive artificial intelligence (AI) workloads. However, limited reconfigurability, high electrical-optical conversion cost, and thermal sensitivity limit the deployment of current optical analog computing engines to support power-restricted, performance-sensitive AI workloads at scale. Sparsity provides a great opportunity for hardware-efficient AI accelerators. However, current dense photonic accelerators fail to fully exploit the power-saving potential of algorithmic sparsity. It requires sparsity-aware hardware specialization with a fundamental re-design of photonic tensor core topology and cross-layer device-circuit-architecture-algorithm co-optimization aware of hardware non-ideality and power bottleneck. To trim down the redundant power consumption while maximizing robustness to thermal variations, we propose SCATTER, a novel algorithm-circuit co-sparse photonic accelerator featuring dynamically reconfigurable signal path via thermal-tolerant, power-efficient in-situ light redistribution and power gating. A power-optimized, crosstalk-aware dynamic sparse training framework is introduced to explore row-column structured sparsity and ensure marginal accuracy loss and maximum power efficiency. The extensive evaluation shows that our cross-stacked optimized accelerator SCATTER achieves a 511X area reduction and 12.4X power saving with superior crosstalk tolerance that enables unprecedented circuit layout compactness and on-chip power efficiency.

cs.AR↗

arXiv2024

Silicon Photonics Foundry Fabricated, Slow-Light Enhanced, Low Power Thermal Phase Shifter

In this research, we developed a low-power silicon photonics foundry-fabricated slow-light thermal phase shifter (SLTPS) where the slow-light (SL) effect is achieved using an integrated Bragg grating (BG) waveguide. Heating the grating induces a red shift in the transmission spectrum, leading to an increased group index $n_g$ during operation, which facilitates a further reduction in the voltage needed for a $π$ phase shift, i.e. $V_π$. Additionally, we investigated a compact Mach-Zehnder Interferometer (MZI) that incorporates the SLTPS in both arms with a phase shifter length of 50 $μ$m. A detailed theoretical analysis was conducted to address the non-idealities of the SL-MZI due to uneven optical power splitting and unbalanced loss in the two MZI arms. The $V_π$ and power consumption for a $π$ phase shift $(P_π)$ of the SL-MZI were quantified for operation in the slow light regime, demonstrating a $V_π$ of 1.1 V and a $P_π$ of 3.63 mW at an operational wavelength near the photonic band edge. The figure of merit (FOM) $P_π \times τ$ is commonly used to assess the performance of thermal optical switches. The SL-MZI in this work has achieved a low $P_π \times τ$ of 5.1 mW $μ$s. Insertion loss of the SL-MZI ranges from 1.3 dB to 4.4 dB depending on the operation wavelength, indicating a trade-off with the $V_π$ reduction.

physics.optics↗

arXiv2024

Characterization of Large Language Model Development in the Datacenter

Large Language Models (LLMs) have presented impressive performance across several transformative tasks. However, it is non-trivial to efficiently utilize large-scale cluster resources to develop LLMs, often riddled with numerous challenges such as frequent hardware failures, intricate parallelization strategies, and imbalanced resource utilization. In this paper, we present an in-depth characterization study of a six-month LLM development workload trace collected from our GPU datacenter Acme. Specifically, we investigate discrepancies between LLMs and prior task-specific Deep Learning (DL) workloads, explore resource utilization patterns, and identify the impact of various job failures. Our analysis summarizes hurdles we encountered and uncovers potential opportunities to optimize systems tailored for LLMs. Furthermore, we introduce our system efforts: (1) fault-tolerant pretraining, which enhances fault tolerance through LLM-involved failure diagnosis and automatic recovery. (2) decoupled scheduling for evaluation, which achieves timely performance feedback via trial decomposition and scheduling optimization.

cs.DC↗