arXiv subjects

Ye Yu

Publications and source records attributed to Ye Yu.

At least 37 records · Page 2Linked to original sources

arXiv2024

Prolonging Carrier Lifetime in P-type 4H-SiC Epilayer by Thermal Oxidation and Hydrogen Annealing

A minority carrier lifetime of 25.46 $μ$s in a P-type 4H-SiC epilayer has been attained through sequential thermal oxidation and hydrogen annealing. Thermal oxidation can enhance the minority carrier lifetime in the 4H-SiC epilayer by reducing carbon vacancies. However, this process also generates carbon clusters with limited diffusivity and contributes to the enlargement of surface pits on the 4H-SiC. High-temperature hydrogen annealing effectively reduces stacking fault and dislocation density. Moreover, electron spin resonance analysis indicates a significant reduction in carbon vacancy defects after hydrogen annealing. The mechanisms of the elimination of carbon vacancies by hydrogen annealing include the decomposition of carbon clusters formed during thermal oxidation and the low-pressure selective etching by hydrogen, which increases the carbon content on the 4H-SiC surface and facilitates carbon diffusion. Consequently, the combination of thermal oxidation and hydrogen annealing eliminates carbon vacancies more effectively, substantially enhancing the minority carrier lifetime in P-type 4H-SiC. This improvement is advantageous for the application of high-voltage SiC bipolar devices.

cond-mat.mtrl-sci↗

arXiv2024

LoSA: Long-Short-range Adapter for Scaling End-to-End Temporal Action Localization

Temporal Action Localization (TAL) involves localizing and classifying action snippets in an untrimmed video. The emergence of large video foundation models has led RGB-only video backbones to outperform previous methods needing both RGB and optical flow modalities. Leveraging these large models is often limited to training only the TAL head due to the prohibitively large GPU memory required to adapt the video backbone for TAL. To overcome this limitation, we introduce LoSA, the first memory-and-parameter-efficient backbone adapter designed specifically for TAL to handle untrimmed videos. LoSA specializes for TAL by introducing Long-Short-range Adapters that adapt the intermediate layers of the video backbone over different temporal ranges. These adapters run parallel to the video backbone to significantly reduce memory footprint. LoSA also includes Long-Short-range Gated Fusion that strategically combines the output of these adapters from the video backbone layers to enhance the video features provided to the TAL head. Experiments show that LoSA significantly outperforms all existing methods on standard TAL benchmarks, THUMOS-14 and ActivityNet-v1.3, by scaling end-to-end backbone adaptation to billion-parameter-plus models like VideoMAEv2~(ViT-g) and leveraging them beyond head-only transfer learning.

cs.CV↗

arXiv2024

LMT-GP: Combined Latent Mean-Teacher and Gaussian Process for Semi-supervised Low-light Image Enhancement

While recent low-light image enhancement (LLIE) methods have made significant advancements, they still face challenges in terms of low visual quality and weak generalization ability when applied to complex scenarios. To address these issues, we propose a semi-supervised method based on latent mean-teacher and Gaussian process, named LMT-GP. We first design a latent mean-teacher framework that integrates both labeled and unlabeled data, as well as their latent vectors, into model training. Meanwhile, we use a mean-teacher-assisted Gaussian process learning strategy to establish a connection between the latent and pseudo-latent vectors obtained from the labeled and unlabeled data. To guide the learning process, we utilize an assisted Gaussian process regression (GPR) loss function. Furthermore, we design a pseudo-label adaptation module (PAM) to ensure the reliability of the network learning. To demonstrate our method's generalization ability and effectiveness, we apply it to multiple LLIE datasets and high-level vision tasks. Experiment results demonstrate that our method achieves high generalization performance and image quality. The code is available at https://github.com/HFUT-CV/LMT-GP.

cs.CV↗

arXiv2023

Rethinking Multimodal Content Moderation from an Asymmetric Angle with Mixed-modality

