Proposes finding missing features in Lasso solutions.
problem Lasso overlooks features not selected in its optimal solution.
method Computes alternate features efficiently without redundant computations.
result Reasonable alternate features found in 20 newsgroup data.
Predictive modeling applications increasingly use data representing people's behavior, opinions, and interactions. Fine-grained behavior data often has different structure from traditional data, being very high-dimensional and sparse. Models built from these data are quite difficult to interpret, since they contain man…
AGF explains feature learning in neural networks through alternating steps.
problem Understanding what features neural networks learn and how they learn them.
method AGF is an algorithmic framework that approximates the dynamics of feature learning in two-layer networks.
result AGF provides a unified framework to understand feature learning in neural networks, matching experimental results across various architectures.
Introduces alternators for modeling sequences, outperforming baselines.
problem Modeling complex sequential data with stability and efficiency.
method Two neural networks (OTN and FTN) alternate between outputting samples in observation and feature spaces, learned via cross-entropy criterion.
result Alternators outperform strong baselines in various domains (Lorenz equations, Neuroscience, Climate Science).
New algorithm recovers feature matrix with strong correlations.
problem Uncertainty in recovering feature matrix with highly correlated weights.
method Provable alternating gradient descent algorithm.
result Algorithm provably recovers ground-truth feature matrix with strong correlations.
Derives FACT, an alternative to NFA for neural networks, explaining feature learning.
problem Understanding how neural networks learn representations.
method First-principles approach using first-order optimality conditions.
result FACT explains why NFA holds and provides a principled alternative.
Paper addresses incorrectness of nearest neighbor in ranking models.
problem Incorrectness of nearest neighbor in ranking models.
method Introducing new algorithms with features constructed from 'global' and 'local' information.
result New algorithms provide correct neighbor identification in ranking models.
Paper proposes an efficient algorithm for clustering with sparse feature selection.
problem Estimating labels and sparse weights in unsupervised clustering.
method Alternating minimization of Frobenius norm criterion with K-sparse algorithm.
result Significantly improves clustering results on single-cell RNA sequencing datasets.
PCMC-Net uses neural networks to estimate transition rates in choice models, improving accuracy over traditional methods.
problem Inference limitations of traditional PCMC models when examples are scarce or new alternatives are observed.
method Amortized inference approach embedding PCMC definition into a neural network.
result Neural network outperforms feature engineered and machine learning models in airline booking prediction.
Proposes a partitioned least squares model for feature grouping.
problem Modeling with feature groups to assess variable importance.
method Two methods: alternating least squares and exact reformulation.
result Exact method provides better results in less time.
New methods improve matching in high-dimensional data.
problem Matching in high-dimensional data is challenging due to the curse of dimensionality.
method Pseudo-metric learning approach using MLPs or siamese neural networks.
result Neural-net-based methods outperform traditional matching methods in simulations and real data.
Proposes LM3FE for multi-modal feature extraction in image classification.
problem High-dimensional features and multi-modal data challenges.
method Large margin multi-modal multi-task feature extraction (LM3FE) framework.
result LM3FE outperforms single-task feature extraction and multi-modal feature extraction.
Paper proposes an algorithm to recover non-negative matrix factorization with mild conditions.
problem Understanding and guaranteeing recovery of non-negative matrix factorization.
method Alternates between updating features and decoding weights using ReLU.
result Proves recovery of ground-truth under mild conditions, including linear independence of features.
Develops a method to continuously audit black-box conditional quantile forecasts.
problem Continuous monitoring of black-box forecasts under changing data streams and regimes.
method Distribution-free and game-theoretic testing framework for non-i.i.d. losses.
result Derives finite-time detection guarantees for miscalibrated forecasts based on features.
Deep learning directly extracts features for emotion recognition, outperforming traditional methods.
problem Emotion recognition using traditional feature encoding methods.
method Used deep learning networks to directly encode features for emotion recognition.
result Highest performance in emotion recognition on EmoDB dataset.
Study uses machine learning to detect sleep disorders by identifying brain patterns.
problem Detecting sleep disorders through EEG patterns in NREM sleep cycles.
method Feature engineering and machine learning model for predicting Cyclic Alternating Patterns (CAP).
result The model accurately predicts CAP sequences associated with sleep disorders.
