Paper finds universal speech command perturbations that fool models.
problem Existence of universal adversarial examples in speech command classification.
method Proposed a novel analytical framework for evaluating universal perturbations and a detailed distortion measurement method.
result Universal perturbations can fool speech command classification models across different models.
Reduces bounded loss learning to binary classification.
problem Universal consistency of non-i.i.d. processes with bounded loss.
method Constructive reduction to binary classification.
result Any bounded loss output setting can be reduced to binary classification.
A new sector classification method outperforms existing ones in risk-adjusted returns.
problem Subjective sector classification heuristics like GICS and NAICS are not optimal.
method Learned sector classification using hierarchical clustering and reIndexer evaluation tool.
result 17-sector learned sector universe outperforms GICS and NAICS in backtests.
This paper studies universal rates of ERM for binary classification under agnostic learning.
problem The challenge of achieving universal rates of ERM for binary classification under agnostic learning.
method The paper explores the agnostic universal rates of ERM for binary classification, revealing three possible rates: e−n, o(n−1/2), or arbitrarily slow. result The paper provides a complete characterization of which concept classes fall into each of the three categories of agnostic universal rates.
New findings show Gaussian universality breaks down in high-dimensional linear factor mixtures.
problem The limitations of Gaussian universality in high-dimensional classification.
method Characterization of empirical risk minimization for classification under linear factor mixture models.
result Gaussian universality breaks down under high-dimensional linear factor mixtures.
This paper improves universal sound separation using sound classification.
problem Separating acoustic sources from an open domain, regardless of their class.
method Utilizing semantic embeddings from a sound classifier to condition a separation network.
result Classifier embeddings provide nearly one dB of SNR gain, and iterative models achieve significant performance.
Framework for universal graph function approximators outperforms existing methods.
problem Graph classification and separation of graph classes.
method Inspired by persistent homology, dependency parsing, and multivalued functions, the framework constructs universal approximators on graph isomorphism classes.
result Achieves state-of-the-art performance on four graph datasets.
MAT combines meta-learning and adversarial training to defend against universal patches.
problem Defending against universal patches that fool models in various contexts.
method Meta adversarial training (MAT) integrates meta-learning with adversarial training.
result MAT increases robustness against universal patch attacks on image classification and traffic-light detection.
Paper interprets ResNets via gate-network controls and deep-layer classifications.
problem Understanding the performance mechanism of ResNets.
method Constructs typical solutions using gate-network controls and deep-layer classifications.
result Proves the universal-approximation capability of ResNets.
URT layer improves few-shot image classification across diverse domains.
problem Few-shot image classification in multi-domain settings.
method Meta-learns to dynamically re-weight and compose domain-specific representations.
result Sets new state-of-the-art on Meta-Dataset.
Study classifies super vector bundles and proves universality.
problem Homotopy classification of super vector bundles.
method Construction of supergrassmannians, Gauss morphism, multilinear algebra, direct and inverse limits.
result Proves the resulting super vector bundle is universal.
New rules found to fool deep neural networks in text classification.
problem Vulnerabilities of deep neural networks in text classification.
method Coevolutionary optimization algorithm to create imperceptible adversarial samples.
result Universal rules for fooling deep neural networks in text classification exist and are sample and method agnostic.
Theory extends optimal learning rates without realizability assumption.
problem Agnostic binary classification without realizability assumption.
method Identifies tetrachotomy of optimal rates and combinatorial structures.
result Optimal universal rates for binary classification in agnostic setting.
PanRep learns universal node embeddings for heterogeneous graphs.
problem Learning universal node embeddings for heterogeneous graphs.
method Graph Neural Network (GNN) model with four decoders capturing different properties.
result PanRep outperforms unsupervised and supervised methods in node classification and link prediction.
Adversarial training helps classifiers resist universal perturbations.
problem Vulnerability of classifiers to universal perturbations.
method Adversarial training with shared adversarial examples.
result Adversarial training reduces sensitivity to universal perturbations.
Unified framework for comparing classification metrics across different imbalance rates.
problem Differences in scale and sensitivity to class imbalance rates in classification metrics.
method Introduces outperformance standardization (OPS) function to map metrics to a common scale.
result Unified o-value metric provides clear comparison across different imbalance rates.