There is a rapidly growing need for multimodal content moderation (CM) as more and more content on social media is multimodal in nature. Existing unimodal CM systems may fail to catch harmful content that crosses modalities (e.g., memes or videos), which may lead to severe consequences. In this paper, we present a novel CM model, Asymmetric Mixed-Modal Moderation (AM3), to target multimodal and unimodal CM tasks. Specifically, to address the asymmetry in semantics between vision and language, AM3 has a novel asymmetric fusion architecture that is designed to not only fuse the common knowledge in both modalities but also to exploit the unique information in each modality. Unlike previous works that focus on representing the two modalities into a similar feature space while overlooking the intrinsic difference between the information conveyed in multimodality and in unimodality (asymmetry in modalities), we propose a novel cross-modality contrastive loss to learn the unique knowledge that only appears in multimodality. This is critical as some harmful intent may only be conveyed through the intersection of both modalities. With extensive experiments, we show that AM3 outperforms all existing state-of-the-art methods on both multimodal and unimodal CM benchmarks.

cs.CV↗

arXiv2023

Rule By Example: Harnessing Logical Rules for Explainable Hate Speech Detection

Classic approaches to content moderation typically apply a rule-based heuristic approach to flag content. While rules are easily customizable and intuitive for humans to interpret, they are inherently fragile and lack the flexibility or robustness needed to moderate the vast amount of undesirable content found online today. Recent advances in deep learning have demonstrated the promise of using highly effective deep neural models to overcome these challenges. However, despite the improved performance, these data-driven models lack transparency and explainability, often leading to mistrust from everyday users and a lack of adoption by many platforms. In this paper, we present Rule By Example (RBE): a novel exemplar-based contrastive learning approach for learning from logical rules for the task of textual content moderation. RBE is capable of providing rule-grounded predictions, allowing for more explainable and customizable predictions compared to typical deep learning-based approaches. We demonstrate that our approach is capable of learning rich rule embedding representations using only a few data examples. Experimental results on 3 popular hate speech classification datasets show that RBE is able to outperform state-of-the-art deep learning classifiers as well as the use of rules in both supervised and unsupervised settings while providing explainable model predictions via rule-grounding.

cs.CL↗

arXiv2022

Mid- and far-infrared localized surface plasmon resonances in chalcogen-hyperdoped silicon

Plasmonic sensing in the infrared region employs the direct interaction of the vibrational fingerprints of molecules with the plasmonic resonances, creating surface-enhanced sensing platforms that are superior than the traditional spectroscopy. However, the standard noble metals used for plasmonic resonances suffer from high radiative losses as well as fabrication challenges, such as tuning the spectral resonance positions into mid- to far-infrared regions, and the compatibility issue with the existing complementary metal-oxide-semiconductor (CMOS) manufacturing platform. Here, we demonstrate the occurrence of mid-infrared localized surface plasmon resonances (LSPR) in thin Si films hyperdoped with the known deep-level impurity tellurium. We show that the mid-infrared LSPR can be further enhanced and spectrally extended to the far-infrared range by fabricating two-dimensional arrays of micrometer-sized antennas in a Te-hyperdoped Si chip. Since Te-hyperdoped Si can also work as an infrared photodetector, we believe that our results will unlock the route toward the direct integration of plasmonic sensors with the one-chip CMOS platform, greatly advancing the possibility of mass manufacturing of high-performance plasmonic sensing systems.

cond-mat.mtrl-sci↗

arXiv2022

BATMAN: Bilateral Attention Transformer in Motion-Appearance Neighboring Space for Video Object Segmentation

Video Object Segmentation (VOS) is fundamental to video understanding. Transformer-based methods show significant performance improvement on semi-supervised VOS. However, existing work faces challenges segmenting visually similar objects in close proximity of each other. In this paper, we propose a novel Bilateral Attention Transformer in Motion-Appearance Neighboring space (BATMAN) for semi-supervised VOS. It captures object motion in the video via a novel optical flow calibration module that fuses the segmentation mask with optical flow estimation to improve within-object optical flow smoothness and reduce noise at object boundaries. This calibrated optical flow is then employed in our novel bilateral attention, which computes the correspondence between the query and reference frames in the neighboring bilateral space considering both motion and appearance. Extensive experiments validate the effectiveness of BATMAN architecture by outperforming all existing state-of-the-art on all four popular VOS benchmarks: Youtube-VOS 2019 (85.0%), Youtube-VOS 2018 (85.3%), DAVIS 2017Val/Testdev (86.2%/82.2%), and DAVIS 2016 (92.5%).

cs.CV↗

arXiv2022

GateHUB: Gated History Unit with Background Suppression for Online Action Detection

