Filters








250 Hits in 3.7 sec

MST: Masked Self-Supervised Transformer for Visual Representation [article]

Zhaowen Li, Zhiyang Chen, Fan Yang, Wei Li, Yousong Zhu, Chaoyang Zhao, Rui Deng, Liwei Wu, Rui Zhao, Ming Tang, Jinqiao Wang
2021 arXiv   pre-print
However, it has not been fully explored in visual self-supervised learning.  ...  Transformer has been widely used for self-supervised pre-training in Natural Language Processing (NLP) and achieved great success.  ...  Appendix: Masked Self-Supervised Transformer for Visual Representation A The setting of computation resources In ablation studies, the MST with 1024 images is trained in 128 AMD DCUs that are publicly  ... 
arXiv:2106.05656v2 fatcat:gnlxfm5a7veupgq4oex5s3ph3i

A State-of-the-Art Survey on Deep Learning Theory and Architectures

Md Zahangir Alom, Tarek M. Taha, Chris Yakopcic, Stefan Westberg, Paheding Sidike, Mst Shamima Nasrin, Mahmudul Hasan, Brian C. Van Essen, Abdul A. S. Awwal, Vijayan K. Asari
2019 Electronics  
Different methods have been proposed based on different categories of learning, including supervised, semi-supervised, and un-supervised learning.  ...  We also included recently developed frameworks, SDKs, and benchmark datasets that are used for implementing and evaluating deep learning approaches.  ...  tasks, competitive with contemporary approaches to un-supervised and self-supervised feature learning [227] .  ... 
doi:10.3390/electronics8030292 fatcat:2i64q7g6kjbjvfalvzwgiggnyq

The History Began from AlexNet: A Comprehensive Survey on Deep Learning Approaches [article]

Md Zahangir Alom, Tarek M. Taha, Christopher Yakopcic, Stefan Westberg, Paheding Sidike, Mst Shamima Nasrin, Brian C Van Esesn, Abdul A S. Awwal, Vijayan K. Asari
2018 arXiv   pre-print
There are different methods have been proposed on different category of learning approaches, which includes supervised, semi-supervised and un-supervised learning.  ...  We have also comprised recently developed frameworks, SDKs, and benchmark datasets that are used for implementing and evaluating deep learning approaches.  ...  Doctoral research scientist on deep Learning, computer vision for remote sensing and hyper spectral imaging (e-mail: pehedings@slu.edu). Brian C Van Esesn 3 and Abdul A S.  ... 
arXiv:1803.01164v2 fatcat:eo353y77tvckbdjcfexpaadeh4

Coarse-to-Fine Sparse Transformer for Hyperspectral Image Reconstruction [article]

Jing Lin, Yuanhao Cai, Xiaowan Hu, Haoqian Wang, Xin Yuan, Yulun Zhang, Radu Timofte, Luc Van Gool
2022 arXiv   pre-print
In this paper, we propose a novel Transformer-based method, coarse-to-fine sparse Transformer (CST), firstly embedding HSI sparsity into deep learning for HSI reconstruction.  ...  Then the selected patches are fed into our customized spectra-aggregation hashing multi-head self-attention (SAH-MSA) for fine pixel clustering and self-similarity capturing.  ...  Please zoom in for better visualization.  ... 
arXiv:2203.04845v1 fatcat:jibizartlncxze2r4q27wbli3a

Neural Execution Engines: Learning to Execute Subroutines [article]

Yujun Yan, Kevin Swersky, Danai Koutra, Parthasarathy Ranganathan, Milad Hashemi
2020 arXiv   pre-print
First, we observe that transformer-based sequence-to-sequence models can learn subroutines like sorting a list of numbers, but their performance rapidly degrades as the length of lists grows beyond those  ...  Second, to generalize to unseen data, we show that encoding numbers with a binary representation leads to embeddings with rich structure once trained on downstream tasks like addition or multiplication  ...  Acknowledgments and Disclosure of Funding We thank Danny Tarlow and the anonymous NeurIPS reviewers for their insightful feedback.  ... 
arXiv:2006.08084v3 fatcat:pqpwd3e2ubcfjan4vxpskv54i4

