Article
Engineering, Electrical & Electronic
Ren Yang, Fabian Mentzer, Luc Van Gool, Radu Timofte
Summary: This paper introduces a Recurrent Learned Video Compression (RLVC) approach with Recurrent Auto-Encoder (RAE) and Recurrent Probability Model (RPM) to better utilize the temporal correlation among video frames, achieving state-of-the-art compression performance.
IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING
(2021)
Article
Computer Science, Information Systems
Ziyang Fu, Chen Huang, Li Zhang, Shihui Wang, Yan Zhang
Summary: This study developed a deep learning model for sleep EEG signal using bidirectional recurrent neural network (BiRNN). The model achieved automatic discrimination of sleep staging of EEG signals through denoising, feature extraction, and time-series information mining, with high recognition rate and stability.
Article
Computer Science, Information Systems
Shridhar B. Devamane, Rajeshwari L. Itagi
Summary: This paper examines the performance of neural Turbo decoder and deep learning-based Turbo decoder, comparing them with the conventional convolutional Viterbi decoder. The study is conducted on different structures and analyzes their performance under different input data lengths and code rates.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Engineering, Electrical & Electronic
Rongrong Fu, Zeyi Wang, Shiwei Wang, Xuechen Xu, Junxiang Chen, Guilin Wen
Summary: In this study, a compact DL decoder combining EEGNet decoder with ProbSparse multihead self-attention mechanism was proposed to enhance the performance of EEG-based intent decoding. Experimental results showed that this decoder outperformed other methods on different datasets.
IEEE SENSORS JOURNAL
(2023)
Article
Environmental Sciences
Zhen Qin, Anyue Jiang, Dave Faulder, Trenton T. Cladouhos, Behnam Jafarpour
Summary: This paper introduces an optimization framework for improving the net power generation of geothermal reservoirs using a data-driven predictive model based on recurrent neural network (RNN). The developed model shows higher efficiency in prediction and optimization compared to traditional numerical simulation models.
WATER RESOURCES RESEARCH
(2023)
Article
Chemistry, Analytical
Muhammad Mostafa Monowar, Md Abdul Hamid, Abu Quwsar Ohi, Madini O. Alassafi, M. F. Mridha
Summary: Image retrieval techniques are gaining popularity due to the availability of multimedia data. This paper introduces AutoRet, a self-supervised image retrieval system based on deep convolutional neural networks (DCNN). The system is trained on pairwise constraints and can work with partially labeled datasets. Benchmarking results show that the proposed method performs well in a self-supervised manner and can handle mixed availability of labeled data.
Article
Multidisciplinary Sciences
Chao Chen, Weiyu Guo, Zheng Wang, Yongkui Yang, Zhuoyu Wu, Guannan Li
Summary: This paper proposes a low-overhead, energy-aware runtime manager for processing RNN tasks in edge cloud computing. By dynamically assigning tasks to edge and cloud computing systems based on QoS requirements and optimizing energy on edge systems using DVFS techniques, experimental results show significant reduction in energy consumption compared to existing methods.
Article
Biology
Elodie Germani, Elisa Fromont, Camille Maumet
Summary: This study examines the advantages of using a large public neuroimaging database in a self-taught learning framework to enhance brain decoding for new tasks. The researchers show that training a convolutional autoencoder and using a pretrained model to classify unknown tasks improves the performance of the classifiers.
Article
Engineering, Electrical & Electronic
Guocai Nan, Zhengkuan Wang, Chenghua Wang, Bi Wu, Zhican Wang, Weiqiang Liu, Fabrizio Lombardi
Summary: This work introduces a hybrid-iterative compression algorithm for LSTM/GRU and proposes an energy-efficient accelerator for bidirectional RNNs. By grouping gating units and using different compression algorithms, significant reduction in storage and computation requirements can be achieved without compromising accuracy. Improvements in the data flow of matrix operation unit and BRAM utilization, along with a timing matching strategy, address the load-imbalance issue and result in enhanced energy efficiency compared to state-of-the-art designs.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS
(2021)
Article
Computer Science, Artificial Intelligence
Xiaoyin Nie, Gang Xie
Summary: The paper introduces a framework called normalized recurrent neural network (NRNN) for noisy label fault diagnosis, which improves the training process with normalized long short-term memory and handles noisy labels with forward crossentropy loss. The effectiveness and superiority of the framework are verified with four datasets of varying noisy label proportions. Additionally, the layer-wise relevance propagation algorithm is used to explore decision-making within the framework, revealing that NRNN does not treat samples equally and prefers signal peaks for classification decisions.
JOURNAL OF INTELLIGENT MANUFACTURING
(2021)
Article
Engineering, Geological
Xuzhen He, Fang Wang, Wengui Li, Daichao Sheng
Summary: The study introduces the use of deep learning to train models for improving computational efficiency in stochastic analysis. Training models with a large dataset allows for accurate results for new data without the need for re-training. The research shows that deep learning models have a competitive edge in complex problems and can extend their capabilities by generating more data and re-training.
