BC learning improves deep sound recognition performance.
problem Improving deep sound recognition using novel training data.
method BC learning: mixing sounds from different classes to generate between-class sounds and train models to recognize these.
result BC learning improves performance on various sound recognition networks, surpassing human level.
BC learning improves image classification by mixing images from different classes.
problem Improving image classification accuracy.
method Generates mixed images from different classes and trains models to output the mixing ratio.
result Significant improvement in image classification performance (19.4% and 2.26% top-1 errors on ImageNet-1K and CIFAR-10, respectively).
ClassSim measures similarity between classes using misclassification ratios of trained classifiers.
problem Evaluating similarities between similar classes in real-world datasets.
method ClassSim metric based on misclassification ratios of trained DNNs.
result ClassSim provides better similarities than existing methods for image recognition.
A new neural network learns from acoustic scenes by suppressing irrelevant patterns.
problem Acoustic scenes are rich and redundant, making classification challenging.
method Spatio-temporal attention pooling layer coupled with a convolutional recurrent neural network.
result The method outperforms a strong convolutional neural network baseline and sets new state-of-the-art performance.
Fisher loss improves deep domain adaptation by learning discriminative within-class compact and between-class separable representations.
problem Improving deep domain adaptation performance by learning discriminative representations.
method Proposes a Fisher loss to learn discriminative representations that are within-class compact and between-class separable.
result Noticeable improvements in deep domain adaptation performance, e.g., 6.67% absolute improvement in mean accuracy on the Office-Home dataset.
SWRLDA improves LDA for multi-class classification with edge classes.
problem LDA's vulnerability to edge classes causing biased mean and large distances.
method Self-weighted robust LDA with l21-norm distance criterion.
result SWRLDA outperforms other methods on synthetic and real-world datasets.
CP-GAN generates images selectively conditioned on class specificity, capturing between-class relationships.
problem Generating images selectively conditioned on class specificity in class-overlapping data.
method Proposed Classifier's Posterior GAN (CP-GAN) that redesigns generator input and objective function for class-overlapping data.
result Demonstrated effectiveness of CP-GAN using both controlled and real-world class-overlapping data.
A new data augmentation method selects mixed classes based on class distances for better performance.
problem Improving recognition accuracy in object recognition using deep learning.
method Calculates class distances and selects mixed data from suitable classes dynamically.
result Improves recognition performance on general and long-tailed image recognition datasets.
This paper is devoted to the systematic investigation of the cone construction for Riemannian G manifolds M, endowed with an invariant metric connection with skew torsion ∇c, a `characteristic connection'. We show how to define a Gˉ structure on the cone $\bar M=M\x \R^+$ with a cone metric, and we prov…
Sharp-SSL uses random projections to identify important variables for semi-supervised learning.
problem High-dimensional semi-supervised learning problems.
method Careful aggregation of low-dimensional results from many axis-aligned random projections.
result Sharp-SSL algorithm can recover signal coordinates with high probability.
Nonnegative Matrix Factorization (NMF) has been a popular representation method for pattern classification problem. It tries to decompose a nonnegative matrix of data samples as the product of a nonnegative basic matrix and a nonnegative coefficient matrix, and the coefficient matrix is used as the new representation. …
Exploits class similarity for better machine learning models with confidence labels and projective loss functions.
problem Poor model performance due to confusing similar classes.
method Exploits class similarity with confidence labels and projective loss functions.
result Improved model performance on noisy labels.
We present a graph-based variational algorithm for multiclass classification of high-dimensional data, motivated by total variation techniques. The energy functional is based on a diffuse interface model with a periodic potential. We augment the model by introducing an alternative measure of smoothness that preserves s…
A number of classification problems need to deal with data imbalance between classes. Often it is desired to have a high recall on the minority class while maintaining a high precision on the majority class. In this paper, we review a number of resampling techniques proposed in literature to handle unbalanced datasets …
New architecture learns class representations from few examples.
problem Few-shot learning with high-quality class representations.
method Conditional embeddings based on target images, flexible network for comparisons.
result Achieves state-of-the-art performance on fine-grained classification task.
