arXiv subjects

Chang Zhang

Publications and source records attributed to Chang Zhang.

At least 19 recordsLinked to original sources

arXiv2025

MoA-Off: Adaptive Heterogeneous Modality-Aware Offloading with Edge-Cloud Collaboration for Efficient Multimodal LLM Inference

Multimodal large language models (MLLMs) enable powerful cross-modal inference but impose significant computational and latency burdens, posing severe challenges for deployment in resource-constrained environments. In this paper, we propose MoA-Off, an adaptive heterogeneous modality-aware offloading framework with edge-cloud collaboration for efficient MLLM inference. MoA-Off introduces a lightweight heterogeneous modality-aware module that estimates the complexity of heterogeneous inputs through multi-dimensional feature analysis. Then, an adaptive edge-cloud collaborative offloading strategy is proposed that dynamically schedules workloads between edge and cloud based on modality-aware complexity scores and real-time system states. The experimental results demonstrate that MoA-Off can achieve over 30% reduction in latency and 30%-65% decrease in resource overhead while maintaining competitive accuracy compared to traditional approaches.

cs.DC↗

arXiv2025

Laser micromachining of arbitrarily complex and overhang-free SiN nanomechanical resonators

Research on silicon nitride (SiN) nanomechanical resonators produces an exceptionally rich variety of resonator geometries, for which there is currently no available rapid prototyping solution. Experimental advances in nanobeam, trampoline, phononic bandgap, and soft-clamping structures all rely on conventional nanofabrication involving e-beam or photolithography, followed by various etching steps. These techniques are typically time-consuming, relatively inflexible, and often result in spurious residual SiN overhang that can degrade mechanical quality factors. In contrast, recent work has shown that simple resonant structures, such as nanobeams, can be prototyped by direct laser ablation of free-standing SiN membranes using a spatially distributed sequence of microholes that limits stress concentration. However, these early demonstrations were restricted to basic shapes, created by manually combining ablation routines for circles and straight lines. Here, we demonstrate the fabrication of arbitrarily complex geometries using an open-source software toolset--released with this publication--that automatically generates laser-ablated hole sequences directly from standard semiconductor layout files (i.e., GDSII). The software includes a layout alignment tool that compensates for the membrane orientation and dimensional variations, limiting material overhang to ~2 um. Using this toolset, we fabricate several resonator geometries, each in under 1 hour, two of which are exhaustively characterized as candidate structures for high-performance radiation sensing. The measured quality factors of these structures closely match finite element simulations and reach values up to 3.7 x 10^6. From these measurements, we extract material quality factors above 3700, which is on par with low-stress SiN unablated plain membranes and with comparable structures produced using conventional fabrication methods.

physics.optics↗

arXiv2025

Properties of the cores and filaments in the Ophiuchus molecular cloud and its L1688 hub-filament system

Analyzing filaments and cores in molecular clouds is key to understanding galactic star formation and its environmental dependence. This paper studies the properties and distribution of dense cores and filaments in the Ophiuchus molecular cloud, with a focus on the L1688 hub-filament system (HFS) and its star formation potential. We extracted sources and filaments from Herschel images and a 13.5 arcsec resolution surface density map using the getsf method, identified prestellar cores among the extracted sources, evaluated core mass segregation, and constructed the core mass function (CMF). We derived properties of the filaments from their radial surface density profiles, constructed the filament linear density function (FLDF), and assessed the mass distribution in the L1688 HFS to estimate the core and filament formation efficiencies (CFE, FFE). The results suggest that the filament-dominated core formation is a key mechanism in star formation within the system.

astro-ph.GA↗

arXiv2025

Enhanced bandwidth in radiation sensors operating at the fundamental temperature fluctuation noise limit

Temperature-based radiation detectors are an essential tool for long optical wavelengths detection even if they often suffer from important bandwidth limitations. Their responsivity, and hence their noise equivalent power (NEP), typically degrade at frequencies exceeding the cutoff set by their characteristic thermal response time ($\tau_\text{th}$), i.e., at $\omega > \tau_\text{th}^{-1}$. Here we show that this bandwidth limitation can be broken when a radiation sensor operates at its fundamental temperature fluctuation noise limit. The key enabler of this demonstration is a nanomechanical sensor in which frequency stability is limited by fundamental temperature fluctuations over an unprecedentedly large bandwidth of 54 $\text{Hz}$. In this range, the sensor performance remains within a factor 3 from its peak detectivity ($D_T^* = 7.4 \times 10^9~\mathrm{cm \cdot Hz^{1/2} W^{-1}}$) even though the thermal cutoff frequency is 30 times lower (i.e., $1/2\mathrm{\pi} \tau_\text{th} = 1.8~\text{Hz}$). We also derive and validate experimentally closed-form expression predicting maximum bandwidth enhancement in the context of nanomechanical resonators interfaced with a closed-loop frequency tracking scheme.