Tree Energy Loss: Towards Sparsely Annotated Semantic Segmentation [article]

Zhiyuan Liang, Tiancai Wang, Xiangyu Zhang, Jian Sun, Jianbing Shen
2022 arXiv   pre-print
By sequentially applying these affinities to the network prediction, soft pseudo labels for unlabeled pixels are generated in a coarse-to-fine manner, achieving dynamic online self-training.  ...  In this paper, we propose a novel tree energy loss for SASS by providing semantic guidance for unlabeled pixels.  ...  Visualizations of block-supervised settings We synthesize the block-wise annotations for ADE20k [43] and Cityscapes [6] datasets.  ... 
arXiv:2203.10739v2 fatcat:mmh6z5d62fg75e6jzlifpeieom

MILES: Visual BERT Pre-training with Injected Language Semantics for Video-text Retrieval [article]

Yuying Ge, Yixiao Ge, Xihui Liu, Alex Jinpeng Wang, Jianping Wu, Ying Shan, Xiaohu Qie, Ping Luo
2022 arXiv   pre-print
In this work, we for the first time investigate masked visual modeling in video-text pre-training with the "dual-encoder" architecture.  ...  We perform Masked visual modeling with Injected LanguagE Semantics (MILES) by employing an extra snapshot video encoder as an evolving "tokenizer" to produce reconstruction targets for masked video patch  ...  Recent works introduce masked visual modeling (MVM) to image self-supervised pre-training, where MVM masks a proportion of the visual patches and optimizes the vision Transformers to reconstruct the missing  ... 
arXiv:2204.12408v1 fatcat:hdl63xqcnvb6bj6zjvvv2rqdty

TS-CAM: Token Semantic Coupled Attention Map for Weakly Supervised Object Localization [article]

Wei Gao, Fang Wan, Xingjia Pan, Zhiliang Peng, Qi Tian, Zhenjun Han, Bolei Zhou, Qixiang Ye
2021 arXiv   pre-print
We introduce the token semantic coupled attention map (TS-CAM) to take full advantage of the self-attention mechanism in visual transformer for long-range dependency extraction.  ...  TS-CAM first splits an image into a sequence of patch tokens for spatial embedding, which produce attention maps of long-range visual dependency to avoid partial activation.  ...  Visual transformer [9] constructs a sequence of tokens by splitting an input image into patches with positional embedding and applying cascaded transformer blocks to extract visual representation.  ... 
arXiv:2103.14862v5 fatcat:rupkwqdn6jfsxm3ustipo7npr4

Neural Fields in Visual Computing and Beyond [article]

Yiheng Xie, Towaki Takikawa, Shunsuke Saito, Or Litany, Shiqin Yan, Numair Khan, Federico Tombari, James Tompkin, Vincent Sitzmann, Srinath Sridhar
2022 arXiv   pre-print
In Part I, we focus on techniques in neural fields by identifying common components of neural field methods, including different representations, architectures, forward mapping, and generalization methods  ...  In Part II, we focus on applications of neural fields to different problems in visual computing, and beyond (e.g., robotics, audio).  ...  We would like to thank Sunny Li for their help in designing the website, and Alexander Rush and Hendrik Strobelt of the Mini-Conf project.  ... 
arXiv:2111.11426v4 fatcat:yteqzbu6gvgdzobnfzuqohix2e

Geometric Regularization of Local Activations for Knowledge Transfer in Convolutional Neural Networks

Ilias Theodorakopoulos, Foteini Fotopoulou, George Economou
2021 Information  
In this work, we propose a mechanism for knowledge transfer between Convolutional Neural Networks via the geometric regularization of local features produced by the activations of convolutional layers.  ...  Furthermore, recent advances in self-supervised [52] learning have revealed great potential for regularization methods to be used in new tasks, beyond the typical knowledge transfer.  ...  In this intersection, transfer learning can be formulated as the quest for an appropriate transformation for the representations learned over the source domain, so that they match the distribution and  ... 
doi:10.3390/info12080333 fatcat:tx257qkcrjbf5kocqzzbybkrm4

