A method to approximate instance-dependent label noise using instance-confidence embedding.
problem Real-world label noise that depends on individual instances.
method Variational approximation with instance embedding to capture instance-specific label corruption.
result ICE method effectively approximates instance-dependent noise and detects ambiguous instances.
A probabilistic method for deep embedding that improves classification accuracy and interpretability.
problem Improving classification accuracy and interpretability in deep learning.
method A probabilistic approach that treats embeddings as random variables, using a product distribution over labeled instances and marginalizing prototype proximity.
result Superior large- and open-set classification accuracy compared to state-of-the-art methods.
Proposes a new method for MIR using kernel mean embeddings.
problem Multiple instance regression (MIR) where bags contain multiple instances with a single label.
method Computes kernel mean embeddings of predicted label distributions and learns a regressor from these embeddings.
result Better results than baseline instance-MIR across all datasets, state-of-the-art on two.
This work models uncertainty in instance embeddings using hedging.
problem Uncertainty in ambiguous inputs is not well represented by traditional embeddings.
method Hedged instance embedding (HIB) models embeddings as random variables and trains under variational information bottleneck.
result Improved performance in image matching and classification tasks, more structured embedding space, and per-exemplar uncertainty measure.
This paper proposes a new method for embedding sequences using Wasserstein distances.
problem Embedding sequences in a metric space for better pattern recognition.
method Develops a deep learning model that embeds sequences as distributions and uses Wasserstein distances for comparison.
result Distributional embeddings using Wasserstein distances outperform traditional vector embeddings.
Method detects anomalies on attributed graphs with few labeled instances.
problem Detecting anomalies on connected instances (attributed graphs) with limited labeled data.
method Embed nodes in latent space using GCNs, training to distinguish normal and anomalous nodes.
result Method outperforms existing methods on real-world attributed graph datasets.
Prototypical Networks improve multi-label classification accuracy.
problem Multi-label classification with nonlinear label dependencies.
method Formulate multi-label learning as class distribution in a non-linear embedding space. For each label, positive and negative embeddings are compactly distributed. Labels are inferred by measuring the distance to prototype positive or negative embeddings.
result Extensive experiments show improved accuracy compared to state-of-the-art algorithms.
New method tackles instance segmentation on 3D point clouds with improved metrics.
problem Evaluation metrics are affected by small regions containing few instances.
method Proposes a new method with O(Np) space complexity that learns embeddings for clusters of instances.
result Achieves state-of-the-art performance using both existing and proposed metrics.
Despite the breakthroughs achieved by deep learning models in conventional supervised learning scenarios, their dependence on sufficient labeled training data in each class prevents effective applications of these deep models in situations where labeled training instances for a subset of novel classes are very sparse -…
Paper tackles novelty detection in text classification.
problem Traditional text classification assumes known classes in testing, but often encounters unexpected instances.
method Converts problem to pair-wise matching, uses CNN with embedding matrices.
result Proposed method outperforms state-of-the-art baselines.
Paper tackles embedding attributed sequences in unsupervised learning.
problem Mining tasks over attributed sequences with dependencies between sequences and attributes.
method Proposes a deep multimodal learning framework, NAS, for unsupervised learning of attributed sequences.
result NAS produces task-independent embeddings for various mining tasks on real-world datasets.
Meta-learning improves relative density-ratio estimation from limited data.
problem Estimating relative density-ratios from few instances.
method Meta-learning using neural networks to extract and embed dataset information for relative DRE.
result Meta-learning enables efficient and effective adaptation to few instances for relative DRE.
Dropout improves MIL performance on noisy WSI classification.
problem Noisy feature embeddings and weak supervision in MIL for WSI classification.
method Empirical exploration of dropout in MIL, proposing MIL-Dropout.
result MIL-Dropout boosts MIL performance with minimal computational cost.
SymNet uses neural models to solve RMDPs efficiently.
problem Limited effectiveness of generalized policies in RMDPs.
method SymNet trains shared parameters for an RDDL domain, converts instances to graphs, uses relational neural models for node embeddings, and scores actions based on embeddings.
result SymNet policies outperform random and state-of-the-art deep reactive policies on nine RDDL domains.
In this work we give a detailed description of Matthias Günther's proof of the Isometric Embedding Theorem of Riemannian manifolds. Subsequently we will use this method to show that it is possible to construct an isometric embedding of a geometric flow, for instance the Ricci-flow, into some Euclidean space.
