Paper proposes a method to estimate true positive proportion without knowing it.
problem Bias in binary classifier performance due to different positive item proportions.
method Maximum likelihood estimator for true proportion of positives.
result Method accurately estimates true positive proportion in data sets.
New method improves false-/true-positive-rate estimation in fraud detection with noisy labels.
problem Estimating FPR/TPR in fraud detection with class-conditional label noise.
method Directly cleaning model's validation data to de-correlate cleaning error with model scores.
result Improves accuracy of FPR/TPR estimates, especially in asymmetric label noise scenarios.
Two kernel Stein tests control decision errors in non-parametric model comparison.
problem Non-parametric multiple model comparison.
method Two statistical tests controlling false positive and false discovery rates.
result The first test has a higher true positive rate than the second under appropriate conditions.
This research examines anomaly detection metrics under class imbalance.
problem Challenges in interpreting evaluation metrics under class imbalance.
method Analysis of four common anomaly detection metrics (AUROC, AUPR, F1-score, MCC) under varying imbalance ratios.
result Visualisations of metric landscapes provide an intuitive view of metric preferences and stability.
The paper shows how to recover true node positions from a graph or similarity matrix.
problem Recovering true distances and positions from a graph or similarity matrix.
method Two steps: matrix factorisation followed by nonlinear dimension reduction.
result Nonlinear dimension reduction can recover latent positions close to a manifold where geodesic distance is encoded.
We prove a central limit theorem for the components of the largest eigenvectors of the adjacency matrix of a finite-dimensional random dot product graph whose true latent positions are unknown. In particular, we follow the methodology outlined in \citet{sussman2012universally} to construct consistent estimates for the …
We examine geometric properties of a knot J that are unchanged by taking a (p,q)-cable K of J. Specifically, we relate w(K) to w(J), where w(K) is the width of K in the sense of Gabai. We use this information to demonstrate that thin position is a minimal bridge position of J if and only if the same is true for K, and …
The doubling conjecture for positive scalar curvature is proven under certain conditions.
problem Determining when a manifold with a specific boundary condition admits positive scalar curvature.
method Surgery techniques for positive scalar and mean curvature, and existence of area-minimizing hypersurfaces.
result The doubling conjecture holds true for manifolds with certain split conditions on fundamental groups.
New neural network outperforms existing methods in scene matching.
problem Automated scene matching with high accuracy and low false positives.
method Convolutional hashing using a new loss function and training scheme.
result Significantly higher true positive rate and 100-fold reduction in false positives.
AdaDetectGPT improves text authorship detection with statistical guarantees.
problem Determining if text is authored by a human or an LLM.
method Adaptive learning of a witness function from training data to enhance logits-based detectors.
result AdaDetectGPT nearly uniformly improves text authorship detection, with up to 37% improvement.
SFS-DA method statistically tests FS reliability under domain adaptation.
problem Feature selection reliability under domain adaptation with limited target data.
method Selective Inference framework to control false positive rate and enhance true positive rate.
result SFS-DA method controls FPR below a pre-specified level α (e.g., 0.05) while maximizing true positive rate. We show that in cohomogeneity 3 there are G-manifolds with any given number of isolated singular orbits and an invariant metric of positive Ricci curvature. We show that the corresponding result is also true in cohomogeneity 5 provided the number of singular orbits is even.
In this paper, we study the following conjecture of Hamilton: Any compact gradient shrinking Ricci soliton with positive curvature operator must be Einstein. We first derive several identities. Then we show that the conjecture is true under an additional condition. Furthermore, such a soliton must be of constant curvat…
A multiple instance dictionary learning method using functions of multiple instances (DL-FUMI) is proposed to address target detection and two-class classification problems with inaccurate training labels. Given inaccurate training labels, DL-FUMI learns a set of target dictionary atoms that describe the most distincti…
We study the topology of the 13 dimensional positively curved Bazaikin spaces. We show that there is only one such manifold which is homotopy equivalent to a homogeneous space, the so called Berger space. This is in contrast to the case of the 7 dimensional positively curved Eschenburg spaces. In addition, we compute t…
Study functional confounders in causal inference, enabling estimable effects.
problem Causal inference challenges with functional confounders violating positivity.
method Functional interventions, functional positivity, gradient fields, Level-set Orthogonal Descent Estimation (LODE).
result Valid causal effect estimation under certain conditions.
