Convolutional LSTM

LSTM: Long Short-Term Memory

Convolutional LSTM (ConvLSTM) combines convolutional layers, which capture local spatial structure, with recurrent memory to model how images or grids change over time. It is used for tasks such as rainfall and temperature forecasting, satellite-image change detection, activity recognition, anomaly detection, and simulation of physical systems. Common approaches include stacked or encoder–decoder ConvLSTMs, multi-resolution ensembles, and hybrids that pair convolutional or latent-space models with recurrent dynamics; evaluations often compare them with persistence, fully connected LSTMs, and other forecasting baselines. A central open tension is whether modeling spatial structure improves results enough to justify added complexity: performance depends on the data and task, and ConvLSTM does not consistently beat simpler alternatives, especially for longer forecasts or when temporal inputs are poorly chosen.

Latest papers

14 papers3 papers in the last four weeks, with none the four weeks before. 0.0% of all new papers.

Motion Equivariant Convolutional LSTM: When Motion Becomes a Gauge Symmetry
Oct 20, 2026Vahid Hosseinzadeh, T. Anderson Keller, Michael PuthawalaEquivariant Neural Networks Convolutional LSTM
Department of Mathematics & Statistics, South Dakota State University · The Kempner Institute at Harvard University · Cusp AI
A Pre-trained Variational Autoencoder for Gyrokinetic Plasma Turbulence Surrogate Modeling
Sep 29, 2026Minglei Yang, Marshall Nicholson, Diego Del-Castillo-Negrete +2Variational Autoencoders Physics-Informed Generative Modeling
Oak Ridge National Laboratory, Oak Ridge, TN, 37831, USA · Grand Valley State University, Allendale, MI, 49401, USA · University of Texas at Austin, Austin, TX, 78712, USA
Temporal Modelling for Burn Scars on Sentinel-3
Sep 28, 2026Luca Barco, Edoardo Arnaudo, Andrea Bragagnolo +2Remote Sensing Image Segmentation Disaster Damage Assessment
Politecnico di Torino, Torino, Italy · Fondazione LINKS, Torino, Italy
Benchmarking ConvLSTM for One-Day-Ahead IMDAA Rainfall-Field Prediction across Four Indian Cities
Jul 29, 2026Tanmay Ghosh, Shaurabh Anand, Rakesh Gomaji Nannewar +1Convolutional RNNs Precipitation Forecasting
National Institute of Advanced Studies, Indian Institute of Science Campus, Bengaluru, India. · School of Development, Azim Premji University, Bengaluru, India. · Complex Systems Programme, National Institute of Advanced Studies, Indian Institute of Science Campus, Bengaluru, India.
An Improved CNN-LSTM Based Intrusion Detection System for IoT Networks
Jun 4, 2026Mohammad Tariq Ikhlas, Pohanyar Khowaja Khil, Malik Muhammad Mueed Aslam +1Network Intrusion Detection Internet of Things
SEECS NUST, Pakistan
Hybrid CNN-LSTM Framework for Intelligent Cyber Attack Detection and Prevention in U.S. Critical Digital Infrastructure: A Comparative Machine Learning Evaluation on CSE-CIC-IDS2018
Jun 4, 2026Md. Iqbal Hossan, Md. Serajul Kabir Chowdhury Rubel, Md. Arifur Rahman +1Autonomous Cyber Defense Network Intrusion Detection
Department of Computer Science, Maharishi International University, Fairfield, Iowa, USA. · Department of Information Studies, Trine University, Angola, Indiana, USA. · Department of Business Information Systems, Central Michigan University, Mount Pleasant, Michigan, USA.
Self-Supervised ConvLSTM for Fermi Large Area Telescope Transient Detection
May 21, 2026Alberto Garinei, Stefano Speziali, Alessandro Vispa +11Convolutional LSTM High-Energy Physics
Idea-Re S.r.l., 06125, Perugia, Italy · bDipartimento Scienze Ingegneristiche, Università Guglielmo Marconi, 00193, Rome, Italy · cIstituto Nazionale Fisica Nucleare, sezione di Perugia, 06123, Perugia, Italy +1
GPU-Accelerated Deep Learning for Heatwave Prediction and Urban Heat Risk Assessment
May 14, 2026Adis AlihodžićConvolutional LSTM Spatiotemporal Forecasting
Department of Mathematical and Computer Sciences, Faculty of Science, University of Sarajevo, Sarajevo, Bosnia and Herzegovina.
Spatiotemporal downscaling and nowcasting of urban land surface temperatures with deep neural networks
May 13, 2026Solomiia Kurchaba, Angela MeyerU-Net Convolutional LSTM
Department of Geoscience and Remote Sensing · School of Engineering and Computer Science, Bern University of Applied Sciences
Multilevel neural networks with dual-stage feature fusion for human activity recognition
Apr 17, 2026Abeer FathAllah Brery, Ascensión Gallardo-Antolín, Israel Gonzalez-Carrasco +1Human Activity Recognition Recurrent Neural Networks
L-UNet: An LSTM Network for Remote Sensing Image Change Detection
Mar 24, 2026Shuting Sun, Lin Mu, Lizhe Wang +1Remote Sensing Change Detection U-Net
College of Marine Science and Technology, China University of Geosciences (CUG), Wuhan 430074, China · College of Life Sciences and Oceanography, Shenzhen University, Shenzhen 518060, China · Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou), Guangzhou 511458, China +2
Spatio-Temporal Forecasting of Retaining Wall Deformation: Mitigating Error Accumulation via Multi-Resolution ConvLSTM Stacking Ensemble
Mar 11, 2026Jihoon Kim, Heejung YounSpatiotemporal Forecasting Long-Term Time Series Forecasting
Department of Civil and Environmental Engineering, Hongik University, Seoul, Republic of Korea · Department of Civil and Environmental Engineering, Hongik University, 94, Wausan-ro, Mapo-gu, Seoul, 04066, Republic of Korea
Weather-Related Crash Risk Forecasting: A Deep Learning Approach for Heterogenous Spatiotemporal Data
Mar 4, 2026Abimbola Ogungbire, Srinivas PulugurthaTraffic Accident Anticipation Spatiotemporal Forecasting
Traffic Engineering, Gresham Smith, Alpharetta, 30009, GA - USA. · Civil Engineering, University of Mississippi, Oxford, 38677, MS - USA.
A Comparative Study of Convolutional and Recurrent Neural Networks for Storm Surge Prediction in Tampa Bay
Aug 11, 2024Mandana Farhang Ghahfarokhi, Anahita Farhang Ghahfarokhi, Mahta Zamanizadeh +1Recurrent Neural Networks Convolutional Neural Networks
Department of Engineering Management and System Engineering, Old Dominion University, Norfolk, VA · Department of Civil and Environmental Engineering, Old Dominion University, Norfolk, VA