Article
Computer Science, Information Systems
Pritom Biswas Udas, Md. Ebtidaul Karim, Kowshik Sankar Roy
Summary: In this paper, a network anomaly detection model called SPIDER is proposed, which combines four updated versions of recurrent neural networks and utilizes principal component analysis for dimension reduction to improve detection performance. Compared to different models, the proposed model shows significant improvement in detecting intrusions.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Computer Science, Interdisciplinary Applications
Huajie Wen, Jian Zhao, Shaohua Xiang, Lin Lin, Chengjian Liu, Tao Wang, Lin An, Lixin Liang, Bingding Huang
Summary: This paper proposes a novel lesion-localization convolution transformer method that improves the classification and localization of ophthalmic diseases in retina OCT images. It combines convolution and self-attention and achieves significant performance improvements.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2022)
Article
Neurosciences
Noriya Watanabe, Kosuke Miyoshi, Koji Jimura, Daisuke Shimane, Ruedeerat Keerativittayayut, Kiyoshi Nakahara, Masaki Takeda
Summary: Perception and categorization of objects in visual scenes can provide a better understanding of the surrounding situation. This study explored the spatial and temporal organization of visual object representations by using concurrent fMRI and EEG data, combined with deep neural networks. The results demonstrate the ability of multimodal deep learning to efficiently classify visual objects and sub-categories, and reveal the mechanisms of object categorization in brain-wide regions.
Article
Engineering, Electrical & Electronic
Sangyeob Kim, Soyeon Kim, Seongyon Hong, Sangjin Kim, Donghyeon Han, Jiwon Choi, Hoi-Jun Yoo
Summary: The article proposes a complementary deep-neural-network processor that combines CNN and SNN, enabling both inference and training. The processor achieves energy efficiency through workload division and integrates various modules for optimal performance. It achieves state-of-the-art results for image classification tasks.
IEEE JOURNAL OF SOLID-STATE CIRCUITS
(2023)
Article
Biochemistry & Molecular Biology
Ben Steemers, Alejandro Vicente-Grabovetsky, Caswell Barry, Peter Smulders, Tobias Navarro Schroder, Neil Burgess, Christian F. Doeller
Article
Neurosciences
H. Freyja Olafsdottir, Francis Carpenter, Caswell Barry
NATURE NEUROSCIENCE
(2016)
Editorial Material
Behavioral Sciences
Francis Carpenter, Caswell Barry
TRENDS IN COGNITIVE SCIENCES
(2016)
Article
Neurosciences
H. Freyja Olafsdottir, Francis Carpenter, Caswell Barry
Article
Multidisciplinary Sciences
Francis Carpenter, Neil Burgess, Caswell Barry
SCIENTIFIC REPORTS
(2017)
Article
Cell Biology
Lorenza Magno, Caswell Barry, Christoph Schmidt-Hieber, Polyvios Theodotou, Michael Hausser, Nicoletta Kessaris
Review
Biochemistry & Molecular Biology
H. Freyja Olafsdottir, Daniel Bush, Caswell Barry
Article
Multidisciplinary Sciences
Andrea Banino, Caswell Barry, Benigno Uria, Charles Blundell, Timothy Lillicrap, Piotr Mirowski, Alexander Pritzel, Martin J. Chadwick, Thomas Degris, Joseph Modayil, Greg Wayne, Hubert Soyer, Fabio Viola, Brian Zhang, Ross Goroshin, Neil Rabinowitz, Razvan Pascanu, Charlie Beattie, Stig Petersen, Amir Sadik, Stephen Gaffney, Helen King, Koray Kavukcuoglu, Demis Hassabis, Raia Hadsell, Dharshan Kumaran
Article
Neurosciences
Giulio Casali, Sarah Shipley, Charlie Dowell, Robin Hayman, Caswell Barry
FRONTIERS IN CELLULAR NEUROSCIENCE
(2019)
Article
Psychology, Biological
Jacob L. S. Bellmund, William De Cothi, Tom A. Ruiter, Matthias Nau, Caswell Barry, Christian F. Doeller
NATURE HUMAN BEHAVIOUR
(2020)
Article
Neurosciences
William de Cothi, Caswell Barry
Article
Biochemistry & Molecular Biology
Daniel Bush, H. Freyja Olafsdottir, Caswell Barry, Neil Burgess
Summary: This study reveals the use of phase coding during neuronal replay, similar to the theta phase precession observed during navigation. Each replay event involves the forward propagation of decoded locations along the recapitulated trajectory.
Article
Biology
Tom M. George, William de Cothi, Kimberly L. Stachenfeld, Caswell Barry
Summary: The predictive map hypothesis suggests that each place cell in the hippocampus encodes the expected future occupancy of its target location. Although it is unclear how these successor representations are learned in the brain, a model using spike-timing dependent plasticity (STDP) and theta sweeps shows promise in rapidly learning a close approximation to the successor representation. This biologically plausible model explains various observed phenomena related to successor representations in place cells and provides insight into the topographical ordering of place field sizes.
Article
Biochemistry & Molecular Biology
James C. R. Whittington, Timothy H. Muller, Shirley Mark, Guifen Chen, Caswell Barry, Neil Burgess, Timothy E. J. Behrens