Study shows infinitely many metrics with nonnegative sectional or positive Ricci curvature on specific 5D quotients.
problem Finding metrics with specific curvature properties on Brieskorn quotients.
method Analyzing moduli spaces of metrics with nonnegative sectional or positive Ricci curvature.
result Moduli spaces have infinitely many path components for both nonnegative sectional and positive Ricci curvature.
We provide a direct proof of a non-collapsing estimate for compact hypersurfaces with positive mean curvature moving under the mean curvature flow: Precisely, if every point on the initial hypersurface admits an interior sphere with radius inversely proportional to the mean curvature at that point, then this remains tr…
Simon Brendle's result extended to manifolds with positive isotropic curvature of dimension at least nine.
problem Proving the diffeomorphism of manifolds with positive isotropic curvature.
method Extending a result by Simon Brendle to manifolds with dimension at least nine.
result The result holds for manifolds with positive isotropic curvature of dimension at least nine.
The paper sets new bounds on metrics with positive scalar curvature.
problem Classifying metrics with positive scalar curvature.
method Lower bounds on the rank of the group of psc metrics over M.
result Lower bounds on the rank of psc metrics up to bordism.
The study detects and classifies touch gestures with high accuracy.
problem Detecting and classifying touch gestures from touch screens.
method Supervised learning techniques using a capacitive sensor array to record touch and swipe gestures.
result Logistic Regression models achieved over 95% accuracy for all gesture types.
A new method predicts true classes from positive and unlabeled data with additional labeled observations.
problem Predicting true classes from positive and unlabeled data with selection bias.
method Introduces augmented PU prediction, allowing feature-dependent labeling, and compares various empirical Bayes rules.
result The variational autoencoder-based method performs similarly or better than other methods and improves accuracy for unlabeled samples.
We establish a boundary connected sum theorem for asymptotically hyperbolic Einstein metrics; this requires no nondegeneracy hypothesis. We also show that if the two metrics have scalar positive conformal infinities, then the same is true for this boundary join.
Optimal systolic inequality proved for manifolds with positive bi-Ricci curvature.
problem Proving optimal systolic inequalities on manifolds with positive bi-Ricci curvature.
method Minimal surfaces method under the Generic Regularity Hypothesis.
result Optimal systolic inequality proved in all dimensions.
This study benchmarks changepoint detection algorithms on cardiac time series data.
problem Identifying state changes in cardiac time series for disease classification.
method Comparison of 8 changepoint detection algorithms on artificial and real cardiac time series data.
result RMDM algorithm achieved highest true positive rate and cross validated accuracy for classification.
We show that an n−dimensional Moishezon manifold is uniruled if and only if it supports a balanced metric ωn−1 of positive total scalar Chern curvature. A similar statement also holds true for class C manifolds of dimension three.
Fractional Sobolev maps with positive distributional Jacobians are continuous.
problem Proving continuity of maps in fractional Sobolev spaces with positive Jacobians.
method Extending known results from W1,n to Ws,sn for s≥n+1n, considering distributional Jacobians. result Fractional Sobolev maps with positive distributional Jacobians are continuous.
Coercivity condition ensures learning of interacting particle systems.
problem Ensuring identifiability of interaction functions in learning systems of interacting particles.
method Equivalence of coercivity condition to strictly positive definiteness of an integral kernel.
result For ergodic systems, the integral kernel is strictly positive definite, satisfying the coercivity condition.
Method reduces bias in occupation classification without protected attribute data.
problem Mitigating bias in occupation classification without access to protected attributes.
method Uses word embeddings to discourage correlation between predicted occupation probability and name.
result Reduces race and gender biases without significant loss in true positive rate.
Geodesic nets with three vertices have at most one balanced vertex.
problem Characterizing geodesic nets with specific vertex configurations.
method Analyzing geodesic nets on non-positively curved planes.
result Geodesic nets with three boundary vertices have at most one balanced vertex.
Study of 2+1 dimensional cosmologies with positive cosmological constant, proving asymptotic convergence to de Sitter.
problem Asymptotic behavior of 2+1 dimensional cosmologies with positive cosmological constant.
method Mean Curvature Flow methods.
result Spatial slices asymptotically converge to de Sitter, becoming physically indistinguishable from it.
Solves overdetermined boundary problems for semilinear equations.
problem Overdetermined boundary problems for semilinear equations with position-dependent nonlinearities.
method Analyzes Euclidean and compact Riemannian manifolds, proving existence of nontrivial solutions.
result Nontrivial solutions exist for a wide range of problems.