Oct 20, 2026J/K move · Enter open · S save

Motion Equivariant Convolutional LSTM: When Motion Becomes a Gauge Symmetry

Vahid Hosseinzadeh, T. Anderson Keller, Michael Puthawala

Department of Mathematics & Statistics, South Dakota State University · The Kempner Institute at Harvard University · Cusp AI

Equivariant Neural Networks Convolutional LSTM Video Prediction

Open paper AskPDF arXiv

Objects move, cameras pan, and weather drifts with the wind: much of what changes in sequential data is motion. Motion changes where things are, not what they are, and is therefore a symmetry of the data. Yet standard recurrent networks are not equivariant to it: their memory stays put while the input moves on. Flow equivariant recurrent networks first brought this symmetry into sequence models, for motions of constant velocity, or flows, by enumerating a bank of velocities. We extend this idea in three directions: from simple recurrent networks to convolutional LSTMs, from flows to motions whose velocity changes in time, and from enumerating velocities to estimating them. Treating a sequence as a field over time, we show that motion becomes a gauge symmetry: a global symmetry made local in time. Built on this view, our motion equivariant convolutional LSTM (MELSTM) uses its velocity model as the gauge field and is provably equivariant to every motion. Across future-frame prediction, classification of digits visible only through their motion, and action recognition under camera shake, MELSTM outperforms a convolutional LSTM and its flow equivariant counterpart with the same number of parameters: it is more accurate, generalizes to longer horizons and to unseen motions, and carries a fraction of the flow equivariant model's state.