# CloudLSTM: A Recurrent Neural Model for Spatiotemporal Point-cloud Stream Forecasting

arXiv: https://arxiv.org/abs/1907.12410 · HTML: https://mlsanity.com/paper/1907.12410/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Jul 2019.

## Built on by (most cited first, 14 of 14)

- [ImDiffusion: Imputed Diffusion Models for Multivariate Time Series Anomaly Detection](https://mlsanity.com/paper/2307.00754/) (2023)
- [LiDAR-based 4D Occupancy Completion and Forecasting](https://mlsanity.com/paper/2310.11239/) (2023)
- [Xpert: Empowering Incident Management with Query Recommendations via Large Language Models](https://mlsanity.com/paper/2312.11988/) (2023)
- [PCPNet: An Efficient and Semantic-Enhanced Transformer Network for Point Cloud Prediction](https://mlsanity.com/paper/2304.07773/) (2023)
- [Learning Scene Dynamics from Point Cloud Sequences](https://mlsanity.com/paper/2111.08755/) (2021)
- [Spatio-temporal Graph-RNN for Point Cloud Prediction](https://mlsanity.com/paper/2102.07482/) (2021)
- [QuickSkill: Novice Skill Estimation in Online Multiplayer Games](https://mlsanity.com/paper/2208.07704/) (2022)
- [Point Cloud Audio Processing](https://mlsanity.com/paper/2105.02469/) (2021)
- [CUPID: Improving Battle Fairness and Position Satisfaction in Online MOBA Games with a Re-matchmaking System](https://mlsanity.com/paper/2406.19720/) (2024)
- [HyPCV-Former: Hyperbolic Spatio-Temporal Transformer for 3D Point Cloud Video Anomaly Detection](https://mlsanity.com/paper/2508.00473/) (2025)
- [AGAR: Attention Graph-RNN for Adaptative Motion Prediction of Point Clouds of Deformable Objects](https://mlsanity.com/paper/2307.09936/) (2023)
- [Predicting Fatigue Crack Growth via Path Slicing and Re-Weighting](https://mlsanity.com/paper/2309.06708/) (2023)
- [packetLSTM: Dynamic LSTM Framework for Streaming Data with Varying Feature Space](https://mlsanity.com/paper/2410.17394/) (2024)
- [Martingale Neural Operators: Learning Stochastic Marginals via Doob-Meyer Factorization](https://mlsanity.com/paper/2605.15806/) (2026)

## Abstract (arXiv)

This paper introduces CloudLSTM, a new branch of recurrent neural models tailored to forecasting over data streams generated by geospatial point-cloud sources. We design a Dynamic Point-cloud Convolution (DConv) operator as the core component of CloudLSTMs, which performs convolution directly over point-clouds and extracts local spatial features from sets of neighboring points that surround different elements of the input. This operator maintains the permutation invariance of sequence-to-sequence learning frameworks, while representing neighboring correlations at each time step -- an important aspect in spatiotemporal predictive learning. The DConv operator resolves the grid-structural data requirements of existing spatiotemporal forecasting models and can be easily plugged into traditional LSTM architectures with sequence-to-sequence learning and attention mechanisms. We apply our proposed architecture to two representative, practical use cases that involve point-cloud streams, i.e., mobile service traffic forecasting and air quality indicator forecasting. Our results, obtained with real-world datasets collected in diverse scenarios for each use case, show that CloudLSTM delivers accurate long-term predictions, outperforming a variety of competitor neural network models.