检索结果-内蒙古大学图书馆

Combined depth space based architecture search for person re-identification

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Li, Hanjun Wu, Gaojie Zheng, Wei-Shi School of Computer Science and Engineering Sun Yat-sen University China Peng Cheng Laboratory Shenzhen518005 China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China Pazhou Lab Guangzhou China

Most works on person re-identification (ReID) take advantage of large backbone networks such as ResNet, which are designed for image classification instead of ReID, for feature extraction. However, these backbones may not be computationally efficient or the most suitable architectures for ReID. In this work, we aim to design a lightweight and suitable network for ReID. We propose a novel search space called Combined Depth Space (CDS), based on which we search for an efficient network architecture, which we call CDNet, via a differentiable architecture search algorithm. Through the use of the combined basic building blocks in CDS, CDNet tends to focus on combined pattern information that is typically found in images of pedestrians. We then propose a low-cost search strategy named the Top-k Sample Search strategy to make full use of the search space and avoid trapping in local optimal result. Furthermore, an effective Fine-grained Balance Neck (FBLNeck), which is removable at the inference time, is presented to balance the effects of triplet loss and softmax loss during the training process. Extensive experiments show that our CDNet (∼1.8 M parameters) has comparable performance with state-of-the-art lightweight networks. Copyright © 2021, The Authors. All rights reserved.

关键词： Network architecture

Syntax-enhanced Pre-trained model

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Xu, Zenan Guo, Daya Tang, Duyu Su, Qinliang Shou, Linjun Gong, Ming Zhong, Wanjun Quan, Xiaojun Jiang, Daxin Duan, Nan School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Microsoft Research Asia Beijing China Microsoft Search Technology Center Asia Beijing China Guangdong Key Laboratory Big Data Analysis and Processing Guangzhou China Key Lab. of Machine Intelligence and Advanced Computing Ministry of Education China

We study the problem of leveraging the syntactic structure of text to enhance pre-trained models such as BERT and RoBERTa. Existing methods utilize syntax of text either in the pre-training stage or in the fine-tuning stage, so that they suffer from discrepancy between the two stages. Such a problem would lead to the necessity of having human-annotated syntactic information, which limits the application of existing methods to broader scenarios. To address this, we present a model that utilizes the syntax of text in both pre-training and fine-tuning stages. Our model is based on Transformer with a syntax-aware attention layer that considers the dependency tree of the text. We further introduce a new pre-training task of predicting the syntactic distance among tokens in the dependency tree. We evaluate the model on three downstream tasks, including relation classification, entity typing, and question answering. Results show that our model achieves state-of-the-art performance on six public benchmark datasets. We have two major findings. First, we demonstrate that infusing automatically produced syntax of text improves pre-trained models. Second, global syntactic distances among tokens bring larger performance gains compared to local head relations between contiguous tokens. Copyright © 2020, The Authors. All rights reserved.

关键词： Syntactics

Coarse-to-Fine Latent Diffusion for Pose-Guided Person Image Synthesis

学校读者我要写书评

暂无评论

Coarse-to-Fine Latent Diffusion for Pose-Guided Person Image...

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Yanzuo Lu Manlin Zhang Andy J Ma Xiaohua Xie Jianhuang Lai School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Guangdong Province Key Laboratory of Information Security Technology China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China Pazhou Lab (HuangPu) Guangzhou China

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

Diffusion model is a promising approach to image generation and has been employed for Pose-Guided Person Image Synthesis (PGPIS) with competitive performance. While existing methods simply align the person appearance to the target pose, they are prone to overfitting due to the lack of a high-level semantic understanding on the source person image. In this paper, we propose a novel Coarse-to-Fine Latent Diffusion (CFLD) method for PGPIS. In the absence of image-caption pairs and textual prompts, we de-velop a novel training paradigm purely based on images to control the generation process of a pre-trained text-to-image diffusion model. A perception-refined decoder is designed to progressively refine a set of learnable queries and extract semantic understanding of person images as a coarse-grained prompt. This allows for the decoupling of fine-grained appearance and pose information controls at different stages, and thus circumventing the potential over-fitting problem. To generate more realistic texture details, a hybrid- granularity attention module is proposed to encode multi-scale fine-grained appearance features as bias terms to augment the coarse-grained prompt. Both quantitative and qualitative experimental results on the DeepFashion benchmark demonstrate the superiority of our method over the state of the arts for PGPIS. Code is availab.e at https://***/YanzuoLu/CFLD.

