Papers for

weather forecast teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Digital typhoon dataset improves satellite storm intensity benchmarks

Digital Typhoon Dataset V3: Benchmarking for Multi-spectral Extension and Analyzing Systematic Bias in Intensity Regression

Abstract: This paper presents the Digital Typhoon Dataset V3, a new version of the longest typhoon satellite image dataset spanning nearly 50 years, designed to benchmark machine learning models on long-term spatio-temporal data. The latest addition to Dataset V3 is multispectral satellite data, enabling multi-channel models. We investigate data-splitting strategies and present benchmarking results for the Northern and Southern Hemispheres. Experimental results show that, under the current dataset construction and experimental setting, multi-channel inputs do not clearly outperform single-channel inputs for pressure regression. Among the evaluated model architectures, image types, and spectral channels, ResNet-18, full-resolution images, and Channel 1 generally provide the best single-channel performance for pressure regression. Performance also differs between the WP and AU datasets by regression target, with lower RMSE on WP for pressure regression and on AU for wind regression. These results highlight the importance of considering dataset-specific and task-specific characteristics when interpreting benchmark performance. Additionally, we analyzed systematic bias in intensity regression, which leads to underestimation of intensity for strong typhoons. We tested imbalance-aware weighting strategies, including inverse-density weighting inspired by BalancedMSE, and showed that these approaches are not sufficient to remove the bias. The dataset is publicly available at http://agora.ex.nii.ac.jp/digital-typhoon/dataset/ and https://github.com/kitamoto-lab/digital-typhoon/.

Thu 15 OctComputer Vision and Pattern Recognition

The gist

Typhoons have been tracked by satellites for almost 50 years, creating a huge amount of image data. The authors updated this dataset to include more types of satellite images, hoping to help machine learning models better predict typhoon intensity. They found that using more image channels didn’t clearly improve accuracy for estimating storm pressure. Also, some models tend to underestimate the strength of very strong typhoons, and their attempts to fix this bias didn’t fully work. This dataset is now publicly available for others to use.

Open → 2610.21654v1

Task-aware memory improves compact models for spatiotemporal prediction

TAM: Task-Aware Memory Distillation for Efficient Spatiotemporal Prediction

Abstract: Knowledge distillation enables efficient spatiotemporal prediction by transferring knowledge from an accurate teacher to a compact student. However, matching outputs or features independently for each sample leaves cross-sample predictive structure underused. Exploiting this structure requires representations and historical references that reflect the dynamics of each task. We propose TAM, a Task-Aware Memory Distillation framework that organizes a frozen teacher's knowledge into a bounded, retrievable history. Memory entries encode latent features, forecast changes, or flow residuals, while task-specific selection rules identify relevant historical references. The student either matches the teacher's similarity distribution over shared references or regresses observation-conditioned residual prototypes. These objectives complement supervised prediction and conventional distillation. The teacher, memory, and auxiliary adapters are used only during training, leaving student inference unchanged. We evaluate TAM on video prediction, weather forecasting, and traffic flow prediction across multiple teacher-student configurations. Averaged over four paired runs, adding TAM improves SSIM on all six video datasets and reduces MSE on five relative to the corresponding KD baselines. Mean paired MSE reductions reach 1.86% on KittiCaltech, 1.93% on WeatherBench with a gSTA teacher, and 1.01% on TaxiBJ. These results demonstrate the utility of historical teacher supervision across distinct forecasting tasks without additional student inference cost.

Thu 8 OctComputer Vision and Pattern RecognitionArtificial Intelligence

The gist

Spatiotemporal prediction means forecasting how things change over time and space, like weather or traffic flow. The authors show that teaching a smaller model by letting it remember patterns the bigger model noticed before helps improve its accuracy. They organize the teacher’s knowledge into a memory of past examples that the smaller model uses during training but not when making predictions. This method, TAM, works better than older techniques on different tasks without making the smaller model slower.

Open → 2610.11617v1

Bayesian data assimilation framework combines models and noisy observations

A Unified Framework for Bayesian Data Assimilation with Generative Models and Observation Interpolants

Abstract: Bayesian data assimilation combines model forecasts with noisy observations, but sampling high-dimensional, non-Gaussian posteriors remains challenging. We introduce an observation-interpolant framework that turns pretrained stochastic interpolant, flow matching, and diffusion models into posterior samplers without retraining. Conditioning the interpolant path on observations yields a shared likelihood-score correction to the drift or velocity, unifying stochastic and deterministic posterior sampling. The resulting SDEs and ODEs sample the exact posterior when the intermediate likelihood score is known. For practical computation, we approximate this score using a closed-form Gaussian surrogate with a bias-corrected mean and covariance inflated by the model's source covariance. Jacobian-free and ensemble-shared approximations make the method tractable in high dimensions. We evaluate the framework on linear-Gaussian dynamics, stochastic two-dimensional Navier-Stokes, and urban airflow with up to $O(10^4)$ degrees of freedom.

Fri 2 OctMachine LearningComputational Engineering, Finance, and Science

The gist

Combining mathematical models with noisy sensor data to make better predictions is important but often hard, especially when dealing with complex systems. The authors present a new method that uses pretrained machine learning models to efficiently sample from the updated predictions without needing additional training. This approach unifies different techniques for refining predictions and offers practical ways to handle high-dimensional data. They tested their method on examples including fluid dynamics and air flow in cities, showing it works well even with thousands of variables.

Open → 2610.03396v1