Online action detection is the task of predicting the action as soon as it happens in a streaming video. A major challenge is that the model does not have access to the future and has to solely rely on the history, i.e., the frames observed so far, to make predictions. It is therefore important to accentuate parts of the history that are more informative to the prediction of the current frame. We present GateHUB, Gated History Unit with Background Suppression, that comprises a novel position-guided gated cross-attention mechanism to enhance or suppress parts of the history as per how informative they are for current frame prediction. GateHUB further proposes Future-augmented History (FaH) to make history features more informative by using subsequently observed frames when available. In a single unified framework, GateHUB integrates the transformer's ability of long-range temporal modeling and the recurrent model's capacity to selectively encode relevant information. GateHUB also introduces a background suppression objective to further mitigate false positive background frames that closely resemble the action frames. Extensive validation on three benchmark datasets, THUMOS, TVSeries, and HDD, demonstrates that GateHUB significantly outperforms all existing methods and is also more efficient than the existing best work. Furthermore, a flow-free version of GateHUB is able to achieve higher or close accuracy at 2.8x higher frame rate compared to all existing methods that require both RGB and optical flow information for prediction.

cs.CV↗

arXiv2022

Asymmetric topological pumping in nonparaxial photonics

Topological photonics was initially inspired by the quantum-optical analogy between the Schrödinger equation for an electron wavefunction and the paraxial equation for a light beam. Here, we reveal an unexpected phenomenon in topological pumping observed in arrays of nonparaxial optical waveguides where the quantum-optical analogy becomes invalid. We predict theoretically and demonstrate experimentally an asymmetric topological pumping when the injected field transfers from one side of the waveguide array to the other side whereas the reverse process is unexpectedly forbidden. Our finding could open an avenue for exploring topological photonics that enables nontrivial topological phenomena and designs in photonics driven by nonparaxiality.

physics.optics↗

arXiv2021

Stronger NAS with Weaker Predictors

Neural Architecture Search (NAS) often trains and evaluates a large number of architectures. Recent predictor-based NAS approaches attempt to alleviate such heavy computation costs with two key steps: sampling some architecture-performance pairs and fitting a proxy accuracy predictor. Given limited samples, these predictors, however, are far from accurate to locate top architectures due to the difficulty of fitting the huge search space. This paper reflects on a simple yet crucial question: if our final goal is to find the best architecture, do we really need to model the whole space well?. We propose a paradigm shift from fitting the whole architecture space using one strong predictor, to progressively fitting a search path towards the high-performance sub-space through a set of weaker predictors. As a key property of the weak predictors, their probabilities of sampling better architectures keep increasing. Hence we only sample a few well-performed architectures guided by the previously learned predictor and estimate a new better weak predictor. This embarrassingly easy framework, dubbed WeakNAS, produces coarse-to-fine iteration to gradually refine the ranking of sampling space. Extensive experiments demonstrate that WeakNAS costs fewer samples to find top-performance architectures on NAS-Bench-101 and NAS-Bench-201. Compared to state-of-the-art (SOTA) predictor-based NAS methods, WeakNAS outperforms all with notable margins, e.g., requiring at least 7.5x less samples to find global optimal on NAS-Bench-101. WeakNAS can also absorb their ideas to boost performance more. Further, WeakNAS strikes the new SOTA result of 81.3% in the ImageNet MobileNet Search Space. The code is available at https://github.com/VITA-Group/WeakNAS.

cs.LG↗

arXiv2021

MUSE: Feature Self-Distillation with Mutual Information and Self-Information

We present a novel information-theoretic approach to introduce dependency among features of a deep convolutional neural network (CNN). The core idea of our proposed method, called MUSE, is to combine MUtual information and SElf-information to jointly improve the expressivity of all features extracted from different layers in a CNN. We present two variants of the realization of MUSE -- Additive Information and Multiplicative Information. Importantly, we argue and empirically demonstrate that MUSE, compared to other feature discrepancy functions, is a more functional proxy to introduce dependency and effectively improve the expressivity of all features in the knowledge distillation framework. MUSE achieves superior performance over a variety of popular architectures and feature discrepancy functions for self-distillation and online distillation, and performs competitively with the state-of-the-art methods for offline distillation. MUSE is also demonstrably versatile that enables it to be easily extended to CNN-based models on tasks other than image classification such as object detection.