关键词： Training Image synthesis Semantics Text to image Process control Diffusion models Generators

Cross-camera feature prediction for intra-camera supervised person re-identification across distant scenes

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Ge, Wenhang Pan, Chunyan Wu, Ancong Zheng, Hongwei Zheng, Wei-Shi School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Pazhou Lab Guangzhou China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China Universtiy of Chinese Academy of Sciences Xinjiang China

Person re-identification (Re-ID) aims to match person images across non-overlapping camera views. The majority of Re-ID methods focus on small-scale surveillance systems in which each pedestrian is captured in different camera views of adjacent scenes. However, in large-scale surveillance systems that cover larger areas, it is required to track a pedestrian of interest across distant scenes (e.g., a criminal suspect escapes from one city to another). Since most pedestrians appear in limited local areas, it is difficult to collect training data with cross-camera pairs of the same person. In this work, we study intra-camera supervised person re-identification across distant scenes (ICS-DS Re-ID), which uses cross-camera unpaired data with intra-camera identity lab.ls for training. It is challenging as cross-camera paired data plays a crucial role for learning camera-invariant features in most existing Re-ID methods. To learn camera-invariant representation from cross-camera unpaired training data, we propose a cross-camera feature prediction method to mine cross-camera self supervision information from camera-specific feature distribution by transforming fake cross-camera positive feature pairs and minimize the distances of the fake pairs. Furthermore, we automatically localize and extract local-level feature by a transformer. Joint learning of global-level and local-level features forms a global-local cross-camera feature prediction scheme for mining fine-grained cross-camera self supervision information. Finally, cross-camera self supervision and intra-camera supervision are aggregated in a framework. The experiments are conducted in the ICS-DS setting on Market-SCT, Duke-SCT and MSMT17-SCT datasets. The evaluation results demonstrate the superiority of our method, which gains significant improvements of 15.4 Rank-1 and 22.3 mAP on Market-SCT as compared to the second best method. Our code is availab.e at https://***/g3956/CCFP. Copyright © 2021, The Authors.

关键词： machine learning

BTI Aging Monitoring based on SRAM Start-up Behavior

学校读者我要写书评

暂无评论

BTI Aging Monitoring based on SRAM Start-up Behavior

Asian Test Symposium (ATS)

作者： Shengyu Duan Peng Wang Gaole Sai School of Computer Engineering and Science Shanghai University Shanghai China State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences China State Key Laboratory of Mathematical Engineering and Advanced Computing Wuxi China Guangdong Provincial Key Lab of Robotics and Intelligent System Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China CAS Key Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institutes of Advanced Technology China

ISBN: (数字)9781728174679

ISBN: (纸本)9781728174686

Bias Temperature Instability (BTI) is one of the dominant CMOS aging mechanisms. It causes time-dependent variation, threatening circuit lifetime reliability. BTI-induced circuit errors are not detectable at the fabrication stage. On-line monitoring schemes are therefore necessary to capture the degradations during the operational time. Traditional aging monitoring techniques exhibit high implementation complexity and low stability. In this paper, we propose a BTI monitoring approach by simply tracking the start-up behavior of SRAM cells. SRAM is a widely used on-chip device in many applications. We study the impact of BTI for SRAM start-up values and age some cells in a manipulated manner. The BTI degradation is evaluated based on the number of SRAM cells starting with a certain value. This technique can be used to estimate the degradation for on-chip logic circuits without introducing additional circuitry, and thus has very low implementation complexity. We use an SRAM array with 1024 cells to estimate the degradations for multiple logic circuits, and show the average mean absolute percentage error as 8.48%. In addition, this technique is robust considering process, voltage and temperature variations.

关键词： Degradation Monitoring Logic circuits Transistors SRAM cells Aging System-on-chip

Gray Learning from Non-IID Data with Out-of-distribution Samples

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Zhao, Zhilin Cao, Longbing Wang, Chang-Dong The Data Science Lab School of Computing and DataX Research Centre Macquarie University SydneyNSW2109 Australia The School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Guangdong Province Key Laboratory of Computational Science Guangzhou China Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China

The integrity of training data, even when annotated by experts, is far from guaranteed, especially for non-IID datasets comprising both in- and out-of-distribution samples. In an ideal scenario, the majority of samples would be in-distribution, while samples that deviate semantically would be identified as out-of-distribution and excluded during the annotation process. However, experts may erroneously classify these out-of-distribution samples as in-distribution, assigning them lab.ls that are inherently unreliable. This mixture of unreliable lab.ls and varied data types makes the task of learning robust neural networks notably challenging. We observe that both in- and out-of-distribution samples can almost invariably be ruled out from belonging to certain classes, aside from those corresponding to unreliable ground-truth lab.ls. This opens the possibility of utilizing reliable complementary lab.ls that indicate the classes to which a sample does not belong. Guided by this insight, we introduce a novel approach, termed Gray Learning (GL), which leverages both ground-truth and complementary lab.ls. Crucially, GL adaptively adjusts the loss weights for these two lab.l types based on prediction confidence levels. By grounding our approach in statistical learning theory, we derive bounds for the generalization error, demonstrating that GL achieves tight constraints even in non-IID settings. Extensive experimental evaluations reveal that our method significantly outperforms alternative approaches grounded in robust statistics. Copyright © 2022, The Authors. All rights reserved.