physics.optics↗

arXiv2025

Well-posedness and global attractor for wave equation with displacement dependent damping and super-cubic nonlinearity

This work investigates the semilinear wave equation featuring the displacement dependent term $\sigma(u)\partial_t u $ and nonlinearity $f(u)$. By developing refined space-time a priori estimates under extended ranges of the nonlinearity exponents with $\sigma(u)$ and $f(u)$, the well-posedness of the weak solution is established. Furthermore, the existence of a global attractor in the naturally phase space $H^1_0(\Omega)\times L^2(\Omega)$ is obtained. Moreover, the regularity of the global attractor is established, implying that it is a bounded subset of $(H^2(\Omega)\cap H^1_0(\Omega))\times H^1_0(\Omega)$.

math.AP↗

arXiv2025

Long-term behavior for wave equation with nonlinear damping and super-cubic nonlinearity

In this paper, we consider the semilinear wave equation involving the nonlinear damping term $g(u_t) $ and nonlinearity $f(u)$. The well-posedness of the weak solution satisfying some additional regularity is achieved under the wider ranges of the exponents $g$ and $f$. Moreover, the existence of global attractor and exponential attractor are proved.

math.AP↗

arXiv2024

Monte Carlo sampling with integrator snippets

Assume interest is in sampling from a probability distribution $\mu$ defined on $(\mathsf{Z},\mathscr{Z})$. We develop a framework for sampling algorithms which takes full advantage of ODE numerical integrators, say $\psi\colon\mathsf{Z}\rightarrow\mathsf{Z}$ for one integration step, to explore $\mu$ efficiently and robustly. The popular Hybrid Monte Carlo (HMC) algorithm \cite{duane1987hybrid,neal2011mcmc} and its derivatives are examples of such a use of numerical integrators. A key idea developed here is that of sampling integrator snippets, that is fragments of the orbit of an ODE numerical integrator $\psi$, and the definition of an associated probability distribution $\bar{\mu}$ such that expectations with respect to $\mu$ can be estimated from integrator snippets distributed according to $\bar{\mu}$. The integrator snippet target distribution $\bar{\mu}$ takes the form of a mixture of pushforward distributions which suggests numerous generalisations beyond mappings arising from numerical integrators, e.g. normalising flows. Very importantly this structure also suggests new principled and robust strategies to tune the parameters of integrators, such as the discretisation stepsize, effective integration time, or number of integration steps, in a Leapfrog integrator. We focus here primarily on Sequential Monte Carlo (SMC) algorithms, but the approach can be used in the context of Markov chain Monte Carlo algorithms. We illustrate performance and, in particular, robustness through numerical experiments and provide preliminary theoretical results supporting observed performance.

stat.CO↗

arXiv2024

High detectivity terahertz radiation sensing using frequency-noise-optimized nanomechanical resonators

We achieve high detectivity terahertz sensing using a silicon nitride nanomechanical resonator functionalized with a metasurface absorber. High performances are achieved by striking a fine balance between the frequency stability of the resonator, and its responsivity to absorbed radiation. Using this approach, we demonstrate a detectivity $D^*=3.4\times10^9~\mathrm{cm\cdot\sqrt{Hz}/W}$ and a noise equivalent power $\mathrm{NEP}=36~\mathrm{pW/\sqrt{Hz}}$ that outperform the best room-temperature on-chip THz detectors (i.e., pyroelectrics). Our optical absorber consists of a 1-mm diameter metasurface, which currently enables a 0.5-3 THz detection range but can easily be scaled to other frequencies in the THz and infrared ranges. In addition to demonstrating high-performance terahertz sensing, our work unveils an important fundamental trade-off between high frequency stability and high responsivity in thermal-based nanomechanical radiation sensors.

physics.optics↗

arXiv2023

Well-posedness and global attractor for wave equation with nonlinear damping and super-cubic nonlinearity