We consider classification tasks in the regime of scarce labeled training data in high dimensional feature space, where specific expert knowledge is also available. We propose a new hybrid optimization algorithm that solves the elastic-net support vector machine (SVM) through an alternating direction method of multipli…
Rare features reduce analysis effectiveness; new method aggregates them.
problem Reducing analysis effectiveness due to rare features in high-dimensional data.
method Aggregating rare features into denser features using a tree-based similarity encoding.
result Our method achieves high accuracy in predicting hotel ratings from reviews.
A faster alternative to ICA methods using randomized features.
problem Efficiently recovering statistically independent signals from observations.
method Randomized Generalized Variance and Randomized Canonical Correlation.
result Optimization of randomized measures yields comparable separation error at a much faster rate.
Improves convex biclustering for high-dimensional data.
problem Discovering meaningful biclusters in high-dimensional data.
method Biconvex modification with adaptive feature weighting.
result Consistently recovers biclusters and selects features appropriately.
A new tensor-based layer reduces neural network dimensions without losing important features.
problem Reducing dimensionality in tensor-structured feature data for deep neural networks.
method TensorProjection layer that projects input tensors into output tensors with reduced dimensions through mode-wise projections.
result The TensorProjection layer outperforms traditional downsampling methods in tasks like medical image classification and segmentation.
A new feature selection method for cost-sensitive classification in Random Forests.
problem Feature selection in Random Forests is challenging due to the complexity of ensemble trees.
method Shallow Tree Selection method for feature selection from small tree structures, adapted for cost-sensitive learning.
result BCR criterion improves performance of feature selection methods.
New methods identify limits of testing in high-dimensional models with non-sparse structures.
problem Understanding statistical inference in high-dimensional models with non-sparse regression coefficients.
method Developed new concepts of uniform and essentially uniform non-testability.
result Identified new tradeoffs between testability and feature correlation, showing that minimax lower bounds can be attained by tests with n \sqrt{n} n power. Tree prototypes improve tree ensemble interpretability.
problem Making tree ensembles interpretable.
method Introducing prototypes, a new distance for GBTs, and adaptive selection methods.
result Prototypes can perform as well as or better than original tree ensembles when used as nearest-prototype classifiers.
The paper addresses interpretability issues in EBM models by improving feature selection and reducing spurious interactions.
problem Interpretability issues in EBM models, especially spurious interactions and single feature dominance.
method Alternate Cross-feature selection, ensemble features, and model configuration alteration techniques.
result Our approach improves interpretability and predictive performance of EBM models, reducing spurious interactions and single feature dominance.
New methods needed for accurate feature importance due to feature dependencies.
problem Misleading variable importance measures from PaP methods due to feature dependencies.
method Alternative approaches involving additional modeling to avoid extrapolation.
result PaP metrics can over-emphasize correlated features, requiring more direct methods.
This paper explores how NLP enhances insurance data analysis.
problem Traditional insurance data limitations and need for alternative data.
method Application of NLP techniques to transform and analyze unstructured text data.
result NLP techniques improve insurance data analysis and risk assessment.
TDA improves accuracy of machine learning models for repeated measurements.
problem Limited accuracy of machine learning models for repeated measurements.
method Samples from data space, builds network graph based on data topology.
result TDA classifier achieves high accuracy (up to 96.8%) in repeated measurement datasets.
New model learns hierarchical features from generative models.
problem Generative models struggle with hierarchical feature learning.
method Proposed an alternative architecture to overcome limitations.
result Model learns interpretable, disentangled hierarchical features.
Fisher score is one of the most widely used supervised feature selection methods. However, it selects each feature independently according to their scores under the Fisher criterion, which leads to a suboptimal subset of features. In this paper, we present a generalized Fisher score to jointly select features. It aims …
Paper proposes a novel unsupervised feature selection method using K-means and ADMM.
problem Finding a subset of features for high-dimensional unsupervised learning problems.
method Developed K-means Derived Unsupervised Feature Selection (K-means UFS) using ADMM to solve NP-hard optimization.
result K-means UFS outperforms baselines in feature selection for clustering.
