SMAC optimizes wildfire suppression policies quickly and effectively.
problem Managing conflicts among stakeholders in wildfire suppression policies.
method Applied SMAC to optimize MDP policies in a high-fidelity simulation environment.
result SMAC finds good policies that stakeholders can understand and validate.
Improved model predicts wildfire spread on slopes.
problem Accurate prediction of wildfire spread on slopes.
method Combines Rothermel model, Huygens' principle, and advanced techniques.
result More precise model of wildfire propagation.
Deep learning model predicts wildfire spread in Australia.
problem Predicting the full distribution of wildfire spread in Australia.
method Graph convolutional neural networks and extended generalized Pareto distribution.
result Efficacy of the model demonstrated through hazard assessment.
Study identifies key drivers and spatio-temporal trends of extreme Mediterranean wildfires.
problem Understanding and predicting the impacts of climate change on wildfire activity.
method Statistical deep-learning model combining meteorological, land cover, and orographic data.
result Vapour-pressure deficit significantly affects wildfire occurrence, while air temperature and drought affect spread.
Paper compares GRU and LSTM for predicting wildfire spread direction.
problem Predicting wildfire spread direction with limited data.
method Comparison of Gated Recurrent Unit (GRU) and Long Short-Term Memory (LSTM) networks.
result GRU performs better for longer time series than LSTM.
Machine learning improves wildfire science and management, but requires expert knowledge.
problem Improving wildfire science and management through AI.
method Review of popular ML approaches and their application in six wildfire science domains.
result Opportunities exist for applying more advanced ML methods in wildfire science.
Modeling wildfire spread using Gielis superformula and Finsler spacetime.
problem Accurately modeling the short-time spread of wildfires.
method Using Gielis superformula and Finsler spacetime to determine firefronts.
result A concise and efficient expression of geodesic equations.
Paper proposes decision-theoretic approach to combat wildfires.
problem Complexity and uncertainty in wildfire management.
method Partially-observable Markov decision process and data-driven model.
result Forecasting model accurately models wildfire spread.
Paper proposes model to assess financial risk of grid-ignited wildfires.
problem Financial risk and solvency threats from grid-ignited wildfires.
method Integrated model to evaluate damage costs and risk levels.
result Identifies high-risk areas for preemptive actions.
Modeling wildfire aerosols using satellite data to predict solar radiation reduction.
problem Accurately estimate and predict AOD propagation from wildfires using multi-source satellite data.
method Physics-informed statistical modeling integrating multi-source satellite data with an advection-diffusion equation.
result The proposed approach accurately predicts AOD propagation and demonstrates model interpretability.
Machine learning helps predict smoke types for safer forest burns.
problem Determining which fuels to burn safely to reduce wildfire risk and minimize smoke.
method Machine learning techniques like spectral clustering and manifold learning.
result Interpretable representations and tools for differentiating smoke types.
Custom loss functions improve accuracy of wildfire rate of spread forecasts.
problem Improving accuracy of wildfire rate of spread forecasts.
method Examined custom loss functions in machine learning models of fuel moisture content.
result Custom loss functions improved accuracy of ROS forecasts by a small amount.
Study creates open-access wildfire dataset for Russia.
problem Data scarcity for comprehensive Eurasian wildfire research.
method Machine learning for exploratory data analysis and predictive modeling.
result Identified key environmental factors influencing fire behavior.
Deep model predicts wildfire data in incomplete regions.
problem Predicting distributions of wildfire frequency and size in missing data regions.
method Conditional Missing data Importance-Weighted Autoencoder (CMIWAE) model trained on incomplete data.
result Maximizes log-likelihood of observed wildfire information.
Improved real-time UAV terrain following with RVM-RLS filter.
problem Accurate real-time waypoints estimation under measurement noise in nonlinear, time-varying systems.
method Residual Variance Matching Recursive Least Squares (RVM-RLS) filter guided by RVME criterion.
result Improved waypoints estimation accuracy by approximately 88% compared to benchmarks.
A new model predicts wildfire spread with wind and slope effects.
problem Predicting wildfire spread with wind and slope effects.
method Geometric model based on Lorentz-Finsler framework, considering wind and slope.
result Infinitesimal wavefronts are no longer restricted to be elliptical, allowing for more accurate predictions.
Efficient algorithms for large Maxent models improve wildfire probability predictions.
problem Training large-scale, non-smooth Maxent models efficiently for big data.
method First-order optimization algorithms using Kullback-Leibler divergence.
result Our algorithms outperform state-of-the-art methods by one order of magnitude.
