Výsledky vyhledávání - We propose a multi-layer variational autoencoder method

Hierarchical Residual Learning Based Vector Quantized Variational Autoencoder for Image Reconstruction and Generation

Autoři: Adiban M. Stefanov K. Siniscalchi S. M. a další

Témata: Settore ING-INF/05 - Sistemi Di Elaborazione Delle Informazioni, I.2, FOS: Computer and information sciences, Computer Science - Machine Learning, I.4, Computer Sciences, namely VQVAE and VQVAE-2. HR-VQVAE can also generate high-quality and diverse images that outper- form state-of-the-art generative models, Computer Vision and Pattern Recognition (cs.CV), the representations at each layer are hierarchically linked to those at previous layers. We evaluate our method on the tasks of image reconstruction and generation. Experimental results demonstrate that the discrete representations learned by HR-VQVAE enable the decoder to reconstruct high-quality images with less distortion than the baseline methods, Computer Science - Computer Vision and Pattern Recognition, 02 engineering and technology, each layer in HR-VQVAE learns a discrete representation of the residual from previous layers through a vector quantized encoder. Furthermore, We propose a multi-layer variational autoencoder method, we call HR-VQVAE, that learns hierarchical discrete representations of the data. By utilizing a novel objective function, each layer in HR-VQVAE learns a discrete representation of the residual from previous layers through a vector quantized encoder. Furthermore, the representations at each layer are hierarchically linked to those at previous layers. We evaluate our method on the tasks of image reconstruction and generation. Experimental results demonstrate that the discrete representations learned by HR-VQVAE enable the decoder to reconstruct high-quality images with less distortion than the baseline methods, namely VQVAE and VQVAE-2. HR-VQVAE can also generate high-quality and diverse images that outper- form state-of-the-art generative models, providing further verification of the efficiency of the learned representations. The hierarchical nature of HR-VQVAE i) reduces the decoing search time, making the method particularly suitable for high-load tasks and ii) allows to increase the codebook size without incurring the codebook collapse problem, that learns hierarchical discrete representations of the data. By utilizing a novel objective function, Machine Learning (cs.LG), 03 medical and health sciences, Datavetenskap (datalogi), 0302 clinical medicine, We propose a multi-layer variational autoencoder method, providing further verification of the efficiency of the learned representations. The hierarchical nature of HR-VQVAE i) reduces the decoing search time, making the method particularly suitable for high-load tasks and ii) allows to increase the codebook size without incurring the codebook collapse problem, 0202 electrical engineering, electronic engineering, information engineering, we call HR-VQVAE

Popis souboru: application/pdf

Přístupová URL adresa: http://arxiv.org/abs/2208.04554
https://hdl.handle.net/10447/637533
http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-324770

Building damage detection from multi-feature fusion of Sentinel-1/2 imagery using variational autoencoder and MLP-Mixer network: insights from the Jishishan earthquake

Autoři: Jing Wang Xiwei Fan Yuanmeng Qi a další

Zdroj: Big Earth Data, Pp 1-40 (2025)

Témata: Building damage detection, variational autoencoder, multi-layer perceptron, InSAR, earthquake, Geography. Anthropology. Recreation, Geology, QE1-996.5

Popis souboru: electronic resource

Relation: https://doaj.org/toc/2096-4471; https://doaj.org/toc/2574-5417

Přístupová URL adresa: https://doaj.org/article/6015177fd5ab4390b318f6a700079ad8

Small sample pipeline DR defect detection based on smooth variational autoencoder and enhanced detection head faster RCNN.

Autoři: Zhang, Ting You, Tianyang Liu, Zhaoying a další

Zdroj: Applied Intelligence; Jul2025, Vol. 55 Issue 10, p1-29, 29p

Témata: AUTOENCODERS, SAMPLE size (Statistics), CLASSIFICATION, CALIBRATION

The Wyner Variational Autoencoder for Unsupervised Multi-Layer Wireless Fingerprinting

Autoři: Huang, Teng-Hui Dahanayaka, Thilini Thilakarathna, Kanchana a další

Zdroj: GLOBECOM 2023 - 2023 IEEE Global Communications Conference. :820-825

Témata: FOS: Computer and information sciences, Computer Science - Machine Learning, Computer Science - Information Theory, Information Theory (cs.IT), 0202 electrical engineering, electronic engineering, information engineering, 02 engineering and technology, Machine Learning (cs.LG)

Přístupová URL adresa: http://arxiv.org/abs/2303.15860

Smoke and mirrors: A variational optuna autoencoder with reinforcement learning for intrusion detection in IoT networks.

Autoři: Asante, Isaac Osei Wu, Libing

Zdroj: Cluster Computing; 2025, Vol. 28 Issue 11, p1-20, 20p

Témata: INTRUSION detection systems (Computer security), INTERNET of things, LATENT variables, MACHINE learning, REINFORCEMENT learning, INTERNET security, COMPUTER network security

Multi-Task Trajectory Prediction Using a Vehicle-Lane Disentangled Conditional Variational Autoencoder.

Autoři: Chen, Haoyang Li, Na Shan, Hangguan a další

Zdroj: Sensors (14248220); Jul2025, Vol. 25 Issue 14, p4505, 22p

Témata: AUTONOMOUS vehicles, PREDICTION models, MACHINE learning, PROBABILISTIC generative models

Unsupervised Anomaly Detection in Time Series Data via Enhanced VAE-Transformer Framework.

