Framework for interpreting ML models to reveal properties of real-world phenomena.
problem Lack of direct interpretability in modern ML models hinders scientific understanding.
method Developed 'property descriptors' grounded in statistical learning theory.
result Property descriptors can reveal relevant properties of joint probability distributions.
New insights into ML models' accuracy and generalization for scientific problems.
problem Quantifying accuracy and generalization of ML models in scientific applications.
method Rigorous numerical analysis and theoretical bounds for linear differential equations.
result Different ML models can have opposing generalization behaviors, contrary to intuition.
The paper argues for prioritizing identifying structure over complex models for scientific discovery.
problem Underdetermination of mechanisms in high-dimensional data, leading to unreliable explanations.
method Proposes concrete standards for 'mechanistic ML' to avoid collapsing explanations.
result Large language models (LLMs) can collapse large equivalence classes of explanations, making it hard to distinguish between mechanisms.
New approach uses low-fidelity data to train ML models efficiently.
problem Training ML models with scarce high-fidelity data leads to high variance and poor generalization.
method Multifidelity linear regression using approximate control variates.
result Multifidelity training achieves similar accuracy with reduced high-fidelity data.
This paper provides a guide to feature importance methods for better scientific inference.
problem Limited understanding of data-generating process due to opaque ML model mechanisms.
method Comprehensive review and new proofs of global feature importance methods.
result Facilitates a thorough understanding and concrete recommendations for FI methods.
DLHub enables sharing and serving of scientific ML models.
problem Lack of specialized ML systems for scientific applications.
method Self-service model repository and scalable serving capabilities.
result DLHub offers better performance and more capabilities than existing systems.
New method for valid inference from ML-predicted data.
problem Invalid scientific conclusions from ML-predicted outcomes.
method Assumption-Lean and Data-Adaptive Post-Prediction Inference (PSPA)
result Valid and powerful inference based on ML-predicted data.
Novel framework for ML-assisted inference valid for any statistical task.
problem Limited validity of existing methods for post-prediction inference.
method Introduces PSPS framework for task-agnostic ML-assisted inference.
result Valid and efficient inference for arbitrary ML models.
The paper aims to improve reproducibility of machine learning experiments.
problem Lack of reproducibility in machine learning experiments.
method Investigates factors affecting reproducibility and proposes FAIR data practices.
result Preliminary results show the role of ProvBook in capturing and comparing provenance.
Machine learning enhances ocean studies by analyzing sparse, multi-scale data.
problem Sparse, multi-scale ocean data limits traditional methods.
method Machine learning techniques applied to ocean observations, theory, and numerical simulations.
result ML can advance theoretical oceanographic exploration and numerical simulations.
This paper analyzes machine learning workflows in climate modeling.
problem Challenges in integrating machine learning with climate modeling.
method Analysis of case studies focusing on design patterns and workflow structure.
result Synthesis of workflow design patterns across diverse projects in ML-enabled climate modeling.
Paper integrates ML with physics models for engineering and environmental challenges.
problem Complex science and engineering problems require new methodologies combining physics-based models and ML.
method Structured overview of integrating physics-based models with ML techniques.
result Taxonomy of existing techniques and potential research gaps identified.
Clarifies the various fairness definitions in ML.
problem Addressing fairness in ML with different definitions.
method Analyzes and clarifies the differences between fairness definitions.
result Provides a clearer understanding of fairness definitions in ML.
Diamond method controls FDR for trustworthy feature interaction discovery in ML models.
problem Limited interpretability of ML models due to black box nature.
method Diamond method integrates model-X knockoffs framework to control FDR for non-additive interactions.
result Diamond method ensures accurate discovery of feature interactions with FDR control.
A new method for designing accurate emulators using deep learning with interval calibration.
problem Designing accurate emulators for scientific processes with modern machine learning methods.
method Learn-by-Calibrating (LbC) approach based on interval calibration.
result Significant improvements in generalization error over widely-used loss functions.
CausalBench aims to advance causal learning research with a transparent platform.
problem Lack of unified benchmark datasets, algorithms, metrics, and evaluation interfaces for causal learning.
method Introduces CausalBench, a flexible benchmark framework for causal analysis and machine learning.
result Promotes scientific collaboration, reproducibility, and awareness in causal learning research.