In accordance with the Bing-Borsuk conjecture \cite{bb}, we show that if X is an n-dimensional homogeneous metric ANR compactum and x∈X, then there is a local basis at x consisting of connected open sets U such that the homological properties of \bar U and bdU are similar to the properties of the closed ball…
A model learns from valid examples while avoiding invalid ones to generate better data.
problem Generative models produce nonsense when fitting to observed data.
method Active distribution learning using an invalidity oracle.
result Improper distribution learning can be done with polynomial queries, unlike proper learning which requires exponentially many.
Enhanced attacks quantify machine learning data leakage.
problem Quantifying how much machine learning models reveal about their training data.
method Hypothesis testing framework for membership inference attacks.
result New attacks achieve higher true positive rates with lower false positive rates.
Given a compact Riemannian manifold, with positive Yamabe quotient, not conformally diffeomorphic to the standard sphere, we prove a priori estimates for solutions to the Yamabe problem. We restrict ourselves to the dimensions less than or equal to 7, where the Positive Mass Theorem is known to be true. We also show th…
Approach collects missing outcomes to improve fairness in classification.
problem Lack of true outcomes for incorrectly classified samples leads to biased classifiers.
method Exploration-based data collection to ensure all subpopulations are represented and fairness properties are encoded.
result Trained classifier converges to a fair classifier with bounded false positives.
We consider the problem of vertex classification for graphs constructed from the latent position model. It was shown previously that the approach of embedding the graphs into some Euclidean space followed by classification in that space can yields a universally consistent vertex classifier. However, a major technical d…
Paper proves conjecture about Einstein metrics on manifolds with positive isotropic curvature.
problem Proving the Besse conjecture for metrics with positive isotropic curvature.
method Analyzing the critical point equation and using properties of metrics with positive isotropic curvature.
result The Besse conjecture is true for metrics with positive isotropic curvature.
Study shows metrics on certain manifolds lose positive curvature under Ricci flow.
problem Understanding the dynamics of positively curved metrics on specific manifolds.
method Analysis of invariant metrics on SU(3)/T2 and SU(m+2p)/S(U(m)imesU(p)imesU(p)) under homogeneous Ricci flow. result Metrics lose positive intermediate Ricci curvature under Ricci flow for certain dimensions.
The paper studies Kähler-Einstein metrics on fiber spaces with positive Kodaira dimension.
problem Understanding Kähler-Einstein metrics on fiber spaces with positive Kodaira dimension.
method Analyzes the properties of singular Kähler-Einstein metrics and their curvature.
result The fiberwise singular Kähler-Einstein metric induces a semipositively curved metric on the relative canonical bundle.
Proposes an EM method for learning from positive and unlabeled data with random selection assumption.
problem Learning from positive and unlabeled data with random selection assumption.
method Proposes an EM method to learn under the assumption that positive examples are selected at random, conditioned on some attributes.
result The proposed method outperforms state-of-the-art methods for learning under the selected completely at random assumption.
A common approach in positive-unlabeled learning is to train a classification model between labeled and unlabeled data. This strategy is in fact known to give an optimal classifier under mild conditions; however, it results in biased empirical estimates of the classifier performance. In this work, we show that the typi…
Study links' arc index and Turaev genus, proving conjectures.
problem Understanding the arc index and Turaev genus of links.
method Computed arc index, established bounds, and conjectured inequalities.
result Proved conjectures linking crossing number, arc index, and Turaev genus.
Active learning improves anomaly detection by tuning ensembles with true labels.
problem Configuring anomaly detectors for minimal effort on false positives.
method Compact description, data drift detection, active learning strategies.
result Active learning significantly improves anomaly discovery and adapts to streaming data.
Paper analyzes privacy risks of learning complex models.
problem Privacy leakage from learning high-dimensional graphical models.
method Theoretical analysis of tracing attacks on Bayesian networks.
result Provides a tight upper bound on tracing attack power.
The paper explores positivity conditions for χy-genus and their implications on Chern numbers and symplectic manifolds.
problem Optimizing Chern number inequalities for almost-complex manifolds.
method Introducing and analyzing positivity conditions for the modified χy-genus and applying them to Chern numbers and symplectic manifolds. result Optimal Chern number inequalities hold for many important Kähler and symplectic manifolds.
New optimization method improves AUC for binary classification and changepoint detection.
problem Non-convex AUC and sub-optimal points in ROC curves.
method AUM (Area Under Min(FP, FN)) surrogate loss function based on sorting and summing ROC curve points.
result AUM minimization learning algorithm improves AUC and speeds up compared to previous methods.