SNE successfully separates well-separated clusters in high dimensions.
problem No theoretical results are known for SNE and its variants.
method Stochastic Neighbor Embedding and variants.
result SNE successfully separates well-separated clusters in high dimensions.
Convex program recovers mixture components in well-separated data.
problem Mixed linear regression with well-separated classes.
method Second-order cone program based on L1 minimization.
result The convex program exactly recovers mixture components under well-separation assumptions.
Consistent estimator for mixtures of nonparametric elliptical distributions helps cluster analysis.
problem Consistency of maximum likelihood estimator for mixtures of nonparametric elliptical distributions.
method Maximum likelihood estimation for mixtures of elliptically-symmetric distributions under nonparametric P P P . result Components of the estimator correspond to well-separated components of the underlying distribution P P P . This paper proves t-SNE can recover well-separated clusters, improving visualization and embedding quality.
problem The lack of mathematical foundations and inner workings of t-SNE.
method Proves t-SNE's ability to recover well-separated clusters, using early exaggeration phase and rigorous analysis.
result t-SNE in the early exaggeration phase can be rigorously analyzed and provides novel ways to set parameters.
Study the geometric structure of graph Laplacian embeddings for manifold data.
problem Identifying coarse structure in manifold data sampled from a mixture model.
method Analyze spectral clustering procedure for data sampled from a manifold, focusing on graph Laplacian embeddings.
result Embedded data concentrates on cones centered around orthogonal vectors when the mixture model is well-separated.
Convex clustering can only learn convex clusters, with significant gaps between clusters.
problem Understanding the limitations and capabilities of convex clustering.
method Analyzing convex clustering solutions, proving properties, and characterizing clusters.
result Convex clustering can only learn convex clusters with significant gaps between clusters.
We inspect a possible clustering structure of the corruption perception among 134 countries. Using the average linkage clustering, we uncover a well-defined hierarchy in the relationships among countries. Four main clusters are identified and they suggest that countries worldwide can be quite well separated according t…
A new approach clusters data first, then embeds each cluster, improving transparency.
problem Visualizing data with latent clusters while preserving global geometry.
method First cluster, then embed each cluster, aligning the clusters.
result The approach is competitive with existing methods and more transparent.
Algorithm distinguishes Gaussian mixtures from pure Gaussians in quasi-polynomial time.
problem Distinguishing mixtures of Gaussian components from pure Gaussians, especially when components are well-separated.
method Sum-of-Squares method, quasi-polynomial time algorithm, bipartitioning sample to separate components.
result Algorithm can reliably distinguish between mixtures and pure Gaussians in quasi-polynomial time.
New method scales features for better clustering.
problem Irregular features disrupt classification.
method Spectral clustering with modified feature scales.
result Outperforms existing methods in experiments.
Adaptive clustering and personalization algorithms minimize regret in multi-agent stochastic linear bandits.
problem Minimizing regret in a multi-agent stochastic linear bandits framework with user heterogeneity.
method Proposes a novel algorithm that refines cluster identities and minimizes regret, adapting to cluster separation and user parameter deviations.
result Regret scales as O ( T / N ) \mathcal{O}(\sqrt{T/N}) O ( T / N ) for well-separated clusters and O ( T 1 2 + ε / ( N ) 1 2 − ε ) \mathcal{O}(T^{\frac{1}{2} + \varepsilon}/(N)^{\frac{1}{2} -\varepsilon}) O ( T 2 1 + ε / ( N ) 2 1 − ε ) for poorly separated clusters. New algorithm for private minimum spanning tree release with improved accuracy.
problem Privacy-preserving minimum spanning tree release for graphs.
method Formal differential privacy definition for graphs, new MST algorithm, combining sanitizing mechanism and MST clustering.
result Improved accuracy in weight approximation compared to state of the art.
New clustering algorithm uses reverse nearest neighbour for better density-based clustering.
problem Density-based clustering of separated high-density regions.
method Uses reverse nearest neighbour (RNN) queries to estimate densities and recover clusters.
result Outperforms DBSCAN and ISDBSCAN on synthetic and real-world data.
Sparse subspace clustering (SSC) is an elegant approach for unsupervised segmentation if the data points of each cluster are located in linear subspaces. This model applies, for instance, in motion segmentation if some restrictions on the camera model hold. SSC requires that problems based on the l 1 l_1 l 1 -norm are solved …
New indices for determining cluster compactness and separability.
problem Challenges in identifying true clusters in data sets.
method Developed absolute cluster indices to measure compactness and separability.
result Demonstrated improved performance compared to existing indices.