Framework evaluates AI proposals for drug discovery, finds no LLM advantage.
problem No principled framework exists for evaluating AI-guided scientific selection under budget constraints.
method Formally verified metric (BSDS/DQS) penalizes false discoveries and excessive abstention.
result LLMs provide no marginal value over existing classifiers in drug discovery.
Unified taxonomy for ML uncertainty in physics, validated.
problem Uncertainty quantification in machine learning for physics.
method Unified taxonomy, principled validation tools.
result Illustrated validation tools with examples.
Machine learning (ML) has become a commodity in our every-day lives. We routinely ask ML empowered smartphones to suggest lovely food places or to guide us through a strange place. ML methods have also become standard tools in many fields of science and engineering. A plethora of ML applications transform human lives a…
A review of ML and DL for ecological data analysis.
problem Understanding the strengths and limitations of ML and DL in ecological research.
method Historical overview, algorithm families, differences, universal principles, and emerging trends.
result ML and DL excel in prediction tasks but are still debated for causal inference.
The preceding three decades have seen the emergence, rise, and proliferation of machine learning (ML). From half-recognised beginnings in perceptrons, neural nets, and decision trees, algorithms that extract correlations (that is, patterns) from a set of data points have broken free from their origin in computational c…
New methods use ML predictions to improve statistical inference.
problem Improving statistical inference using machine learning predictions.
method Prediction-Augmented Trees (PART, PAQ) for reliable statistical analysis.
result PART and PAQ outperform existing methods in various datasets.
PIML model improves hydrological predictions by blending physics and ML.
problem Hydrological models either lack predictive accuracy or fail to maintain physical consistency.
method Physics Informed Machine Learning (PIML) that integrates physics-based models and ML algorithms.
result PIML model outperforms both physics-based and ML models in predicting streamflow and evapotranspiration.
New ML methods improve physical system understanding by quantifying uncertainty across diverse regimes.
problem Capturing multi-regime physical systems with standard ML techniques.
method Coverage-oriented uncertainty quantification (UQ) methods.
result Coverage-oriented UQ models deliver physically consistent uncertainty estimates.
Machine learning's data-centric philosophy conflicts with natural sciences' standards.
problem Conflict between machine learning's ontology and epistemology and natural sciences' practices.
method Identifying and analyzing contexts where ML can be beneficial or harmful in natural sciences.
result ML can enhance trustworthiness in causal inference but introduces biases in emulation and labeling.
WISCA generates consensus explanations from conflicting model-agnostic interpretability methods.
problem Conflicting explanations from diverse interpretability algorithms.
method WISCA integrates class probability and normalized attributions to generate consistent explanations.
result WISCA consistently aligns with the most reliable individual method, improving explanation reliability.
Extracts StarCraft II tournament data for AI and ML studies.
problem Lack of accessible esports data for scientific use.
method Gathered and processed StarCraft II tournament replays using an API parser library.
result The largest publicly available StarCraft II esports dataset.
ML surrogates speed up Bayesian inverse problem solving.
problem Infer source location from noisy acoustic wave equation data.
method Use neural network as surrogate for PDE, apply MCMC to posterior.
result Accurately infers source location from noisy data.
Interpretable ML helps discover insights from big data.
problem Validating data-driven discoveries from complex datasets.
method Statistical and machine learning techniques for interpretable models.
result Challenges in validating data-driven discoveries remain.
Clarifies challenges in machine learning uncertainty quantification.
problem Inconsistent terminology and diverse technical requirements for trustworthy uncertainties.
method Examines estimation targets, uncertainty constructs, and problematic mappings.
result Advocates for alignment between intent and implementation in UQ.
Challenge to separate Earth's magnetic field from vehicle's magnetic field for accurate navigation.
problem Separate Earth's magnetic field from vehicle's magnetic field for accurate magnetic navigation.
method Use machine learning (ML) and integrate physics of magnetic navigation (SciML) to remove aircraft magnetic field from total magnetic field.
result A model can be constructed to effectively remove aircraft magnetic field from the dataset.
Brief history and challenges of interpretable machine learning.
problem Challenges in interpreting machine learning models, especially in scientific applications.
method Overview of state-of-the-art methods and discussion of challenges.
result Interpretable machine learning has a rich history but faces significant challenges.