This study investigates a semilinear wave equation characterized by nonlinear damping $g(u_t) $ and nonlinearity $f(u)$. First, the well-posedness of weak solutions across broader exponent ranges for $g$ and $f$ is established, by utilizing a priori space-time estimates. Moreover, the existence of a global attractor in the phase space $H^1_0(\Omega)\times L^2(\Omega)$ is obtained. Furthermore, it is proved that this global attractor is regular, implying that it is a bounded subset of $(H^2(\Omega)\cap H^1_0(\Omega))\times H^1_0(\Omega)$.

math.AP↗

arXiv2023

Properties of the dense cores and filamentary structures in the Vela C molecular cloud

The initial and boundary conditions of the Galactic star formation in molecular clouds are not well understood. In an effort to shed new light on this long-standing problem, we measured properties of dense cores and filamentary structures in the Vela C molecular cloud, observed with Herschel. We applied the getsf extraction method to separate the components of sources and filaments from each other and their backgrounds, before detecting, measuring, and cataloging the structures. The cores and filamentary structures constitute 40% of the total mass of Vela C, most of the material is in the low-density molecular background cloud. We selected 570 reliable cores, of which 149 are the protostellar cores and 421 are the starless cores. Almost 78% of the starless cores were identified with the gravitationally bound prestellar cores. The exponent of the CMF (alpha = 1.35) is identical to that of the Salpeter IMF. We selected 68 filaments with at least one side that appeared not blended with adjacent structures. The filament widths are in the range of 0.15 pc to 0.63 pc, and have a median value of W = 0.3(0.11) pc. The surface densities of filaments are well correlated with their contrasts and linear densities. Within uncertainties of the filament instability criterion, many filaments may well be both supercritical and subcritical. A large fraction of filaments may definitely be considered supercritical, in which are found 94 prestellar cores, 83 protostellar cores, and only 1 unbound starless core. Taking into account the uncertainties, the supercritical filaments contain only prestellar and protostellar cores. Our findings support the idea that there exists a direct relationship between the CMF and IMF and that filaments play a key role in the formation of prestellar cores, which is consistent with the previous Herschel results.

astro-ph.GA↗

arXiv2023

Maximal co-occurrence nonoverlapping sequential rule mining

The aim of sequential pattern mining (SPM) is to discover potentially useful information from a given se-quence. Although various SPM methods have been investigated, most of these focus on mining all of the patterns. However, users sometimes want to mine patterns with the same specific prefix pattern, called co-occurrence pattern. Since sequential rule mining can make better use of the results of SPM, and obtain better recommendation performance, this paper addresses the issue of maximal co-occurrence nonoverlapping sequential rule (MCoR) mining and proposes the MCoR-Miner algo-rithm. To improve the efficiency of support calculation, MCoR-Miner employs depth-first search and backtracking strategies equipped with an indexing mechanism to avoid the use of sequential searching. To obviate useless support calculations for some sequences, MCoR-Miner adopts a filtering strategy to prune the sequences without the prefix pattern. To reduce the number of candidate patterns, MCoR-Miner applies the frequent item and binomial enumeration tree strategies. To avoid searching for the maximal rules through brute force, MCoR-Miner uses a screening strategy. To validate the per-formance of MCoR-Miner, eleven competitive algorithms were conducted on eight sequences. Our experimental results showed that MCoR-Miner outperformed other competitive algorithms, and yielded better recommendation performance than frequent co-occurrence pattern mining. All algorithms and datasets can be downloaded from https://github.com/wuc567/Pattern-Mining/tree/master/MCoR-Miner.

cs.DB↗

arXiv2023

Demonstration of Frequency Stability limited by Thermal Fluctuation Noise in Silicon Nitride Nanomechanical Resonators

The frequency stability of nanomechanical resonators (NMR) dictates the performance level of many state-of-the-art sensors (e.g., mass, force, temperature, radiation) that relate an external physical perturbation to a resonance frequency shift. While this is obviously of fundamental importance, accurate models and understandings of sources of frequency instability are not always available. The contribution of thermomechanical noise to frequency stability has been well studied in recent years and is often the fundamental performance limitation. Frequency stability limited by thermal fluctuation noise has attracted less interest but is nevertheless of fundamental importance notably in temperature sensing applications. In particular, temperature-sensitive NMR have become promising candidates for replacing traditional bolometers in infrared radiation sensing. However, reaching the ultimate detectivity limit of thermal radiation sensors requires their noise to be dominated by fundamental thermal fluctuation, which has not been demonstrated to date. In this work, we first develop a theoretical model for computing the frequency stability of NMR by considering the effect of both additive phase noise (i.e., thermomechanical, and experimental detection noise) and thermal fluctuation noise in a close-loop frequency tracking scheme. We thereafter validate this model experimentally and observe thermal fluctuation noise in SiN drum resonators of various sizes. Our work shows that by using resonators of specific characteristics--such as high temperature sensitivity, high mechanical quality factors, and high mass-to-thermal-conductance ratio--one can minimize additive phase noise below thermal fluctuation noise. This paves the way for NMR-based radiation sensors that can reach the fundamental detectivity limit of thermal radiation sensing and outperform existing technologies.