Graph neural networks improve MIL performance without losing interpretability.
problem Learning bag-level labels from bags of instances.
method Proposed a GNN-based algorithm treating bags as graphs to learn embeddings and predict labels.
result Achieves state-of-the-art performance on MIL datasets.
Disk Embeddings tackle embedding DAGs with exponential growth.
problem Embedding DAGs with exponentially increasing ancestors and descendants.
method Disk Embeddings framework for quasi-metric spaces, including Hyperbolic Disk Embeddings.
result Disk Embeddings outperform existing methods in complex DAGs.
New method samples triplets from data distributions for training Triplet networks.
problem Training robust Triplet networks with discriminative triplets.
method Bayesian updating of multivariate normal distributions for dynamic class embedding sampling.
result Experimental validation on MNIST and histopathology CRC datasets shows effectiveness of the proposed method.
ReliefE ranks features faster and better in high-dimensional data.
problem Feature ranking in high-dimensional spaces.
method Adapting Relief algorithms to manifold embeddings.
result ReliefE outperforms traditional Relief algorithms in feature ranking.
We construct embedded minimal surfaces which are n-periodic in Rn. They are new for codimension n−2≥2. We start with a Jordan curve of edges of the n-dimensional cube. It bounds a Plateau minimal disk which Schwarz reflection extends to a complete minimal surface. Studying the group of Schwarz refl…
Low-dimensional embedding, manifold learning, clustering, classification, and anomaly detection are among the most important problems in machine learning. The existing methods usually consider the case when each instance has a fixed, finite-dimensional feature representation. Here we consider a different setting. We as…
This work embeds annotations into a multidimensional space to measure classification difficulty.
problem Uncertainty in machine learning models during annotation phase.
method Develops a Bayesian Dirichlet-Multinomial framework to embed annotations and uses stochastic Expectation Maximization with MCMC.
result Embeddings reflect semantic similarities of original classes, aiding in measuring classification difficulty.
Paper proposes FOFE for efficient WSD.
problem Word sense disambiguation (WSD) problem.
method Fixed-size ordinally forgetting encoding (FOFE) combined with FFNN.
result FOFE-based FFNN achieves comparable performance to state-of-the-art at lower cost.
New method prevents class collapse in metric learning with margin-based losses.
problem Class collapse in metric learning due to diverse intra-class samples.
method Proposed a sampling method to select nearest same-class samples as positive elements in tuple.
result Demonstrated clear benefits on various fine-grained image retrieval datasets.
Model learns and generalizes new concepts efficiently from few labeled instances.
problem Efficient continual learning of new concepts in AI.
method Develops a computational model inspired by learning theories, using embedding space and generative distribution.
result Model efficiently expands learned concepts to new domains using few labeled samples.
Using the Plucker map between grassmannians, we study basic aspects of classic grassmannian geometries. For `hyperbolic' grassmannian geometries, we prove some facts (for instance, that the Plucker map is a minimal isometric embedding) that were previously known in the `elliptic' case.
AMI-Net+ tackles medical diagnosis from incomplete, imbalanced data.
problem Medical diagnosis from incomplete and imbalanced data.
method AMI-Net+ uses multi-instance neural network with embedding, multi-head attention, and gated attention-based pooling. It also employs focal loss and self-adaptive multi-instance pooling.
result AMI-Net+ outperforms state-of-the-art models on real-world medical datasets.
This text is the extended version of a talk given at the conference Geometry, Topology, QFT and Cosmology hold from May 28 to May 30, 2008 at the Observatoire de Paris. Using exterior differential systems, I provide a positive answer to the generalized isometric embedding problem of vector bundles, and show how conserv…
We give results on when a finitely generated group has only indiscrete embeddings in SL(2,C), with particular reference to 3-manifold groups. For instance if we glue two copies of the figure 8 knot along its torus boundary then the fundamental group of the resulting closed 3-manifold sometimes embeds in SL(2,C) and som…
We define additional gradings on two generalisations of Khovanov homology (one due to the first author, the other due to the second), and use them to define invariants of various kinds of embeddings. These include invariants of links in thickened surfaces and of surfaces embedded in thickened 3-manifolds. In particul…
HCRL learns hierarchical embeddings from deep embeddings of hierarchy components.
problem Flat clustering limits cohesive instance relations in hierarchical data.
method Simultaneously optimizes representation learning and hierarchical clustering in the embedding space.
result HCRL achieves best hierarchical clustering and data reconstruction.