Pyrocast predicts pyrocumulonimbus clouds six hours before they form.
problem Predicting pyrocumulonimbus clouds to adapt to climate change.
method Database of 148 pyroCb events, Random Forests, CNNs, and Auto-Encoders.
result Best model predicts pyroCb with 90% accuracy.
New method uses neural networks to predict extreme wildfires, improving accuracy over traditional models.
problem Predicting extreme wildfires using complex, non-linear relationships.
method Partially-interpretable neural networks for extreme quantile regression.
result Significant improvement in predictive performance over traditional methods.
A new method suppresses echo more effectively with lower latency.
problem Echo cannot be fully removed by linear filters due to nonlinear relationships.
method Uses a modified Conv-TasNet with multiple streams of input signals.
result Efficacy validated in both single-talk and double-talk situations.
New analysis identifies key factors in wildfire-generated thunderstorms.
problem Understanding the causes of pyrocumulonimbus (pyroCb) storms.
method Invariant Causal Prediction, conditional independence test, greedy-ICP search algorithm.
result Identified seven causal predictors for pyroCb formation.
Contrastive learning struggles with class collapse and feature suppression, revealing bias towards simpler solutions.
problem Contrastive learning struggles with class collapse and feature suppression, especially in supervised and unsupervised settings.
method Unified theoretical framework to determine which features are learnt by CL, revealing bias towards simpler solutions.
result Bias towards simpler solutions is a key factor in class collapse and feature suppression.
Human cochlear models improve DNN noise suppression systems.
problem DNN-based noise suppression systems lack robustness to unseen noise conditions.
method Coupled cochlear models with DNNs to improve noise suppression.
result Cochlear models enhance DNN generalizability to various noise conditions.
Paper uses deep learning to suppress bones on chest X-rays.
problem Improving pathologies classification by suppressing bones on chest X-rays.
method Conditional Generative Adversarial Network (GAN) and Haar 2D wavelet decomposition.
result Achieves state-of-the-art performance on bone suppression.
Paper proposes a statistical model for detecting mu-suppression in EEG signals.
problem Detecting mu-suppression in motor imagery EEG signals.
method Proposes a statistical model based on the generalized extreme value distribution (GEV) and a linear classifier.
result Preliminary results show good classification accuracy in detecting mu-suppression and distinguishing EEG events.
Paper introduces VSSGD to balance deep learning training with variance suppression.
problem Unbalanced data leads to unbalanced training process in deep learning.
method Adaptively adjusts error variance and mean to balance training.
result VSSGD accelerates training, prevents overfitting, and improves learning from small samples.
MFMCi factors out state variables to make Monte Carlo simulation work in high-dimensional MDPs.
problem High-dimensional MDPs make Monte Carlo simulation impractical.
method Factoring out some state and action variables to enable Model-Free Monte Carlo.
result MFMCi enables Monte Carlo simulation in high-dimensional MDPs.
Contrastive learning properties studied, including feature suppression and hierarchical learning.
problem Feature suppression and hierarchical learning in contrastive learning.
method Generalized contrastive loss, instance-based contrastive learning, explicit and controllable competing features.
result Contrastive learning can suppress and prevent the learning of competing features.
Paper proposes methods to reduce adversarial vulnerability in neural networks.
problem Adversarial vulnerability of deep neural networks in safety-critical applications.
method Defining vulnerability, proposing Vulnerability Suppression (VS) loss, and a Bayesian feature pruning method.
result Improves adversarial robustness and clean example performance with minimal parameter reduction.
Geodesic sprays and frozen metrics defined for time-dependent Lagrange manifolds.
problem Modeling wildfire spread under varying conditions.
method Define systems of pre-extremals for energy functional, induce Hamilton orthogonal nets, and derive frozen metrics.
result Energy pre-extremals become Finsler geodesics of the frozen metric.
The paper proposes a new dictionary learning method for faster and more accurate image classification.
problem Efficient and accurate image classification with compact dictionaries.
method Cross-label suppression and group regularization to learn a discriminative dictionary.
result The proposed method achieves better classification accuracy and computational efficiency compared to existing methods.
Time-warping improves RNN transfer learning for diverse time scales.
problem Transfer learning for RNNs with varying time scales.
method Time-warping rescales time in LSTM models for better transfer.
result Time-warping maintains accuracy in transferring RNNs between different time scales.