Collective classification models attempt to improve classification performance by taking into account the class labels of related instances. However, they tend not to learn patterns of interactions between classes and/or make the assumption that instances of the same class link to each other (assortativity assumption).…
Distilled teacher model transfers knowledge to student model on new datasets.
problem Improving model quality using unlabeled data.
method Knowledge distillation with an unlabeled teacher model.
result Teacher model knowledge transfers to student model on out-of-distribution datasets.
Paper proposes a new loss function for PU learning without negative examples.
problem Traditional machine learning struggles with negative examples, leading to biased predictions.
method Developed a collective loss function (cPU) for positive and unlabeled data.
result The cPU consistently outperforms existing methods in PU learning benchmarks and real-world datasets.
Proposes a new loss function for distributional learning.
problem Learning sparse and singular distributions.
method Entropy-regularized optimal transport and Fenchel duality.
result Geometric loss results in unconstrained convex objective functions.
Novel L1-norm and L2-norm LDA methods improve discriminant analysis.
problem Improving linear discriminant analysis for robustness and adaptability.
method Proposes L1BLDA and L2BLDA using Bhattacharyya error bound, maximizing between-class scatters and minimizing within-class scatters.
result Proposed methods avoid SSS and have no rank limit, demonstrating robust performance and effectiveness.
Extends L2-norm LDA to 2D inputs using Bhattacharyya bound.
problem L2-norm LDA loses useful image information for 2D inputs.
method 2DBLDA maximizes matrix-based between-class distance and minimizes within-class distance, optimizing Bhattacharyya error bound.
result 2DBLDA improves image recognition and face reconstruction.
A new DR method for HSI classification improves accuracy with limited samples.
problem Challenges in DR for HSI classification with limited training samples.
method Graph-based spatial and spectral regularized local scaling cut (SSRLSC).
result Improved classification accuracy compared to spectral-only methods.
We present a graph-based variational algorithm for classification of high-dimensional data, generalizing the binary diffuse interface model to the case of multiple classes. Motivated by total variation techniques, the method involves minimizing an energy functional made up of three terms. The first two terms promote a …
New method visualizes decision boundaries of classification models.
problem Difficulty in understanding how classification models interpret data.
method Hybrid supervised-unsupervised technique for visualizing decision boundaries.
result Provides interpretable maps for qualitative and quantitative analysis.
Paper optimizes classification of distributions using Wasserstein metric.
problem Classifying instances represented by distributions on a vector space.
method Maximizing Fisher's ratio in the Wasserstein metric space through iterative algorithm.
result The method enhances classification performance and is robust to variations in distribution summaries.
Optimizes natural frequencies of cellular composites with various microstructures.
problem Designing cellular composites with diverse microstructures for maximizing natural frequencies.
method Data-driven topology optimization with a latent-variable Gaussian process model.
result Cellular designs with multiclass microstructures achieve higher natural frequencies.
A new method improves few-shot learning by combining ProtoNet with LFD.
problem Few-shot learning struggles with high variance support sets.
method Combines ProtoNet with Local Fisher Discriminant Analysis.
result Superior classification accuracy on miniImageNet and tieredImageNet.
TzK model learns from multiple datasets efficiently.
problem Efficiently learning from multiple heterogeneous datasets.
method Flow-based conditional generative model trained with maximum likelihood.
result Comparable log likelihood to state-of-the-art models.
Improved k-NN active learning with local smoothness assumption.
problem Active learning convergence rates under smoothness assumptions.
method Designing an active learning algorithm with better convergence rate using local smoothness assumption for k-NN.
result Better convergence rate than in passive learning.
Maximizes coding rate difference for robust, discriminative features.
problem Learning robust, discriminative features from high-dimensional data.
method Maximal Coding Rate Reduction (MCR^2) principle.
result Significantly more robust to label corruptions in classification.
We study prediction and estimation problems using empirical risk minimization, relative to a general convex loss function. We obtain sharp error rates even when concentration is false or is very restricted, for example, in heavy-tailed scenarios. Our results show that the error rate depends on two parameters: one captu…
Proposes Structuring AutoEncoders to learn structured latent spaces.
problem Traditional Autoencoders fail to discover semantic structure in raw data.
method Enhances traditional Autoencoders using weak supervision to form a structured latent space.
result Structured latent space allows for more efficient data representation and tasks like classification.