Autoři: Zhang, Chunhao Xie, Bin Huo, Zhibin

Zdroj: Computers, Materials & Continua; 2025, Vol. 84 Issue 1, p843-860, 18p

Témata: LONG short-term memory, ANOMALY detection (Computer security), TIME series analysis, AUTOENCODERS, FEATURE extraction

Small-data image classification via drop-in variational autoencoder.

Autoři: Mahdian, Babak Nedbal, Radim

Zdroj: Signal, Image & Video Processing; Sep2025, Vol. 19 Issue 9, p1-9, 9p

Optimizing path reasoning scheme to improve conversational recommendation system.

Autoři: Chen, Xinyu Huang, Xianying Liu, Fengjin a další

Zdroj: Evolving Systems; Dec2025, Vol. 16 Issue 4, p1-18, 18p

An Effective Method for Detecting Unknown Types of Attacks Based on Log-Cosh Variational Autoencoder.

Autoři: Yu, Li Xu, Liuquan Jiang, Xuefeng

Zdroj: Applied Sciences (2076-3417); Nov2023, Vol. 13 Issue 22, p12492, 21p

Témata: FEATURE extraction, EXTRACTION techniques, DATA distribution, INTRUSION detection systems (Computer security)

Video Colorization Based on Variational Autoencoder.

Autoři: Zhang, Guangzi Hong, Xiaolin Liu, Yan a další

Zdroj: Electronics (2079-9292); Jun2024, Vol. 13 Issue 12, p2412, 19p

Témata: GRAYSCALE model, VIDEO processing, VIDEOS, VIDEO surveillance

MESM: integrating multi-source data for high-accuracy protein-protein interactions prediction through multimodal language models.

Autoři: Wang, Feng Chu, Jinming Shen, Liyan a další

Zdroj: BMC Biology; 8/11/2025, Vol. 23 Issue 1, p1-24, 24p

Témata: PROTEIN-protein interactions, DEEP learning, LANGUAGE models, GRAPH neural networks, BIOINFORMATICS, SACCHAROMYCES cerevisiae

Multimodal variational autoencoder for inverse problems in geophysics: application to a 1-D magnetotelluric problem.

Autoři: Rodriguez, Oscar Taylor, Jamie M Pardo, David

Zdroj: Geophysical Journal International; Dec2023, Vol. 235 Issue 3, p2598-2613, 16p

Témata: INVERSION (Geophysics), GEOPHYSICS, INVERSE problems, MULTIMODAL user interfaces, DEEP learning

Diverse Image Captioning via Conditional Variational Autoencoder and Dual Contrastive Learning.

Autoři: JING XU BING LIU YONG ZHOU a další

Zdroj: ACM Transactions on Multimedia Computing, Communications & Applications; Jan2024, Vol. 20 Issue 1, p1-16, 16p

Témata: SEQUENTIAL learning, LEARNING, ENTROPY

Model Predictive Control with Variational Autoencoders for Signal Temporal Logic Specifications.

Autoři: Im, Eunji Choi, Minji Cho, Kyunghoon

Zdroj: Sensors (14248220); Jul2024, Vol. 24 Issue 14, p4567, 17p

Témata: DIFFERENTIABLE dynamical systems, PREDICTION models, LOGIC, ITERATIVE learning control, SATISFACTION

Diffangle-Grasp: Dexterous Grasp Synthesis via Fine-Grained Contact Generation and Natural Pose Optimization.

Autoři: Ning, Meng Deng, Chong Zhan, Ziheng a další

Zdroj: Biomimetics (2313-7673); Aug2025, Vol. 10 Issue 8, p492, 20p

Témata: PREHENSION (Physiology), HUMANOID robots, LATENT variables, POSTURE, STOCHASTIC models

Disentangled Variational Autoencoder for Emotion Recognition in Conversations.

Autoři: Yang, Kailai Zhang, Tianlin Ananiadou, Sophia

Zdroj: IEEE Transactions on Affective Computing; Apr-Jun2024, Vol. 15 Issue 2, p508-518, 11p

Variational Autoencoders-Based Algorithm for Multi-Criteria Recommendation Systems.

Autoři: Fraihat, Salam Shambour, Qusai Al-Betar, Mohammed Azmi a další

Zdroj: Algorithms; Dec2024, Vol. 17 Issue 12, p561, 15p

Témata: NATURAL language processing, PATTERN recognition systems, RECOMMENDER systems, STANDARD deviations, IMAGE analysis

CVGAE: A Self-Supervised Generative Method for Gene Regulatory Network Inference Using Single-Cell RNA Sequencing Data.

Autoři: Liu, Wei Teng, Zhijie Li, Zejun a další

Zdroj: Interdisciplinary Sciences: Computational Life Sciences; Dec2024, Vol. 16 Issue 4, p990-1004, 15p

Témata: GRAPH neural networks, MULTILAYER perceptrons, GENE expression, RNA sequencing, GAUSSIAN distribution

Self-supervised variational autoencoder towards recommendation by nested contrastive learning.

Autoři: Wang, Jing Wu, Jun Jia, Caiyan a další

Zdroj: Applied Intelligence; Aug2023, Vol. 53 Issue 15, p18887-18897, 11p

Témata: DEEP learning, RECOMMENDER systems