DG separates successes and failures by gating updates with advantage and surprisal.
problem Negative learning from surprising data in distributed reinforcement learning.
method DG gates each update with the product of advantage and surprisal, suppressing failures and preserving successes.
result DG outperforms other methods in various challenging reinforcement learning tasks.
New method improves fairness in DP learning by preventing excessive gradient suppression.
problem Disparate impact on model predictions for minority groups in DP learning.
method Bounded adaptive clipping to prevent excessive gradient suppression.
result Improves worst-class accuracy by over 10 percentage points compared to existing methods.
New method improves autofocus in CBCT scans by 93%.
problem Improper geometry information leads to misplaced signals in CBCT.
method Learning-based motion estimation combined with CBCT consistency constraint.
result Average artifact suppression of 93% achieved.
Localized Multidirectional Correction improves non-refusal target-response behavior in foundation models.
problem Controlled post-training refusal suppression in routed MoE and hybrid-MoE foundation models.
method Introduce Localized Multidirectional Correction (LoMC), a support-gated intervention framework.
result Substantially improves non-refusal target-response behavior while maintaining general capability under a compact intervention footprint.
Bayesian method suppresses low-frequency pulses in audio recordings.
problem Suppressing long pulses caused by mechanical defects in audio recordings.
method Bayesian approach using Gaussian Process for pulse location, signal interpolation, and tail estimation.
result Perceptual results similar to previous methods, performs well on naturally degraded signals.
Improved market state classification for risk assessment.
problem Classifying financial market states for better risk assessment.
method Modified selection criteria for market states, clustering optimization, and visualization of correlation matrices.
result Statistically significant results in SP 500 and Nikkei 225 markets.
New model LMRC tackles class incremental learning without needing old classes.
problem Softmax suppression problem in class incremental learning.
method Label Mapping with Response Consolidation (LMRC) and multi-head neural network.
result LMRC achieves better performance than related methods in different scenarios.
Solves exploding and vanishing gradient problem in LSTMs.
problem Exploding and vanishing gradient problem in LSTM optimization.
method Introduces a simple stochastic algorithm (h-detach) to prevent suppression of gradient components through the cell state path in LSTM.
result Significant improvements in convergence speed, robustness, and generalization over vanilla LSTM training.
DDGM generates realistic ECG signals for clinical use.
problem Generating accurate ECG signals from noisy data.
method Bayesian ECG reconstruction using DDGM trained on healthy ECG data.
result DDGM successfully generates realistic ECG signals for clinical applications.
AI system narrows human decision options for better outcomes.
problem Improving human decision-making in sequential tasks.
method Developed a decision support system using a pre-trained AI agent to limit human action choices.
result Participants outperformed AI and solo play in a wildfire mitigation game.
Regularization methods can overregulate, suppressing causal features.
problem Mitigating shortcuts in models exploiting spurious correlations.
method Analysis of regularization methods and their effects on causal features.
result Regularization can overregulate, suppressing causal features.
New methods validate a hypothesis explaining how neural nets generalize well.
problem Why over-parameterized nets generalize well despite memorizing training data.
method Developed new algorithms to suppress weak gradient directions without per-example gradients.
result Validated a hypothesis about gradient directions and their role in generalization.
M2M tackles zero-shot structured noise suppression in images.
problem Structured noise with strong anisotropic correlations in real-world images.
method M2M introduces a novel sampling strategy that generates pseudo-independent sub-image pairs from a single noisy input, using directional interpolation and generalized median filtering.
result M2M consistently outperforms state-of-the-art zero-shot methods under correlated noise.
SNAP improves robust computation by emphasizing trustworthy items and downweighting outliers.
problem Improving robustness in computation, especially in high-dimensional settings.
method SNAP assigns weights based on mutual agreement, suppressing outlier contributions.
result SNAP ensures outliers contribute negligibly to computations, even in high-dimensional settings.
New method estimates GGLM parameters, overcoming non-convexity.
problem Estimating parameters in GGLM with dependencies.
method Monotone operator-based variational inequality method.
result Guarantees for parameter recovery in GLM and GGLM.
Robust STAP with coprime arrays reduces clutter using sparse modeling.
problem Limited performance due to training samples support in practical applications.
method Two-stage approach: 1) RD virtual snapshot, 2) RD sparse measurement modeling with OMP-like recovery.
result Robust to prior knowledge errors, good clutter suppression performance.