cs.CV↗

arXiv2021

Unsupervised Few-Shot Action Recognition via Action-Appearance Aligned Meta-Adaptation

We present MetaUVFS as the first Unsupervised Meta-learning algorithm for Video Few-Shot action recognition. MetaUVFS leverages over 550K unlabeled videos to train a two-stream 2D and 3D CNN architecture via contrastive learning to capture the appearance-specific spatial and action-specific spatio-temporal video features respectively. MetaUVFS comprises a novel Action-Appearance Aligned Meta-adaptation (A3M) module that learns to focus on the action-oriented video features in relation to the appearance features via explicit few-shot episodic meta-learning over unsupervised hard-mined episodes. Our action-appearance alignment and explicit few-shot learner conditions the unsupervised training to mimic the downstream few-shot task, enabling MetaUVFS to significantly outperform all unsupervised methods on few-shot benchmarks. Moreover, unlike previous few-shot action recognition methods that are supervised, MetaUVFS needs neither base-class labels nor a supervised pretrained backbone. Thus, we need to train MetaUVFS just once to perform competitively or sometimes even outperform state-of-the-art supervised methods on popular HMDB51, UCF101, and Kinetics100 few-shot datasets.

cs.CV↗

arXiv2021

Self-supervised Outdoor Scene Relighting

Outdoor scene relighting is a challenging problem that requires good understanding of the scene geometry, illumination and albedo. Current techniques are completely supervised, requiring high quality synthetic renderings to train a solution. Such renderings are synthesized using priors learned from limited data. In contrast, we propose a self-supervised approach for relighting. Our approach is trained only on corpora of images collected from the internet without any user-supervision. This virtually endless source of training data allows training a general relighting solution. Our approach first decomposes an image into its albedo, geometry and illumination. A novel relighting is then produced by modifying the illumination parameters. Our solution capture shadow using a dedicated shadow prediction map, and does not rely on accurate geometry estimation. We evaluate our technique subjectively and objectively using a new dataset with ground-truth relighting. Results show the ability of our technique to produce photo-realistic and physically plausible results, that generalizes to unseen scenes.

cs.CV↗

arXiv2021

Revisiting Dynamic Convolution via Matrix Decomposition

Recent research in dynamic convolution shows substantial performance boost for efficient CNNs, due to the adaptive aggregation of K static convolution kernels. It has two limitations: (a) it increases the number of convolutional weights by K-times, and (b) the joint optimization of dynamic attention and static convolution kernels is challenging. In this paper, we revisit it from a new perspective of matrix decomposition and reveal the key issue is that dynamic convolution applies dynamic attention over channel groups after projecting into a higher dimensional latent space. To address this issue, we propose dynamic channel fusion to replace dynamic attention over channel groups. Dynamic channel fusion not only enables significant dimension reduction of the latent space, but also mitigates the joint optimization difficulty. As a result, our method is easier to train and requires significantly fewer parameters without sacrificing accuracy. Source code is at https://github.com/liyunsheng13/dcd.

cs.CV↗

arXiv2021

Outdoor inverse rendering from a single image using multiview self-supervision

In this paper we show how to perform scene-level inverse rendering to recover shape, reflectance and lighting from a single, uncontrolled image using a fully convolutional neural network. The network takes an RGB image as input, regresses albedo, shadow and normal maps from which we infer least squares optimal spherical harmonic lighting coefficients. Our network is trained using large uncontrolled multiview and timelapse image collections without ground truth. By incorporating a differentiable renderer, our network can learn from self-supervision. Since the problem is ill-posed we introduce additional supervision. Our key insight is to perform offline multiview stereo (MVS) on images containing rich illumination variation. From the MVS pose and depth maps, we can cross project between overlapping views such that Siamese training can be used to ensure consistent estimation of photometric invariants. MVS depth also provides direct coarse supervision for normal map estimation. We believe this is the first attempt to use MVS supervision for learning inverse rendering. In addition, we learn a statistical natural illumination prior. We evaluate performance on inverse rendering, normal map estimation and intrinsic image decomposition benchmarks.