physics.app-ph↗

arXiv2022

WISE Green Objects (WGOs): the massive star candidates in the whole Galactic Plane ($\mid b \mid <2^\circ$)

Massive young stellar objects (MYSOs) play a crucial role in star formation. Given that MYSOs were previously identified based on the extended structure and the observational data for them is limited, screening the Wide-field Infrared Survey Explorer (WISE) objects showing green features (for the common coding of the 4.6 $μ$m band as green channel in three-color composite WISE images) will yield more MYSO candidates. Using WISE images in the whole Galactic Plane ($ 0^\circ 60^\circ$. Matching results with various masers show that those three groups of WGOs are at different evolutionary stages. After cross-matching WGOs with published YSO survey catalogs, we infer that $\sim$50% of WGOs are samples of newly discovered YSOs. In addition, 1260 WGOs are associated with Hi-GAL sources, according to physical parameters estimated by spectral energy distribution fitting, of which 231 are classified as robust MYSOs and 172 as candidate MYSOs.

astro-ph.GA↗

arXiv2022

$Herschel$ investigation of cores and filamentary structures in the Perseus molecular cloud

Cores and filamentary structures are the prime birthplaces of stars, and play key roles in the process of star formation. Latest advances in the methods of multi-scale source and filament extraction, and in making high-resolution column density map from $Herschel$ multi-wavelength observations enable us to detect the filamentary network structures in highly complex molecular cloud environments. The statistics for physical parameters shows that core mass strongly correlates with core dust temperature, and $M/L$ strongly correlates with $M/T$, which is in line with the prediction of the blackbody radiation, and can be used to trace evolutionary sequence from unbound starless cores to robust prestellar cores. Crest column densities of the filamentary structures are clearly related with mass per unit length ($M_{\rm line}$), but are uncorrelated by three orders ranging from $\sim 10^{20}$ to $\sim 10^{22}$ $ \rm cm^{-2}$ with widths. Full width at half maximum (FWHM) have a median value of 0.15 pc, which is consistent with the 0.1 pc typical inner width of the filamentary structures reported by previous research. We find $\sim $70\% of robust prestellar cores (135/199) embedded in supercritical filaments with $M_{\rm line}>16~M_{\odot}/{\rm pc}$, which implies that the gravitationally bound cores come from fragmentation of supercritical filaments. And on the basis of observational evidences that probability distribution function (PDF) with power-law distribution in the Perseus south is flatter than north, YSO number is significantly less than that in the north, and dust temperature difference. We infer that south region is more gravitationally bound than north region.

astro-ph.GA↗

arXiv2022

ATOMS: ALMA Three-millimeter Observations of Massive Star-forming regions -- VII. A catalogue of SiO clumps from ACA observations

To understand the nature of SiO emission, we conducted ACA observations of the SiO (2-1) lines toward 146 massive star-forming regions, as part of the ALMA Three-millimeter Observations of Massive Star-forming regions (ATOMS) survey. We detected SiO emission in 128 (87.7$\%$) sources and identified 171 SiO clumps, 105 of which are spatially separated from 3 mm continuum emission. A large amount of the SiO line profiles (60$\%$) are non-Gaussian. The velocity dispersion of the SiO lines ranges from 0.3 to 5.43 km s$^{-1}$. In 63 sources the SiO clumps are associated with H$_\rm{II}$ regions characterized by H40$α$ emission. We find that 68$\%$ (116) of the SiO clumps are associated with strong outflows. The median velocity dispersion of the SiO line for outflow sources and non-outflow sources is 1.91 km s$^{-1}$ and 0.99 km s$^{-1}$, respectively. These results indicate that outflow activities could be connected to strongly shocked gas. The velocity dispersion and [SiO]/[H$^{13}$CO$^+$] intensity ratio do not show any correlation with the dust temperature and particle number density of clumps. We find a positive correlation between the SiO line luminosity and the bolometric luminosity, implying stronger shock activities are associated with more luminous proto-clusters. The SiO clumps in associations with H$_\rm{II}$ regions were found to show a steeper feature in $L_\rm{sio}$/$L_\rm{bol}$. The SiO line luminosity and the fraction of shocked gas have no apparent evidence of correlation with the evolutionary stages traced by luminosity to mass ratio ($L_\rm{bol}/M$).

