arXiv research

Paper analyzes trade-offs in top-k classification accuracies and proposes a new loss function.

Top/O's first two k-invariants are zero.

Class ambiguity is typical in image classification problems with a large number of classes. When classes are difficult to discriminate, it makes sense to allow k guesses and evaluate classifiers based on the top-k error instead of the standard zero-one loss. We propose top-k multiclass SVM as a direct method to optimiz…

The top-$k$ error is often employed to evaluate performance for challenging classification tasks in computer vision as it is designed to compensate for ambiguity in ground truth labels. This practical success motivates our theoretical analysis of consistent top-$k$ classification. Surprisingly, it is not rigorously und…

Paper introduces a new loss function for deep imbalanced classification.

Unified model for prediction and deferral selects top-k entities efficiently.

Proposes top-label calibration and M2B framework for multiclass to binary calibration.

In order to push the performance on realistic computer vision tasks, the number of classes in modern benchmark datasets has significantly increased in recent years. This increase in the number of classes comes along with increased ambiguity between the class labels, raising the question if top-1 error is the right perf…

Smoothed top-k operator improves model training efficiency.

In this paper, we introduce a geometric structure called top, which is a trivialized bundle of plane pencils over a Riemannian 3-manifold, defined as the set of kernels of a circle of 1-forms (e.g. of contact and integrable forms) with particular properties with respect to the metric. We classify the manifolds which ad…

New algorithm reduces sample complexity for Top Two method.

Paper introduces efficient top-k selection with differential privacy.

Top-k error is currently a popular performance measure on large scale image classification benchmarks such as ImageNet and Places. Despite its wide acceptance, our understanding of this metric is limited as most of the previous research is focused on its special case, the top-1 error. In this work, we explore two direc…

The topological fundamental group $π_{1}^{top}$ is a topological invariant that assigns to each space a quasi-topological group and is discrete on spaces which are well behaved locally. For a totally path-disconnected, Hausdorff, unbased space $X$, we compute the topological fundamental group of the "hoop earring" spac…

Establishing criteria for top cell inertness in complexes.

It is well-known that classifiers are vulnerable to adversarial perturbations. To defend against adversarial perturbations, various certified robustness results have been derived. However, existing certified robustnesses are limited to top-1 predictions. In many real-world applications, top-$k$ predictions are more rel…

Extends homotopical theory to locally compact groups, refining their compactness properties.

For information retrieval and binary classification, we show that precision at the top (or precision at k) and recall at the top (or recall at k) are maximised by thresholding the posterior probability of the positive class. This finding is a consequence of a result on constrained minimisation of the cost-sensitive exp…

Motivated by applications in recommender systems, web search, social choice and crowdsourcing, we consider the problem of identifying the set of top $K$ items from noisy pairwise comparisons. In our setting, we are non-actively given $r$ pairwise comparisons between each pair of $n$ items, where each comparison has noi…

Top-two algorithm improved for best-k-arm selection.

Deep Neural Networks (DNNs) are vulnerable to adversarial attacks, especially white-box targeted attacks. One scheme of learning attacks is to design a proper adversarial objective function that leads to the imperceptible perturbation for any test image (e.g., the Carlini-Wagner (C&W) method). Most methods address targ…

We propose Top-N-Rank, a novel family of list-wise Learning-to-Rank models for reliably recommending the N top-ranked items. The proposed models optimize a variant of the widely used discounted cumulative gain (DCG) objective function which differs from DCG in two important aspects: (i) It limits the evaluation of DCG …

Paper characterizes minimax regret rates for online ranking with top-k feedback.

Top-H decoding improves text generation by balancing creativity and coherence.

Attention mechanism combines bottom-up and top-down signals in neural networks.

Paper introduces MPES for top-k ranking BO with preferential observations.

We present online boosting algorithms for multilabel ranking with top-k feedback, where the learner only receives information about the top k items from the ranking it provides. We propose a novel surrogate loss function and unbiased estimator, allowing weak learners to update themselves with limited information. Using…

Improved theoretical guarantees for Top Two algorithms.

Study improves top-k set prediction with low cardinality.

Unified framework for deferring queries to top-k experts, improving accuracy-cost trade-offs.

Paper recovers top-two answers and confusion probability in multi-choice crowdsourcing.

Paper extends top-k Mallows model for better user preference analysis.

Proposes differentiable and sparse top-k operators for neural networks.

This paper compares average-K and top-K classification methods under ambiguity.

Efficiently selects top-m designs for various contexts using sequential sampling.

Using the method of De Lellis-Topping, we prove some almost Schur type results. For example, one of our results gives a quantitative measure of how close the higher mean curvature of a submanifold is to its average value. We also derive another sharp Andrews-De Lellis-Topping type inequality involving the Riemannian cu…

In the top-down approach to multi-name credit modeling, calculation of singe name sensitivities appears possible, at least in principle, within the so-called random thinning (RT) procedure which dissects the portfolio risk into individual contributions. We make an attempt to construct a practical RT framework that enab…

Bottom-up algorithms outperform top-down in hierarchical community detection at intermediate levels.

Study on top-$k$ classification with new loss functions and algorithms.

This paper explores the preference-based top-$K$ rank aggregation problem. Suppose that a collection of items is repeatedly compared in pairs, and one wishes to recover a consistent ordering that emphasizes the top-$K$ ranked items, based on partially revealed preferences. We focus on the Bradley-Terry-Luce (BTL) model…

Optimizes identifying top-k items from comparisons with minimal comparisons.

New scoring rules compare probabilistic top lists in classification.

The paper proposes methods to identify and sample from mixtures of Mallows models for top-k rankings.

The paper addresses calibration in label ranking, a structured prediction task.

DeepTopPush improves accuracy at the top for complex classification tasks.

This paper tackles distributed estimation of the top-L eigenspace in PCA for large data sets.

The top word list, i.e., the top-M words with highest marginal probability in a given topic, is the standard topic representation in topic models. Most of recent automatical topic labeling algorithms and popular topic quality metrics are based on it. However, we find, empirically, words in this type of top word list ar…

We propose the Limited Multi-Label (LML) projection layer as a new primitive operation for end-to-end learning systems. The LML layer provides a probabilistic way of modeling multi-label predictions limited to having exactly k labels. We derive efficient forward and backward passes for this layer and show how the layer…