A Geometric Dictionary Learning Based Approach for Fluorescence Spectroscopy Image Fusion

Zhiqin Zhu, Guanqiu Qi, Yi Chai, Penghua Li
2017 Applied Sciences  
The fused images of sparse-representation-based image fusion methods show great performance. Constructing an informative dictionary is a key step for sparsity-based image fusion method.  ...  In order to ensure sufficient number of useful bases for sparse representation in the process of informative dictionary construction, image patches from all source images are classified into different  ...  Multi-scale transform (MST) and wavelet based algorithms are the most conventional transform approaches applied to transform-domain-based image fusion [14, 15] .  ... 
doi:10.3390/app7020161 fatcat:3cf7f2vcmjcrveo4mjqsuh5cwq

Everything at Once – Multi-modal Fusion Transformer for Video Retrieval [article]

Nina Shvetsova, Brian Chen, Andrew Rouditchenko, Samuel Thomas, Brian Kingsbury, Rogerio Feris, David Harwath, James Glass, Hilde Kuehne
2021 arXiv   pre-print
into a joined multi-modal representation to obtain an embedding that aggregates multi-modal temporal information.  ...  Moreover, the implicit properties of the transformer allow to process inputs of different lengths.  ...  Only ∼ 50% videos mention an object or an action that is visually seen in the video clip [30] . We use this dataset to train our model in a self-supervised way as described in Sec. 3.3.  ... 
arXiv:2112.04446v1 fatcat:uniekznmmzc3noenb3nonpe3fu

The Embodied Brain of SOVEREIGN2: From Space-Variant Conscious Percepts During Visual Search and Navigation to Learning Invariant Object Categories and Cognitive-Emotional Plans for Acquiring Valued Goals

Stephen Grossberg
2019 Frontiers in Computational Neuroscience  
Noisy and incomplete visual sensor data are transformed into representations of visual form and motion. Invariant recognition categories are learned in the What stream.  ...  Controllers for both animals and animats need reactive mechanisms for exploration, and learned plans to efficiently reach goal objects once an environment becomes familiar.  ...  Noisy and incomplete visual sensor data are transformed into representations of visual form and motion. Invariant recognition categories are learned in the What stream.  ... 
doi:10.3389/fncom.2019.00036 pmid:31333437 pmcid:PMC6620614 fatcat:6igudrojjrarjeldiw6on4pwxy

Outpatient Psychotherapy for Borderline Personality Disorder: Randomized Trial of Schema-Focused Therapy vs Transference-Focused Psychotherapy

J.C. Markowitz
2007 Yearbook of Psychiatry and Applied Mental Health  
Survival analyses revealed a higher dropout risk for TFP patients than for SFT patients (P=.01).  ...  Patient assessments were made before randomization and then every 3 months for 3 years. Results: Data on 44 SFT patients and 42 TFP patients were available.  ...  We thank Jeffrey Young, PhD (SFT), and Frank Yeomans, MD, PhD (TFP), for training and supervising the therapists.  ... 
doi:10.1016/s0084-3970(08)70368-2 fatcat:qucsvgaop5bp7ld7gb2x7o4igu

Outpatient Psychotherapy for Borderline Personality Disorder

Josephine Giesen-Bloo, Richard van Dyck, Philip Spinhoven, Willem van Tilburg, Carmen Dirksen, Thea van Asselt, Ismay Kremers, Marjon Nadort, Arnoud Arntz
2006 Archives of General Psychiatry  
Survival analyses revealed a higher dropout risk for TFP patients than for SFT patients (P=.01).  ...  Patient assessments were made before randomization and then every 3 months for 3 years. Results: Data on 44 SFT patients and 42 TFP patients were available.  ...  We thank Jeffrey Young, PhD (SFT), and Frank Yeomans, MD, PhD (TFP), for training and supervising the therapists.  ... 
doi:10.1001/archpsyc.63.6.649 pmid:16754838 fatcat:npzxbx55yvbjnhgvvriq65otbm
« Previous Showing results 1 — 15 out of 250 results