astro-ph.GA↗

arXiv2021

Observation of Silicon Nitride Nanomechanical Resonator Actuation Using Capacitive Substrate Excitation

We observe the actuation of silicon nitride (SiN) nanomechanical resonators by electrical excitation of metal-dielectric-semiconductor (MDS) capacitors on their supporting silicon substrate. We develop first-principle models explaining this actuation mechanism by acoustic waves resulting from voltage-dependent electrostatic forces in the MDS capacitors. Models are developed for actuation in the charge accumulation (Ni-pSi) and charge depletion (Al-pSi) regimes. Experimental observations confirm our prediction that charge accumulation (Ni-pSi) is more efficient at actuation than charge depletion. For a 2 V actuation signal, Ni-pSi capacitors achieve 10 nm actuation amplitude in square (1.7 $\times$ 1.7 mm) low-stress (~100 MPa) SiN membrane resonators. In this case, electrical power dissipation in the chip is on the order of 0.1 $μ$W, and spurious heating is less than 1 mK. Both these values could be further reduced by doping the substrate to minimize resistive dissipation. The actuation method is remarkably simple and only requires attachment of wires to the chip with vacuum-compatible nickel paste, with no extra photolithography step. All the chips presented in this work are fabricated in-house, and a detailed fabrication procedure is provided.

physics.app-ph↗

arXiv2021

Heat Transport in Silicon Nitride Drum Resonators and its Influence on Thermal Fluctuation-induced Frequency Noise

Silicon nitride (SiN) drumhead resonators offer a promising platform for thermal sensing due to their high mechanical quality factor and the high temperature sensitivity of their resonance frequency. As such, gaining an understanding of heat transport in SiN resonators as well as their sensing noise limitations is of interest, both of which are goals of the present work. We first present new experimental results on radiative heat transport in SiN membrane, which we use for benchmarking two recently proposed theoretical models. We measure the characteristic thermal response time of square SiN membranes with a thickness of 90 $\pm$ 1.7 nm and side lengths from 1.5 to 12 mm. A clear transition between radiation and conduction dominated heat transport is measured, in close correspondence with theory. In the second portion of this work, we use our experimentally validated heat transport model to provide a closed-form expression for thermal fluctuation-induced frequency noise in SiN membrane resonators. We find that, for large area SiN membranes, thermal fluctuations can be greater than thermomechanical contributions to frequency noise. For the specific case of thermal radiation sensing applications, we also derive the noise equivalent power resulting from thermal fluctuation-induced frequency noise, and we show in which conditions it reduces to the classical detectivity limit of thermal radiation sensors. Our work therefore provides a path towards achieving thermal radiation sensors operating at the never attained fundamental detectivity limit of bolometric sensing. We also identify questions that remain when attempting to push the limits of radiation sensing, in particular, the effect of thermal fluctuation noise in closed-loop frequency tracking schemes remains to be clarified.

physics.app-ph↗

arXiv2021

Radiative Heat Transfer in Free-Standing Silicon Nitride Membranes

Free-standing silicon nitride (SiN) mechanical resonators are of central interests in applications such as temperature and mass sensing, and for fundamental optomechanical reasearch. Understanding thermal coupling between a membrane resonator and its environment is required for predicting thermal noise, frequency noise, as well as sensors responses to temperature changes. In this work, we provide closed-form derivations of intrinsic thermal coupling quantities in free-standing thin films, namely total thermal conductance with the surroundings, thermal response time, and the relative contribution of thermal radiation. Our model is valid for any free-standing thin film anchored on all sides, although we particularly emphasize the specific case of SiN for which spectral emissivity is thoroughly investigated as a function of thickness and temperature. We find that radiative heat exchanges can play a non-negligible role, and even dominate thermal coupling for membranes of sizes commonly employed in optomechanics experiments. We experimentally confirm the validity of our model by measuring radiative thermal coupling between a SiN mechanical resonator and a ceramic heater in high vacuum.

physics.app-ph↗