We depart from the usual methods for pricing contracts with the counterparty credit risk found in most of the existing literature. In effect, typically, these models do not account for either systemic effects or at-first-default contagion and postulate that the contract value at default equals either the risk-free valu…
CDANs classify unordered feature sets efficiently and invariantly.
problem Classifying unordered feature sets with traditional neural networks leads to spurious patterns.
method Convolutional deep averaging networks (CDANs) for permutation-invariant classification.
result CDANs outperform linear embeddings and other methods in classifying unordered feature sets.
This research reverses feature visualization in neural networks to optimize for specific feature objectives.
problem The invertibility of feature visualization in neural networks is not well understood.
method The approach involves optimizing for the feature objective that generates the input used in feature visualization, using the gradient of a specific objective function.
result A closed-form solution is found to minimize the gradient, providing an alternative view on network sensitivity.
Method learns hierarchical representations of samples and features simultaneously.
problem Hierarchical structures in samples and features not considered by existing methods.
method Jointly learns hierarchical representations via Tree-Wasserstein Distance alternating between samples and features.
result Method improves performance in link prediction and node classification tasks.
FRI identifies relevant features in high-dimensional data for biomedical experiments.
problem Spurious feature selection in high-dimensional data.
method Feature relevance method for identifying all-relevant variables in linear classification and regression.
result FRI can identify causal features in biomedical experiments.
OOMP selects features online for sparse linear regression.
problem Feature selection in high-dimensional sparse linear models.
method Online algorithm that alternates between feature selection and coefficient estimation.
result Theoretical guarantees and computational complexity analysis of OOMP.
The problem of learning a sparse model is conceptually interpreted as the process of identifying active features/samples and then optimizing the model over them. Recently introduced safe screening allows us to identify a part of non-active features/samples. So far, safe screening has been individually studied either fo…
CIBP models feature abundance in latent feature models.
problem Modeling feature abundance in latent feature models.
method Proposes a new Bayesian nonparametric prior, the CIBP, for latent feature models.
result The expected number of features is bounded even as the number of objects increases.
Value selection reduces model size while maintaining accuracy.
problem Space efficiency in model size reduction.
method Two probabilistic methods based on information theory's metric: PVS and P + VS.
result Value selection achieves balance between accuracy and model size reduction.
Selecting important features in non-linear or kernel spaces is a difficult challenge in both classification and regression problems. When many of the features are irrelevant, kernel methods such as the support vector machine and kernel ridge regression can sometimes perform poorly. We propose weighting the features wit…
Paper proposes a method to solve sparse Bayesian learning problems efficiently.
problem Finding sparsest solutions in high-dimensional settings.
method Sparse Bayesian learning with screening test to identify and remove zero coefficients.
result The method accelerates the solution process for sparse Bayesian learning problems.
Enhances NMF for better time series recovery and prediction using side information.
problem Reconstruct and predict electricity consumption time series.
method Extends NMF with side information, proposes HALSX algorithm.
result Improved recovery and prediction performance validated on various datasets.
Alternative perspective on mean-field LIBOR market model, maintaining practicality and applicability.
problem Maintaining practicality and applicability of mean-field LIBOR market model.
method Embedding mean-field model in a classical setup, controlling term rate variances over large time horizons.
result Framework can be directly applied to model term rates from SOFR, ESTR, or other nearly risk-free overnight rates.
AI improves MSME credit scoring using bank statement data.
problem Lack of access to financing for MSMEs due to traditional credit scoring methods.
method Developed a cash flow-based pipeline using bank statement data for machine learning credit scoring.
result Bank statement features significantly improve credit scoring models, achieving AUROC of 0.806.
Investigates feature reduction in finance using Deep Learning models.
problem Improving trend prediction in finance with reduced feature sets.
method Examines Restricted Boltzmann Machines (RBM) and Auto-Encoders (AE) for feature reduction in finance.
result Architectural and input space characteristics influence prediction quality.
Study shows diverse data types improve SARS-COV-2 case surge predictions.
problem Improving pandemic case surge predictions using multimodal data.
method Investigated the effectiveness of biological, public health, and behavioral features.
result Diverse feature sets enhance prediction accuracy, varying by country and phase.
Solves kernel dimension reduction while making features interpretable.
problem Making kernel dimension reduction methods interpretable.
method Projects onto a subspace before kernel feature mapping, using ISM for optimization.
result Extends ISM's theoretical guarantees to a family of kernels, enabling broader applicability.