arXiv subjects

Zhiwei Jia

Publications and source records attributed to Zhiwei Jia.

22 records · Page 2Linked to original sources

arXiv2020

One-pixel Signature: Characterizing CNN Models for Backdoor Detection

We tackle the convolution neural networks (CNNs) backdoor detection problem by proposing a new representation called one-pixel signature. Our task is to detect/classify if a CNN model has been maliciously inserted with an unknown Trojan trigger or not. Here, each CNN model is associated with a signature that is created by generating, pixel-by-pixel, an adversarial value that is the result of the largest change to the class prediction. The one-pixel signature is agnostic to the design choice of CNN architectures, and how they were trained. It can be computed efficiently for a black-box CNN model without accessing the network parameters. Our proposed one-pixel signature demonstrates a substantial improvement (by around 30% in the absolute detection accuracy) over the existing competing methods for backdoored CNN detection/classification. One-pixel signature is a general representation that can be used to characterize CNN models beyond backdoor detection.

cs.CV↗

arXiv2020

Information-Theoretic Local Minima Characterization and Regularization

Recent advances in deep learning theory have evoked the study of generalizability across different local minima of deep neural networks (DNNs). While current work focused on either discovering properties of good local minima or developing regularization techniques to induce good local minima, no approach exists that can tackle both problems. We achieve these two goals successfully in a unified manner. Specifically, based on the observed Fisher information we propose a metric both strongly indicative of generalizability of local minima and effectively applied as a practical regularizer. We provide theoretical analysis including a generalization bound and empirically demonstrate the success of our approach in both capturing and improving the generalizability of DNNs. Experiments are performed on CIFAR-10, CIFAR-100 and ImageNet for various network architectures.

cs.LG↗

arXiv2018

Controllable Top-down Feature Transformer

We study the intrinsic transformation of feature maps across convolutional network layers with explicit top-down control. To this end, we develop top-down feature transformer (TFT), under controllable parameters, that are able to account for the hidden layer transformation while maintaining the overall consistency across layers. The learned generators capture the underlying feature transformation processes that are independent of particular training images. Our proposed TFT framework brings insights to and helps the understanding of, an important problem of studying the CNN internal feature representation and transformation under the top-down processes. In the case of spatial transformations, we demonstrate the significant advantage of TFT over existing data-driven approaches in building data-independent transformations. We also show that it can be adopted in other applications such as data augmentation and image style transfer.

cs.CV↗

arXiv2016

Phase-locked array of quantum cascade lasers with an intracavity spatial filter

Phase-locking an array of quantum cascade lasers is an effective way to achieve higher output power and beam shaping. In this article, based on Talbot effect, we show a new-type phase-locked array of mid-infrared quantum cascade lasers with an integrated spatial- filtering Talbot cavity. All the arrays show stable in-phase operation from the threshold current to full power current. The beam divergence of the array device is smaller than that of a single-ridge laser. We use the multi-slit Fraunhofer diffraction mode to interpret the far-field radiation profile and give a solution to get better beam quality. The maximum power is just about 5 times that of a single-ridge laser for eleven-laser array device and 3 times for seven-laser array device. Considering the great modal selection ability, simple fabricating process and the potential for achieving better beam quality and smaller cavity loss, this new-type phase-locked array may be a hopeful and elegant solution to get high power or beam shaping.

physics.optics↗