Paper connects rejection learning to Bhattacharyya divergence.
problem Learning models to abstain from predictions.
method Developed a link between rejection and thresholding different statistical divergences, focusing on Bhattacharyya divergence.
result Rejector obtained by joint ideal distribution corresponds to thresholding of skewed Bhattacharyya divergence.
New method builds robust trees from noisy data.
problem Building accurate classification trees from noisy labeled data.
method Combines SVM-like splitting rules and label noise detection.
result Effective in detecting and mitigating label noise.
A statistical model predicts generalization in few-shot learning.
problem Lack of validation sets in few-shot learning makes generalization estimation difficult.
method Introduced a Gaussian model of feature distribution and an unbiased estimator for class-conditional density distances.
result Our approach outperforms alternatives like leave-one-out cross-validation.
The study finds conditions for positive braid knots to be Gordian adjacent and explores their unknotting sequences.
problem Understanding Gordian adjacency in positive braid knots.
method Manipulating braid words to find conditions for Gordian adjacency and exploring unknotting sequences.
result There are only finitely many positive braid knots for a given unknotting number.
A framework learns dynamic soft labels to improve model generalization and accuracy.
problem Models trained on one-hot labels overfit and are sensitive to noisy annotations.
method Proposes a framework where labels are treated as learnable parameters, adapting dynamically during optimization.
result Consistent gains across different datasets and architectures, improving ResNet18 by 2.1% on CIFAR100.
AHA model mimics animal episodic learning without labels.
problem Machine learning's slow, statistical learning vs. animals' fast episodic learning.
method Biologically-plausible computational model of the Hippocampus trained without labels.
result AHA model performs image classification comparably to deep ANNs.
Novel model identifies unseen classes with a single example.
problem Weakly supervised one-shot detection of unseen classes.
method Siamese similarity network with attention mechanism.
result Significantly outperforms baseline methods in experiments.
The paper analyzes and proposes an algorithm for multi-modal nonlinear embeddings with theoretical performance bounds.
problem Generalizability of multi-modal nonlinear embeddings to unseen data.
method Theoretical analysis and a multi-modal nonlinear representation learning algorithm motivated by performance bounds.
result The proposed algorithm yields promising performance in multi-modal image classification and cross-modal image-text retrieval applications.
This report works out the details of a closed-form, fully Bayesian, multiclass, openset, generative pattern classifier using multivariate Gaussian likelihoods, with conjugate priors. The generative model has a common within-class covariance, which is proportional to the between-class covariance in the conjugate prior. …
Paper develops a new classifier for time series using topological signatures and Sinkhorn divergences.
problem Classifying time series from chaotic systems with unknown models and noise.
method Topological signatures as weighted KDEs over persistent homology diagrams, predicting labels with Sinkhorn divergences.
result The method accurately discriminates between chaotic system states close in parameter space, robust to noise.
WAR method improves classifier robustness in noisy label datasets.
problem Learning robust classifiers in presence of noisy labels.
method Adversarial regularization based on Wasserstein distance.
result WAR method outperforms state-of-the-art competitors on noisy label datasets.
Paper defines class discrepancy for machine learning problems and provides coresets.
problem Addressing discrepancies in machine learning models.
method Defines class discrepancy, provides techniques for bounding discrepancy, and develops coresets and streaming sketches.
result Establishes coresets of size O(sqrt{d}/epsilon) for various machine learning problems.
New hashing method improves document retrieval precision.
problem Efficiently retrieving similar documents from large text databases.
method Pairwise supervised hashing with Bernoulli VAE and unbiased gradient estimator.
result Superior performance compared to existing methods.
InstanceFlow visualizes classifier confusion over training epochs.
problem Limited model interpretability through aggregate performance measures.
method Dual-view visualization tool showing instance-level learning behavior.
result Allows temporal analysis of training process and instance-level performance.
Proposes a new method for learning flexible nonparametric kernels.
problem Improving model flexibility in margin-based kernel methods.
method Data-adaptive non-parametric kernel learning framework with two constraints.
result Enhanced model flexibility and improved performance on benchmark data sets.
Compared to machines, humans are extremely good at classifying images into categories, especially when they possess prior knowledge of the categories at hand. If this prior information is not available, supervision in the form of teaching images is required. To learn categories more quickly, people should see important…