关键词： machine learning

Exploring architectural ingredients of adversarially robust deep neural networks 21

学校读者我要写书评

暂无评论

Exploring architectural ingredients of adversarially robust ...

Proceedings of the 35th International Conference on Neural Information Processing Systems

作者： Hanxun Huang Yisen Wang Sarah Erfani Quanquan Gu James Bailey Xingjun Ma School of Computing and Information Systems The University of Melbourne Victoria Australia Key Lab. of Machine Perception School of Artificial Intelligence Peking University Beijing China and Institute for Artificial Intelligence Peking University Beijing China University of California Los Angeles School of Computer Science Fudan University Shanghai China

ISBN: (纸本)9781713845393

Deep neural networks (DNNs) are known to be vulnerable to adversarial attacks. A range of defense methods have been proposed to train adversarially robust DNNs, among which adversarial training has demonstrated promising results. However, despite preliminary understandings developed for adversarial training, it is still not clear, from the architectural perspective, what configurations can lead to more robust DNNs. In this paper, we address this gap via a comprehensive investigation on the impact of network width and depth on the robustness of adversarially trained DNNs. Specifically, we make the following key observations: 1) more parame- ters (higher model capacity) does not necessarily help adversarial robustness; 2) reducing capacity at the last stage (the last group of blocks) of the network can actually improve adversarial robustness; and 3) under the same parameter budget, there exists an optimal architectural configuration for adversarial robustness. We also provide a theoretical analysis explaning why such network configuration can help robustness. These architectural insights can help design adversarially robust DNNs.

关键词：

Multi-Stage Speaker Extraction with Utterance and Frame-Level Reference Signals

学校读者我要写书评

暂无评论

Multi-Stage Speaker Extraction with Utterance and Frame-Leve...

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Meng Ge Chenglin Xu Longbiao Wang Eng Siong Chng Jianwu Dang Haizhou Li Tianjin Key Laboratory of Cognitive Computing and Application College of Intelligence and Computing Tianjin University Tianjin China School of Computer Science and Engineering Nanyang Technological University Singapore National University of Singapore Singapore Japan Advanced Institute of Science and Technology Ishikawa Japan Machine Listening Lab University of Bremen Germany

ISBN: (纸本)9781728176055;9781728176062

Speaker extraction requires a sample speech from the target speaker as the reference. However, enrolling a speaker with a long speech is not practical. We propose a speaker extraction technique, that performs in multiple stages to take full advantage of short reference speech sample. The extracted speech in early stages is used as the reference speech for late stages. For the first time, we use frame-level sequential speech embedding as the reference for target speaker. This is a departure from the traditional utterance-based speaker embedding reference. In addition, a signal fusion scheme is proposed to combine the decoded signals in multiple scales with automatically learned weights. Experiments on WSJ0-2mix and its noisy versions (WHAM! and WHAMR!) show that SpEx++ consistently outperforms other state-of-the-art baselines.

关键词： Conferences Signal processing Acoustics Noise measurement Time-domain analysis Speech processing

MINI-Net: Multiple instance ranking network for video highlight detection

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Hong, Fa-Ting Huang, Xuanteng Li, Wei-Hong Zheng, Wei-Shi School of Data and Computer Science Sun Yat-sen University China Peng Cheng Laboratory Shenzhen518005 China VICO Group University of Edinburgh United Kingdom Pazhou Lab Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education China

We address the weakly supervised video highlight detection problem for learning to detect segments that are more attractive in training videos given their video event lab.l but without expensive supervision of manually annotating highlight segments. While manually averting localizing highlight segments, weakly supervised modeling is challenging, as a video in our daily life could contain highlight segments with multiple event types, e.g., skiing and surfing. In this work, we propose casting weakly supervised video highlight detection modeling for a given specific event as a multiple instance ranking network (MINI-Net) learning. We consider each video as a bag of segments, and therefore, the proposed MINI-Net learns to enforce a higher highlight score for a positive bag that contains highlight segments of a specific event than those for negative bags that are irrelevant. In particular, we form a max-max ranking loss to acquire a reliable relative comparison between the most likely positive segment instance and the hardest negative segment instance. With this max-max ranking loss, our MINI-Net effectively leverages all segment information to acquire a more distinct video feature representation for localizing the highlight segments of a specific event in a video. The extensive experimental results on three challenging public benchmarks clearly validate the efficacy of our multiple instance ranking approach for solving the problem. Copyright © 2020, The Authors. All rights reserved.

关键词： machine learning