Filters








801 Hits in 6.0 sec

Compression of Acoustic Event Detection Models with Low-rank Matrix Factorization and Quantization Training [article]

Bowen Shi, Ming Sun, Chieh-Chi Kao, Viktor Rozgic, Spyros Matsoukas, Chao Wang
2019 arXiv   pre-print
In this paper, we present a compression approach based on the combination of low-rank matrix factorization and quantization training, to reduce complexity for neural network based acoustic event detection  ...  For a three-layer long short-term memory (LSTM) based AED model, the original model size can be reduced to 1% with negligible loss of accuracy.  ...  After the low-rank matrix factorization is applied (equation 1), the model is quantized and fine-tuned with quantization training.  ... 
arXiv:1905.00855v1 fatcat:f3bwc7l54zfsnguht3cdpeukey

Compression of Acoustic Event Detection Models with Quantized Distillation

Bowen Shi, Ming Sun, Chieh-Chi Kao, Viktor Rozgic, Spyros Matsoukas, Chao Wang
2019 Interspeech 2019  
Acoustic Event Detection (AED), aiming at detecting categories of events based on audio signals, has found application in many intelligent systems.  ...  In this paper, we present a simple yet effective compression approach which jointly leverages knowledge distillation and quantization to compress larger network (teacher model) into compact network (student  ...  [21] combines quantization and low-rank matrix factorization technique to compress multi-layer recurrent neural network.  ... 
doi:10.21437/interspeech.2019-1747 dblp:conf/interspeech/ShiSKRMW19 fatcat:cppoyd7orfgenpcyvpogtbf4dq

Relevance-based quantization of scattering features for unsupervised mining of environmental audio

Vincent Lostanlen, Grégoire Lafay, Joakim Andén, Mathieu Lagrange
2018 EURASIP Journal on Audio, Speech, and Music Processing  
To decide which parts to inspect further, we need tools that automatically mine the data, identifying recurring patterns and isolated events.  ...  The scattering coefficients capture short-scale structure, while the cluster model captures longer time scales, allowing for more accurate characterization of sparse events.  ...  As the detection of events is still an open problem [22] , we consider in this paper a generic quantization scheme in order to identify and represent time intervals of the scene that are coherent, thus  ... 
doi:10.1186/s13636-018-0138-4 fatcat:euhpapchxra67abwua3aazijpq

Table of Contents

2021 IEEE Transactions on Signal Processing  
Hao Positive Semidefinite Matrix Factorization: A Connection With Phase Retrieval and Affine Rank Minimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ...  Saruwatari Low-Rank Matrix Recovery With Scaled Subgradient Methods: Fast and Robust Convergence Without the Condition Number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ... 
doi:10.1109/tsp.2021.3136798 fatcat:kzkdhzcz3fgx3jv6gfjofooseq

2020 Index IEEE Signal Processing Letters Vol. 27

2020 IEEE Signal Processing Letters  
., +, LSP 2020 1680-1684 RGB-T Saliency Detection via Low-Rank Tensor Learning and Unified Collaborative Ranking. Huang, L., +, LSP 2020 1585-1589 Robust CP Tensor Factorization With Skew Noise.  ...  ., +, LSP 2020 366-370 The Global Geometry of Centralized and Distributed Low-rank Matrix Recovery Without Regularization.  ... 
doi:10.1109/lsp.2021.3055468 fatcat:wfdtkv6fmngihjdqultujzv4by

2020 Index IEEE/ACM Transactions on Audio, Speech, and Language Processing Vol. 28

2020 IEEE/ACM Transactions on Audio Speech and Language Processing  
., +, TASLP 2020 461-473 Quantization (signal) Neural Network Language Model Compression With Product Quantization and Soft Binarization.  ...  ., +, TASLP 2020 2837-2847 Neural Network Language Model Compression With Product Quantization and Soft Binarization.  ...  T Target tracking Multi-Hypothesis Square-Root Cubature Kalman Particle Filter for Speaker Tracking in Noisy and Reverberant Environments. Zhang, Q., +, TASLP 2020 1183 -1197  ... 
doi:10.1109/taslp.2021.3055391 fatcat:7vmstynfqvaprgz6qy3ekinkt4

Application of Compressive Sensing Techniques in Distributed Sensor Networks: A Survey [article]

Thakshila Wimalajeewa, Pramod K. Varshney
2019 arXiv   pre-print
reconstruction in centralized as well in decentralized settings, (ii) solve a variety of inference problems such as detection, classification and parameter estimation, with compressed data without signal  ...  We start the discussion with a brief introduction to the theory of CS and then describe the motivational factors behind the potential use of CS in WSN applications.  ...  In particular, the matrix X can be assumed to be low rank [175] in the presence of spatio-temporal correlations which can be recovered reliably from a compressed version of it (or a subset of its entries  ... 
arXiv:1709.10401v2 fatcat:fnk6vwwykvc5radcgrl22g42hu

