skip to main content


The NSF Public Access Repository (NSF-PAR) system and access will be unavailable from 10:00 PM ET on Friday, December 8 until 2:00 AM ET on Saturday, December 9 due to maintenance. We apologize for the inconvenience.

Search for: All records

Creators/Authors contains: "Zhao, Yang"

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. Abstract

    The photothermal effect in nanomaterials, resulting from resonant optical absorption, finds wide applications in biomedicine, cancer therapy, and microscopy. Despite its prevalence, the photothermal effect in light-absorbing nanoparticles has typically been assessed using bulk measurements, neglecting near-field effects. Beyond standard imaging and therapeutic uses, nanosecond-transient photothermal effects have been harnessed for bacterial inactivation, neural stimulation, drug delivery, and chemical synthesis. While scanning probe microscopy and electron microscopy offer single-particle imaging of photothermal fields, their slow speed limits observations to milliseconds or seconds, preventing nanoscale dynamic investigations. Here, we introduce decoupled optical force nanoscopy (Dofn), enabling nanometer-scale mapping of photothermal forces by exploiting unique phase responses to temporal modulation. We employ the photothermal effect’s back-action to distinguish various time frames within a modulation period. This allows us to capture the dynamic photothermal process of a single gold nanorod in the nanosecond range, providing insights into non-stationary thermal diffusion at the nanoscale.

    more » « less
  2. Free, publicly-accessible full text available April 1, 2024
  3. Free, publicly-accessible full text available June 17, 2024
  4. Vision Transformers (ViTs) have achieved state-of-the-art performance on various vision tasks. However, ViTs’ self-attention module is still arguably a major bottleneck, limiting their achievable hardware efficiency and more extensive applications to resource constrained platforms. Meanwhile, existing accelerators dedicated to NLP Transformers are not optimal for ViTs. This is because there is a large difference between ViTs and Transformers for natural language processing (NLP) tasks: ViTs have a relatively fixed number of input tokens, whose attention maps can be pruned by up to 90% even with fixed sparse patterns, without severely hurting the model accuracy (e.g., <=1.5% under 90% pruning ratio); while NLP Transformers need to handle input sequences of varying numbers of tokens and rely on on-the-fly predictions of dynamic sparse attention patterns for each input to achieve a decent sparsity (e.g., >=50%). To this end, we propose a dedicated algorithm and accelerator co-design framework dubbed ViTCoD for accelerating ViTs. Specifically, on the algorithm level, ViTCoD prunes and polarizes the attention maps to have either denser or sparser fixed patterns for regularizing two levels of workloads without hurting the accuracy, largely reducing the attention computations while leaving room for alleviating the remaining dominant data movements; on top of that, we further integrate a lightweight and learnable auto-encoder module to enable trading the dominant high-cost data movements for lower-cost computations. On the hardware level, we develop a dedicated accelerator to simultaneously coordinate the aforementioned enforced denser and sparser workloads for boosted hardware utilization, while integrating on-chip encoder and decoder engines to leverage ViTCoD’s algorithm pipeline for much reduced data movements. Extensive experiments and ablation studies validate that ViTCoD largely reduces the dominant data movement costs, achieving speedups of up to 235.3×, 142.9×, 86.0×, 10.1×, and 6.8× over general computing platforms CPUs, EdgeGPUs, GPUs, and prior-art Transformer accelerators SpAtten and Sanger under an attention sparsity of 90%, respectively. Our code implementation is available at 
    more » « less
    Free, publicly-accessible full text available February 1, 2024
  5. Abstract The jet composition in gamma-ray bursts (GRBs) is still an unsolved issue. We try to provide some clues to the issue by analyzing the spectral properties of GRB 160509A and GRB 130427A with a main burst and a postburst. We first perform Bayesian time-resolved spectral analysis and compare the spectral components and spectral properties of the main bursts and postbursts of the two bursts and find that both bursts have the thermal components, and the thermal components are mainly found in the main bursts, while the postbursts are mainly dominated by the nonthermal components. We also find that the low-energy spectral indices of some time bins in the main bursts of these two GRBs exceed the so-called synchronous dead line, and in the postburst, only GRB 160509A has four time bins exceeding the dead line, while none of GRB 130427A exceed the dead line. We then constrain the outflow properties of both bursts and find that the main bursts is consistent with the typical properties of photosphere radiation. Therefore, our results support the transition of the GRB jet component from the fireball to the Poynting-flux-dominated jet. Finally, after analyzing the correlation and parameter evolution of the spectral parameters of the two bursts, we find that the correlations of the spectral parameters have different behaviors in the main bursts and postbursts. The parameter evolution trends of the main bursts and postbursts also show consistent and inconsistent behavior; therefore, we currently cannot determine whether the main bursts and postbursts come from the same origin. 
    more » « less
  6. We present a double-yield-surface plasticity theory for transversely isotropic rocks that distinguishes between plastic deformation through the solid matrix and localized plasticity along the weak bedding planes. A recently developed anisotropic modified Cam-Clay model is adopted to model the plastic response of the solid matrix, while the Mohr-Coulomb friction law is used to represent the sliding mechanism along the weak bedding planes. For its numerical implementation, we derive an implicit return mapping algorithm for both the semi-plastic and fully plastic loading processes, as well as the corresponding algorithmic tangent operator for finite element problems. We validate the model with triaxial compression test data for three different transversely isotropic rocks and reproduce the undulatory variation of rock strength with bedding plane orientation. We also implement the proposed model in a finite element setting and investigate the deformation of rock surrounding a borehole subjected to fluid injection. We compare the results of simulations using the proposed double-yield-surface model with those generated using each single yield criterion to highlight the features of the proposed theory. 
    more » « less
  7. We present a first-of-its-kind ultra-compact intelligent camera system, dubbed i-FlatCam, including a lensless camera with a computational (Comp.) chip. It highlights (1) a predict-then-focus eye tracking pipeline for boosted efficiency without compromising the accuracy, (2) a unified compression scheme for single-chip processing and improved frame rate per second (FPS), and (3) dedicated intra-channel reuse design for depth-wise convolutional layers (DW-CONV) to increase utilization. i-FlatCam demonstrates the first eye tracking pipeline with a lensless camera and achieves 3.16 degrees of accuracy, 253 FPS, 91.49 µJ/Frame, and 6.7mm×8.9mm×1.2mm camera form factor, paving the way for next-generation Augmented Reality (AR) and Virtual Reality (VR) devices. 
    more » « less