cs.CV↗

arXiv2020

Beyond Adiabatic Elimination in Topological Floquet Engineering

In quantum mechanics, adiabatic elimination is a standard tool that produces a low-lying reduced Hamiltonian for a relevant subspace of states, incorporating effects of its coupling to states with much higher energy. Suppose this powerful elimination approach is applied to quasi-energy states in periodically-driven systems, a critical question then arises that the violation of the adiabatic condition caused by driven forces challenges such a presence of spectral reduction in the non-equilibrium driven system. Here, both theoretically and experimentally, we newly reported two kinds of driven-induced eliminations universal in topologically-protected Floquet systems. We named them "quasi-adiabatic elimination" and "high-frequency-limited elimination", in terms of different driven frequencies that deny the underlying requirement for the adiabatic condition. Both two non-adiabatic eliminations are observed in our recently developed microwave Floquet simulator, a programmable test platform composed of periodically-bending ultrathin metallic coupled corrugated waveguides. Through the near-field imaging on our simulator, the mechanisms between the adiabatic and driven-induced eliminations are revealed, indicating the ubiquitous spectral decomposition for tailoring and manipulating Floquet states with quasi-energies. Finally, we hope our findings may open up profound and applicable possibilities for further developing Floquet engineering in periodically-driven systems, ranging from condensed matter physics to photonics.

quant-ph↗

arXiv2020

SPRING: A Sparsity-Aware Reduced-Precision Monolithic 3D CNN Accelerator Architecture for Training and Inference

CNNs outperform traditional machine learning algorithms across a wide range of applications. However, their computational complexity makes it necessary to design efficient hardware accelerators. Most CNN accelerators focus on exploring dataflow styles that exploit computational parallelism. However, potential performance speedup from sparsity has not been adequately addressed. The computation and memory footprint of CNNs can be significantly reduced if sparsity is exploited in network evaluations. To take advantage of sparsity, some accelerator designs explore sparsity encoding and evaluation on CNN accelerators. However, sparsity encoding is just performed on activation or weight and only in inference. It has been shown that activation and weight also have high sparsity levels during training. Hence, sparsity-aware computation should also be considered in training. To further improve performance and energy efficiency, some accelerators evaluate CNNs with limited precision. However, this is limited to the inference since reduced precision sacrifices network accuracy if used in training. In addition, CNN evaluation is usually memory-intensive, especially in training. In this paper, we propose SPRING, a SParsity-aware Reduced-precision Monolithic 3D CNN accelerator for trainING and inference. SPRING supports both CNN training and inference. It uses a binary mask scheme to encode sparsities in activation and weight. It uses the stochastic rounding algorithm to train CNNs with reduced precision without accuracy loss. To alleviate the memory bottleneck in CNN evaluation, especially in training, SPRING uses an efficient monolithic 3D NVM interface to increase memory bandwidth. Compared to GTX 1080 Ti, SPRING achieves 15.6X, 4.2X and 66.0X improvements in performance, power reduction, and energy efficiency, respectively, for CNN training, and 15.5X, 4.5X and 69.1X improvements for inference.

cs.AR↗

arXiv2020

Software-defined Design Space Exploration for an Efficient DNN Accelerator Architecture

Deep neural networks (DNNs) have been shown to outperform conventional machine learning algorithms across a wide range of applications, e.g., image recognition, object detection, robotics, and natural language processing. However, the high computational complexity of DNNs often necessitates extremely fast and efficient hardware. The problem gets worse as the size of neural networks grows exponentially. As a result, customized hardware accelerators have been developed to accelerate DNN processing without sacrificing model accuracy. However, previous accelerator design studies have not fully considered the characteristics of the target applications, which may lead to sub-optimal architecture designs. On the other hand, new DNN models have been developed for better accuracy, but their compatibility with the underlying hardware accelerator is often overlooked. In this article, we propose an application-driven framework for architectural design space exploration of DNN accelerators. This framework is based on a hardware analytical model of individual DNN operations. It models the accelerator design task as a multi-dimensional optimization problem. We demonstrate that it can be efficaciously used in application-driven accelerator architecture design. Given a target DNN, the framework can generate efficient accelerator design solutions with optimized performance and area. Furthermore, we explore the opportunity to use the framework for accelerator configuration optimization under simultaneous diverse DNN applications. The framework is also capable of improving neural network models to best fit the underlying hardware resources.

cs.DC↗