Rubin LSST DESC uses AI/ML for dark energy research.
problem Challenges in uncertainty quantification and model robustness for AI/ML in DESC.
method Bayesian inference, physics-informed methods, validation frameworks, active learning.
result AI/ML methods are essential but require rigorous evaluation and governance.
CNNs learn about input data uncertainties, improving classification performance.
problem Errors in measurements are often neglected in ML, affecting classification performance.
method Constructed a model with known noise levels and used CNNs to learn about uncertainties.
result CNNs can incorporate information about input data uncertainties, improving classification performance.
Unified ML imputation framework for missing data.
problem Invalid inference with incomplete data and ML imputations.
method Stratified by missingness patterns, weights and aggregates complete case information.
result Valid statistical inference for general missingness patterns.
XAI methods fail to explain ML models reliably.
problem Current XAI methods fail to provide reliable explanations for ML models.
method Formally define problems and design methods accordingly.
result Diverse notions of explanation correctness and metrics needed.
The study defines backdoor detection in ML and proves its infeasibility.
problem Backdoor detection in machine learning systems.
method Formal statistical definition and analysis of feasibility.
result Backdoor detection is impossible except for very small alphabet sizes.
We present an unsupervised approach for discovering semantic representations of mathematical equations. Equations are challenging to analyze because each is unique, or nearly unique. Our method, which we call equation embeddings, finds good representations of equations by using the representations of their surrounding …
Transformative machine learning improves model accuracy and explainability with limited data.
problem Improving model accuracy and interpretability with limited data in scientific tasks.
method Transforming intrinsic data representations to extrinsic ones based on model predictions.
result Transformative machine learning significantly outperforms intrinsic representations in drug-design, gene expression prediction, and meta-learning.
Proposes SGShift to identify shifted features causing model performance degradation under concept shift.
problem Concept shift leading to miscalibration in ML models across domains.
method SGShift method for identifying sparse set of shifted features using feature selection and statistical tools.
result SGShift identifies shifted features more accurately than baseline methods, requires few samples in the shifted domain, and is robust to complex cases.
Overparameterized models generalize well despite fitting noisy data.
problem Understanding why overparameterized models generalize well despite fitting noisy data.
method Statistical signal processing perspective.
result Overparameterized models often outperform underparameterized models in test performance.
A new ML method speeds up PDE simulations without needing classical training.
problem Accelerating transient PDE simulations using machine learning.
method Online-learned preconditioners using a bandit algorithm.
result One-shot acceleration of PDE simulations.
Paper proposes MDER model to extract ML methods and datasets from papers.
problem Lack of focus on methods and datasets in literature analysis.
method Proposes MDER model for entity recognition of methods and datasets.
result Demonstrates preliminary performance of MDER model.
New method improves statistical inference using machine learning-imputed data.
problem Improving statistical inference with imputed data from machine learning.
method Two-phase sampling approach for Z-estimation with ML-imputed outcomes.
result Guaranteed efficiency matching or exceeding classical inference, regardless of prediction quality.
Deep learning models detect nanopore translocation events with high accuracy.
problem Manual parameter selection for nanopore signal analysis is prone to error.
method Developed a synthetic signal generator for training ML models.
result Deep learning models achieve over 99% true event detection.
This paper connects deep neural networks to game theory, revealing their congestion game properties.
problem Understanding the behavior of deep neural networks using game theory.
method Analyzing deep neural networks as congestion games and applying game theory results.
result Deep neural networks exhibit congestion game properties, linking them to game theory.
Study uses SAR data to estimate forest vegetation indices, improving monitoring of temperate forests.
problem Limitations of optical satellite data in monitoring forest ecosystems, especially due to atmospheric effects.
method Estimating four vegetation indices (LAI, FAPAR, EVI, NDVI) using multitemporal Sentinel-1 SAR and ancillary data.
result Accurate estimation of forest vegetation indices using SAR data, achieving high R2 and low MAE values. MLSys aims to bridge ML and systems research.
problem Designing ML systems for real-world deployment is challenging.
method Foster a new conference and research community.
result MLSys conference focuses on intersection of systems and ML.