Adversarial framework enforces fairness constraints on graph embeddings.
problem Fairness constraints in graph embeddings, especially age and gender.
method Adversarial framework for compositional fairness constraints.
result Framework allows for flexible combinations of fairness constraints.
Word embeddings are a popular approach to unsupervised learning of word relationships that are widely used in natural language processing. In this article, we present a new set of embeddings for medical concepts learned using an extremely large collection of multimodal medical data. Leaning on recent theoretical insigh…
In this paper we consider a version of the zero-shot learning problem where seen class source and target domain data are provided. The goal during test-time is to accurately predict the class label of an unseen target domain instance based on revealed source domain side information (\eg attributes) for unseen classes. …
The paper explores trading off consistency and dimensionality in convex surrogates for multiclass classification.
problem Designing consistent surrogate losses for multiclass classification with high-dimensional outcomes.
method Investigates embedding outcomes into convex polytopes and examining consistency under low-noise assumptions.
result Consistency can be achieved with less than n−1 dimensions, but hallucination occurs for some distributions. Paper proposes an unbiased classifier from triplet comparison data.
problem Learning a classifier from triplet comparison data.
method Empirical risk minimization framework with an unbiased estimator.
result The proposed method achieves better performance than baseline methods.
We consider minimal surfaces M which are complete, embedded and have finite total curvature in R3, and bounded, entire solutions with finite Morse index of the Allen-Cahn equation Δu+f(u)=0inR3. Here f=−W′ with W bistable and balanced, for instance W(u)=41(1−u2)2. We assume that …
Paper proves families of singularities can be topologically trivialized.
problem Understanding behavior of singularities under small perturbations.
method Establishes sufficient conditions for embedded topological trivialization.
result New instances of topological stability, including μ-constant deformations. GNNs can learn multiple graph centrality measures from a single embedding.
problem Estimating network centrality measures from graph data.
method Training a GNN to refine a single set of multidimensional embeddings and decode them into multiple outputs.
result GNN achieves 89% accuracy on random instances with up to 128 vertices.
A new graph kernel uses Wasserstein distance for better graph comparison.
problem Graph kernels often discard valuable information and struggle with continuous attributes.
method Proposes a novel graph kernel using Wasserstein distance for node feature vector distributions.
result Improves prediction performance on graph classification tasks.
This paper finds a linear relationship between t-SNE perplexity and data set size.
problem Choosing the right perplexity for t-SNE embeddings.
method Analyzed the relationship between perplexity and data set size.
result Embeddings remain structurally consistent when perplexity is adjusted accordingly.
Introduces Wasserstein Transform for updating distance structures.
problem Enhancing features and denoising data sets.
method Representing data points as probability measures and updating distances via Wasserstein distance.
result GT is computationally cheaper and has a closed-form solution for ℓ2-Wasserstein distance. Network embedding converts network data into vectors for machine learning.
problem Machine learning tasks on networks require vector representations of nodes and links.
method Various unsupervised and supervised methods for converting network data into vectors.
result Network embedding methods aim to preserve network structure in learned feature representations.
Meta-SAGE improves deep RL scalability for CO tasks by adapting pre-trained models to larger-scale problems.
problem Improving scalability of deep reinforcement learning models for combinatorial optimization tasks.
method Meta-SAGE combines a scale meta-learner and scheduled adaptation with guided exploration to adjust model parameters for larger-scale problems.
result Meta-SAGE outperforms previous methods and significantly improves scalability in CO tasks.
The study proves the existence of free boundary minimal disks in convex regions.
problem Proving the existence of free boundary minimal disks in convex regions.
method Based on a multiplicity-one theorem for the free boundary Simon-Smith min-max theory.
result Existence of at least three embedded free boundary minimal disks in strictly convex domains with nonnegative Ricci curvature.
Multi-label classification has received considerable interest in recent years. Multi-label classifiers have to address many problems including: handling large-scale datasets with many instances and a large set of labels, compensating missing label assignments in the training set, considering correlations between labels…
New theorem shows embedding impossibility for certain 3-manifolds.
problem Embedding restrictions in 3-manifolds.
method Construction of hyperbolic 3-manifolds with complicated ideals.
result Existence of non-embeddable manifolds in certain 3-manifolds.
Network analysis of human brain connectivity is critically important for understanding brain function and disease states. Embedding a brain network as a whole graph instance into a meaningful low-dimensional representation can be used to investigate disease mechanisms and inform therapeutic interventions. Moreover, by …