Unique domain found in Einstein universe, simplifying manifold classification.
problem Classifying closed conformally flat manifolds with proper development.
method Identifying and analyzing almost-homogeneous domains in the Einstein universe.
result Found a unique domain (diamond) in the Einstein universe that simplifies manifold classification.
The orbifold group of the Borromean rings with singular angle 90 degrees, U, is a universal group, because every closed oriented 3--manifold M3 occurs as a quotient space M3=H3/G, where G is a finite index subgroup of U. Therefore, an interesting, but quite difficult problem, is to classify the fin…
Universal perturbations misclassify text with high accuracy.
problem Vulnerability of text classifiers to small perturbations.
method Algorithm to compute universal adversarial perturbations.
result Deep neural networks are highly vulnerable to universal adversarial perturbations.
Universal audio perturbations fool multiple classification models.
problem Creating audio adversarial perturbations that work across different models.
method Two methods: greedy iterative approach and novel penalty formulation.
result The penalty method produces more successful attacks with limited training data.
Paper establishes a universal growth rate for smooth surrogate losses in classification.
problem Analyzing growth rates of consistency bounds for various surrogate losses.
method Proves square-root growth rate for smooth margin-based losses; extends to multi-class classification.
result Demonstrates a universal square-root growth rate for smooth comp-sum and constrained losses.
Study classifies twist knots with maximal self-linking number in S^3.
problem Classifying transverse-universal knots in S^3.
method Classification of transverse twist knots with maximal self-linking number.
result Obtained an infinite family of non-transverse-universal knots.
In this work we show that, using the eigen-decomposition of the adjacency matrix, we can consistently estimate feature maps for latent position graphs with positive definite link function κ, provided that the latent positions are i.i.d. from some distribution F. We then consider the exploitation task of vertex classi…
This paper classifies periodic weaves and their universal cover, extending Tait's conjectures.
problem Classifying periodic weaves and their universal cover in thickened surfaces.
method Introducing hyperbolic periodic weaves, extending Tait's conjectures, and using a generalized Kauffman bracket polynomial.
result Tait's conjectures are extended to minimal reduced alternating weaving motifs.
Prototype rules simplify multiclass classification in metric spaces, achieving consistency and reduced complexity.
problem Multiclass classification in metric spaces, focusing on universal consistency and convergence rates.
method Novel Proto-NN and hybrid rules for multiclass classification in metric spaces, analyzing convergence rates.
result Proto-NN is universally consistent and simpler to implement, with similar computational advantages.
A new Universal Activation Function improves performance across various machine learning tasks.
problem Achieving near optimal performance in different machine learning tasks.
method Optimization algorithms evolve the UAF's parameters to match the optimal activation function for each task.
result The UAF converges to near optimal performance in classification, quantification, and reinforcement learning tasks.
We study the problem of learning classifiers robust to universal adversarial perturbations. While prior work approaches this problem via robust optimization, adversarial training, or input transformation, we instead phrase it as a two-player zero-sum game. In this new formulation, both players simultaneously play the s…
In the present paper, we study the finite type invariants of Gauss words. In the Polyak algebra techniques, we reduce the determination of the group structure to transformation of a matrix into its Smith normal form and we give the simplified form of a universal finite type invariant by means of the isomorphism of this…
WWe define the notion of a random metric space and prove that with probability one such a space is isometricto the Urysohn universal metric space. The main technique is the study of universal and random distance matrices; we relate the properties of metric (in particulary universal) space to the properties of distance …
Study shows fine-tuned linear models outperform pretrained ones in transfer learning.
problem Transfer learning and fine-tuning in linear models for regression and binary classification.
method Stochastic gradient descent on pretrained linear models with small target data sets.
result Fine-tuned models outperform pretrained ones under certain conditions.
A new generalization of Grassmannians, called ν-grassmannians, and a canonical super vector bundle over this new space, say Γ, are introduced. Then, constructing a Gauss supermap of a super vector bundle, the universal property of Γ is discussed. Finally, we generalize one of the main theorems of homotopy classificatio…
We describe the orbits of the irreducible action of PSL(2, R) on the 3-dimensional Einstein universe Ein 1,2. This work completes the study in [2], and is one element of the classification of cohomo-geneity one actions on Ein 1,2 ([5]).
