skip to main content

Search for: All records

Creators/Authors contains: "Xin, Jack"

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. Free, publicly-accessible full text available September 1, 2023
  2. It has been shown by many researchers that transformers perform as well as convolutional neural networks in many computer vision tasks. Meanwhile, the large computational costs of its attention module hinder further studies and applications on edge devices. Some pruning methods have been developed to construct efficient vision transformers, but most of them have considered image classification tasks only. Inspired by these results, we propose SiDT, a method for pruning vision transformer backbones on more complicated vision tasks like object detection, based on the search of transformer dimensions. Experiments on CIFAR-100 and COCO datasets show that the backbones with 20%more »or 40% dimensions/parameters pruned can have similar or even better performance than the unpruned models. Moreover, we have also provided the complexity analysis and comparisons with the previous pruning methods.« less
    Free, publicly-accessible full text available June 1, 2023
  3. Free, publicly-accessible full text available June 1, 2023
  4. Free, publicly-accessible full text available April 1, 2023
  5. Free, publicly-accessible full text available March 1, 2023
  6. Deep Neural Networks (DNNs) need to be both efficient and robust for practical uses. Quantization and structure simplification are promising ways to adapt DNNs to mobile devices, and adversarial training is one of the most successful methods to train robust DNNs. In this work, we aim to realize both advantages by applying a convergent relaxation quantization algorithm, i.e., Binary-Relax (BR), to an adversarially trained robust model, i.e. the ResNets Ensemble via Feynman-Kac Formalism (EnResNet). We discover that high-precision quantization, such as ternary (tnn) or 4-bit, produces sparse DNNs. However, this sparsity is unstructured under adversarial training. To solve the problemsmore »that adversarial training jeopardizes DNNs’ accuracy on clean images and break the structure of sparsity, we design a trade-off loss function that helps DNNs preserve natural accuracy and improve channel sparsity. With our newly designed trade-off loss function, we achieve both goals with no reduction of resistance under weak attacks and very minor reduction of resistance under strong adversarial attacks. Together with our model and algorithm selections and loss function design, we provide an integrated approach to produce robust DNNs with high efficiency and accuracy. Furthermore, we provide a missing benchmark on robustness of quantized models.« less
    Free, publicly-accessible full text available February 1, 2023
  7. We developed an integrated recurrent neural network and nonlinear regression spatio-temporal model for vector-borne disease evolution. We take into account climate data and seasonality as external factors that correlate with disease transmitting insects (e.g. flies), also spill-over infections from neighboring regions surrounding a region of interest. The climate data is encoded to the model through a quadratic embedding scheme motivated by recommendation systems. The neighboring regions’ influence is modeled by a long short-term memory neural network. The integrated model is trained by stochastic gradient descent and tested on leishmaniasis data in Sri Lanka from 2013-2018 where infection outbreaks occurred. Ourmore »model out-performed ARIMA models across a number of regions with high infections, and an associated ablation study renders support to our modeling hypothesis and ideas.« less
    Free, publicly-accessible full text available January 1, 2023
  8. Free, publicly-accessible full text available January 1, 2023
  9. In this paper, we develop structure assisted nonnegative matrix factorization (NMF) methods for blind source separation of degenerate data. The motivation originates from nuclear magnetic resonance (NMR) spectroscopy, where a multiple mixture NMR spectra are recorded to identify chemical compounds with similar structures. Consider the linear mixing model (LMM), we aim to identify the chemical compounds involved when the mixing process is known to be nearly singular. We first consider a class of data with dominant interval(s) (DI) where each of source signals has dominant peaks over others. Besides, a nearly singular mixing process produces degenerate mixtures. The DI conditionmore »implies clustering structures in the data points. Hence, the estimation of the mixing matrix could be achieved by data clustering. Due to the presence of the noise and the degeneracy of the data, a small deviation in the estimation may introduce errors in the output. To resolve this problem and improve robustness of the separation, methods are developed in two aspects. One is to find better estimation of the mixing matrix by allowing a constrained perturbation to the clustering output, and it can be achieved by a quadratic programming. The other is to seek sparse source signals by exploiting the DI condition, and it solves an 1 optimization. If no source information is available, we propose to adopt the nonnegative matrix factorization approach by incorporating the matrix structure (parallel columns of the mixing matrix) into the cost function and develop multiplicative iteration rules for the numerical solutions. We present experimental results of NMR data to show the performance and reliability of the method in the applications arising in NMR spectroscopy.« less
    Free, publicly-accessible full text available January 1, 2023
  10. Multi-resolution paths and multi-scale feature representation are key elements of semantic segmentation networks. We develop two techniques for efficient networks based on the recent FasterSeg network architecture. One is to use a state-of-the-art high resolution network (e.g. HRNet) as a teacher to distill a light weight student network. Due to dissimilar structures in the teacher and student networks, distillation is not effective to be carried out directly in a standard way. To solve this problem, we introduce a tutor network with an added high resolution path to help distill a student network which improves FasterSeg student while maintaining its parameter/FLOPsmore »counts. The other finding is to replace standard bilinear interpolation in the upscaling module of FasterSeg student net by a depth-wise separable convolution and a Pixel Shuffle module which leads to 1.9% (1.4%) mIoU improvements on low (high) input image sizes without increasing model size. A combination of these techniques will be pursued in future works.« less
    Free, publicly-accessible full text available January 1, 2023