Table of Contents

2021 IEEE Transactions on Signal Processing  
Wang Positive Semidefinite Matrix Factorization: A Connection With Phase Retrieval and Affine Rank Minimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ...  Zhao Low-Rank Matrix Recovery With Scaled Subgradient Methods: Fast and Robust Convergence Without the Condition Number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ... 
doi:10.1109/tsp.2021.3136800 fatcat:zhf46mb3rbdlnnh3u2xizgxof4

Table of Contents

2020 IEEE Signal Processing Letters  
Yokota, and Q. Shi 810 On Recoverability of Randomly Compressed Tensors With Low CP Rank . . . . . . . . . . . . . . . . .S. Ibrahim, X. Fu, and X.  ...  Pyzer Knapp, and S. Maskell 1570 Clustering Event Streams With Low Rank Hawkes Processes . . . . . . . . . . . . . . A. C. Türkmen, G. Ç apan, and A. T.  ...  Bovik 2144 Low-Rank Regularized Deep Collaborative Matrix Factorization for Micro-Video Multi-Label Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ... 
doi:10.1109/lsp.2020.3040844 fatcat:xpovskhrvfgctk3hhufuvpyyne

Knowledge Distillation: A Survey [article]

Jianping Gou, Baosheng Yu, Stephen John Maybank, Dacheng Tao
2021 arXiv   pre-print
To this end, a variety of model compression and acceleration techniques have been developed.  ...  As a representative type of model compression and acceleration, knowledge distillation effectively learns a small student model from a large teacher model.  ...  For efficient acoustic event detection, a quantized distillation method is proposed by using both knowledge distillation and quantization .  ... 
arXiv:2006.05525v6 fatcat:aedzaeln5zf3jgjsgsn5kvjrri

Table of Contents

2020 IEEE Signal Processing Letters  
Dunn, and C. J. Rozell 1120 On Recoverability of Randomly Compressed Tensors With Low CP Rank . . . . . . . . . . . . . . . . .S. Ibrahim, X. Fu, and X.  ...  Yokota, and Q. Shi 810 Negative Binomial Matrix Factorization . J. C.  ...  Li 1550 On the Identifiability of Transform Learning for Non-Negative Matrix Factorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ... 
doi:10.1109/lsp.2020.3040840 fatcat:ezrfzwo6tjbkfhohq2tgec4m6y

TOWARDS DISTILLATION OF DEEP NEURAL NETWORKS FOR SATELLITE ON-BOARD IMAGE SEGMENTATION

F. de Vieilleville, A. Lagrange, R. Ruiloba, S. May
2020 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences  
This method is presented through a ship detection example comparing accuracy and inference costs for several networks.  ...  To use them, we must reduce the size of DNN while accommodating efficiency in terms of both accuracy and inference cost.  ...  to perform a significant part of the work presented in this paper.  ... 
doi:10.5194/isprs-archives-xliii-b2-2020-1553-2020 fatcat:fwlos5zfvzgtbmjw35xl37hcv4

A parametric prosody coding approach for Mandarin speech using a hierarchical prosodic model

Chen-Yu Chiang
2018 EURASIP Journal on Audio, Speech, and Music Processing  
It employs a hierarchical prosodic model (HPM) as a prosody-generating model in the encoder to analyze the speech prosody of the input utterance to obtain a parametric representation of four prosodic-acoustic  ...  An informal listening test confirmed that both converted speeches of high and low speaking rate sounded very smooth.  ...  Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.  ... 
doi:10.1186/s13636-018-0129-5 fatcat:223ynzawozdojbtuxxy24vgwqu

A discriminative CNN video representation for event detection

Zhongwen Xu, Yi Yang, Alexander G. Hauptmann
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
The integration of the two contributions results in a new state-of-the-art performance in event detection over the largest video datasets.  ...  The focus of this paper is to effectively leverage deep Convolutional Neural Networks (CNNs) to advance event detection, where only frame level static descriptors can be extracted by the existing CNN toolkits  ...  We gratefully acknowledge the support of NVIDIA Corporation with the donation of the GPUs used for this research.  ... 
doi:10.1109/cvpr.2015.7298789 dblp:conf/cvpr/XuYH15 fatcat:jmdbsqdlmbfedah3yl556mnzyy

A Discriminative CNN Video Representation for Event Detection [article]

Zhongwen Xu, Yi Yang, Alexander G. Hauptmann
2014 arXiv   pre-print
The integration of the two contributions results in a new state-of-the-art performance in event detection over the largest video datasets.  ...  The focus of this paper is to effectively leverage deep Convolutional Neural Networks (CNNs) to advance event detection, where only frame level static descriptors can be extracted by the existing CNN toolkit  ...  performance in event detection, as evidenced by the reports of the top ranked teams in the TRECVID Multimedia Event Detection (MED) competition [3, 21, 28, 29] and research papers [30, 39, 45] that  ... 
arXiv:1411.4006v1 fatcat:3fkymo5jsvc73orwtjfwflafzu
« Previous Showing results 1 — 15 out of 801 results