Proposes a method to discover graph structure in cluster results.
problem Discovering the relations between clusters in clustering results.
method Pairwise overlapping k-means clustering with adjustable parameters.
result Works well on real datasets like financial indices and restaurants.
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.
Tree Index evaluates cluster quality by creating decision trees from data.
problem Evaluating the quality of cluster results from various techniques.
method Tree Index creates a decision tree from clustered data, combining entropy and depth of leaves.
result Tree Index discriminates between sensible and non-sensible clusters on brain dataset.
OSil algorithm optimizes clustering quality using ASW.
problem Optimizing clustering quality using ASW.
method Distance-based objective function optimizing ASW for clustering.
result OSil algorithm outperforms other clustering methods in clustering quality.
TDA improves FX clustering quality over traditional methods.
problem Capturing complex currency co-movements in FX markets.
method Topological Data Analysis (TDA) compared to traditional statistical methods on monthly FX returns.
result TDA-based clustering yields more compact and well-separated clusters.
The moduli space metric and its Kahler potential for well-separated non-Abelian vortices are obtained in U(N) gauge theories with N Higgs fields in the fundamental representation.
GC-Flow uses graph flows for better clustering than traditional GCNs.
problem Traditional GCNs miss useful clustering information.
method Designing normalizing flows to replace GCN layers, creating a generative model.
result GC-Flow produces well-separated clusters while maintaining predictive power.
A new Wasserstein K K K -means method for clustering probability distributions.
problem Clustering probability distributions using the Wasserstein metric.
method Distance-based K K K -means with SDP relaxation for Wasserstein barycenters. result Distance-based K K K -means outperforms centroid-based K K K -means for clustering probability distributions. EM algorithm achieves optimal sample complexity for well-separated Gaussian mixtures.
problem Estimating parameters of well-separated Gaussian mixtures.
method New EM convergence proof for well-separated Gaussian mixtures.
result EM algorithm converges with Ω ( log k ) Ω(\sqrt{\log k}) Ω ( log k ) separation, achieving O ( k d / ε 2 ) O(kd/ε^2) O ( k d / ε 2 ) samples. EDD uses entropy of distance distributions to cluster unlabeled data.
problem Challenges in clustering unlabeled high-dimensional data.
method EDD employs Shannon entropy to quantify distance distribution peaks.
result EDD detects varying degrees of clustering sensitivity.
The paper proposes a method to select clusters, models, and algorithms based on quadratic discriminant scores.
problem Selecting the number of clusters, models, and algorithms in cluster analysis.
method Develops quadratic scores for cluster quality, uses bootstrap resampling, and compares partitions.
result The proposed method achieves better overall performance compared to other state-of-the-art methods.
New findings on robust learning with well-separated data.
problem Learning robust classifiers with well-separated classes.
method Analyzing sample complexity for linear classifiers with robustness.
result For linear classifiers on well-separated data, robust loss is at least $Ω( rac{d}{n})$ .
Constructs classifiers for neural networks with specific data configurations.
problem Finding global minima of deep ReLU neural networks on sequentially separable data.
method Explicitly constructs zero loss neural network classifiers using cumulative parameters and truncation maps.
result Global minimizers can be described with a limited number of parameters based on the data structure.
A new method for clustering heterogeneous data using likelihood-adjusted SDP.
problem Clustering heterogeneous data with different cluster shapes and sizes.
method Iterative likelihood-adjusted semidefinite programming (iLA-SDP) method.
result iLA-SDP achieves lower mis-clustering errors compared to other methods.
New insights into spurious local minima in k-means clustering.
problem Understanding and mitigating spurious local minima in k-means clustering.
method Investigating spurious local minima under a probabilistic generative model.
result Proven structures of spurious local minima for k-means clustering.
Genie clusters faster and resists outliers.
problem Hierarchical clustering's sensitivity to outliers and slow computation.
method Genie uses an economic inequity measure to link clusters, balancing speed and quality.
result Genie outperforms other linkage methods in clustering quality and speed.
Paper proves noise-tolerant SSC using greedy methods under coherence conditions.
problem Proving noise-tolerant SSC using greedy methods under coherence conditions.
method Derives coherence-based sufficient conditions for correct neighbor identification using MP/OMP in the presence of bounded noise.
result MP/OMP succeed in identifying correct neighbors under certain noise levels, leading to higher clustering accuracy.