Study tight contact structures on figure-eight knot surgeries.
problem Classify tight contact structures on surgeries of figure-eight knot.
method Analyzes surgeries on figure-eight knot, determining tightness, symplectic fillability, and universality.
result First classification of tight contact structures on surgeries of figure-eight knot.
Adversarially trained transformers can learn robustly across tasks with minimal tuning.
problem Adversarial attacks and the high cost of adversarial training.
method Adversarial pretraining followed by in-context learning.
result Single-layer linear transformers can generalize robustly to unseen tasks.
Classifies orbits of Hurwitz actions on dihedral quandles.
problem Classifying orbits of Hurwitz actions on dihedral quandles.
method Introduced three computable invariants to classify orbits.
result Complete classification of orbits under Hurwitz action.
Graph homomorphism numbers embed graphs for classification.
problem Graph classification using graph homomorphisms.
method Embed graphs into vectors using homomorphism numbers.
result Homomorphism vectors are universal for approximating graph invariants.
Paper develops a privacy-preserving nonparametric regression method.
problem Nonparametric regression with local differential privacy constraints.
method Privatised discretisation and Laplace noise applied to feature vectors and responses.
result Strongly universally consistent estimator for regression and classification.
Improves domain classification across multiple locales with shared language.
problem Improves domain classification accuracy in Spoken Language Understanding across multiple locales with shared language.
method Selective multi-task learning to create a joint representation of utterances over locales with different sets of domains.
result The proposed approach outperforms other baselines models especially when classifying locale-specific domains and low-resourced domains.
Paper creates universal adversarial attacks.
problem Creating universal, transferable, and targeted adversarial attacks.
method Learn a universal mapping to map sources to adversarial examples.
result Examples can fool networks into classifying all into one targeted class and have strong transferability.
This research classifies singular foliations and finds a universal deformation.
problem Classifying singular foliations on (C2,0). method Topological universal deformation through fixed invariants.
result Every equisingular deformation uniquely factors through the topological universal deformation.
In hyperbolic space, the angle of intersection and distance classify pairs of totally geodesic hyperplanes. A similar algebraic invariant classifies pairs of hyperplanes in the Einstein universe. In dimension 3, symplectic splittings of a 4-dimensional real symplectic vector space model Einstein hyperplanes and the inv…
Inductive transfer learning has greatly impacted computer vision, but existing approaches in NLP still require task-specific modifications and training from scratch. We propose Universal Language Model Fine-tuning (ULMFiT), an effective transfer learning method that can be applied to any task in NLP, and introduce tech…
Paper introduces MRCs that minimize worst-case 0-1 loss, providing tight performance guarantees.
problem Minimizing worst-case 0-1 loss in classification.
method MRCs that minimize worst-case 0-1 loss with uncertainty sets of distributions.
result MRCs provide tight performance guarantees and are strongly universally consistent.
Authors disagree with recent findings on random Gaussian weights in DNNs.
problem The relationship between angle and distance shrinkage in DNNs with random Gaussian weights is incorrect.
method Comparison of recent findings with new observations on random Gaussian weights in DNNs.
result Theorem 3 and Figure 5 in the recent paper are not accurate.
Proposes DCADL for efficient image classification with reduced complexity.
problem Efficiency and discriminative capability in DL methods for image classification.
method Jointly learns a convolutional analysis dictionary and a universal classifier, reducing time complexity.
result Achieves competitive accuracy with reduced computational cost.
Implementing k-NN classification using Gromov--Wasserstein distances
problem Comparing metric measure spaces
method Gromov--Wasserstein and fused Gromov--Wasserstein distances
result Universal consistency of k-NN classifiers While deep learning is remarkably successful on perceptual tasks, it was also shown to be vulnerable to adversarial perturbations of the input. These perturbations denote noise added to the input that was generated specifically to fool the system while being quasi-imperceptible for humans. More severely, there even exi…
A Hilbert space embedding for probability measures has recently been proposed, wherein any probability measure is represented as a mean element in a reproducing kernel Hilbert space (RKHS). Such an embedding has found applications in homogeneity testing, independence testing, dimensionality reduction, etc., with the re…