Study reveals structure of local minima in GMMs, identifying key cluster centers.
problem Identifying optimal cluster centers in non-convex GMM landscapes.
method Analyzing the negative log-likelihood function of GMMs in the population limit.
result Local minima share a common structure that partially identifies true cluster centers.
A new point process for clustering distributions with repulsion.
problem Clustering distributions with repulsion.
method Distributional Determinantal Point Process (dDPP) with sliced Wasserstein kernel.
result Validated dDPP as a well-defined point process and applied to gene expression and epilepsy data.
Improved sample complexity for Gaussian process approximations.
problem Efficiently approximating Gaussian processes with sparse spectrum.
method Improved sample complexity analysis and auto-encoding algorithm.
result Gaussian process predictions and model evidence can be well-approximated with low sample complexity.
This work introduces a novel method to evaluate generative model novelty.
problem Evaluating the novelty of generative models compared to a reference model.
method Spectral approach to differential clustering and Kernel-based Entropic Novelty (KEN) score.
result The KEN score effectively detects novel modes and compares generative models.
Paper tackles learning mixture of RUMs from partial data.
problem Learning a mixture of Random Utility Models (RUMs) from pairwise comparisons.
method PCA-based spectral clustering to reduce mixture to single component.
result Algorithm correctly clusters data from a mixture of RUMs with high probability.
New method estimates density ratio for well-separated distributions using multi-class logistic regression.
problem Challenges in estimating density ratio for well-separated distributions.
method Uses multi-class logistic regression with auxiliary densities to estimate log(p/q).
result Demonstrates superior performance on density ratio estimation, mutual information, and representation learning tasks.
New algorithm recovers mixture means even with many outliers.
problem Estimating mixture means when outliers overwhelm small groups.
method Proposes an algorithm for robust mixture learning with minimal list-size overhead.
result Order-optimal error guarantees for each mixture mean.
Adversarial online multi-task RL with task separation.
problem Minimize regret in an adversarial online multi-task setting with unknown MDPs.
method Prove minimax and instance-specific lower bounds, develop a clustering algorithm with optimal sample complexity and regret.
result Tight sample complexity and regret bounds for adversarial online multi-task RL.
Single neural network learns multiple tasks from combined data.
problem Can a single neural network learn multiple unrelated tasks?
method Investigates how task representations affect joint learning; uses various task encoding methods.
result Single neural network can learn multiple tasks from combined data, even when tasks are unrelated and different.
Gaussian sketching preserves kernel inner products in low dimensions.
problem Preserving kernel inner products in low-dimensional spaces.
method Gaussian sketching of kernel Gram matrices and random projections in RKHS.
result Sketching yields a random projection operator that preserves weighted RKHS inner products.
At critical coupling, the interactions of Ginzburg-Landau vortices are determined by the metric on the moduli space of static solutions. The asymptotic form of the metric for two well separated vortices is shown here to be expressible in terms of a Bessel function. A straightforward extension gives the metric for N vor…
Improved sample efficiency with normalized RBF kernels in neural networks.
problem Learning more with less data in deep learning models.
method Two-phase method to train neural networks with normalized RBF kernels as output layer.
result Normalized RBF kernel networks achieve higher sample efficiency, compactness, and separability.
SGD learns a simple network for multi-class classification from mixtures of well-separated distributions.
problem Learning overparameterized neural networks for multi-class classification.
method Stochastic Gradient Descent (SGD) on structured data.
result SGD learns a network with small generalization error from mixtures of well-separated distributions.
The large-scale structure of the universe is comprised of virialized blob-like clusters, linear filaments, sheet-like walls and huge near empty three-dimensional voids. Characterizing the large scale universe is essential to our understanding of the formation and evolution of galaxies. The density range of clusters, wa…
Simple Deep LDA models achieve accuracy competitive with softmax baselines.
problem Training Deep LDA models by maximum likelihood estimation leads to overlapping or collapsed class clusters.
method Proposed a constrained Deep LDA formulation with geometric constraints to fix class means and covariance.
result MLE becomes stable under geometric constraints, yielding well-separated class clusters.
This study examines when non-parametric methods are robust to adversarial examples.
problem Understanding when non-parametric methods are robust to adversarial examples.
method Examined general non-parametric methods and established conditions for r-consistency.
result Non-parametric methods like nearest neighbors and kernel classifiers are r-consistent when data is well-separated, while histograms are not.