arXiv research

This paper shows how optimizing with hard negative examples improves image retrieval.

New unsupervised method selects hard negative samples for contrastive learning.

WEINCE improves contrastive learning by correcting softmax biases.

Proposes a novel method for generating hard negatives near time series data boundaries.

GCBS improves contrastive learning performance efficiently.

Unified framework for hard affine SDP constraints in vRKHSs.

Three hard diagrams of the unknot require extra crossings to simplify.

Detecting adversarial examples is as hard as classifying them.

This work connects hardness of approximation and learning.

Developed a new thresholding method that connects soft and hard thresholding.

The Hard Lefschetz Theorem extends to certain Kähler Lie Algebroids with ellipticity.

Many graph clustering quality functions suffer from a resolution limit, the inability to find small clusters in large graphs. So called resolution-limit-free quality functions do not have this limit. This property was previously introduced for hard clustering, that is, graph partitioning. We investigate the resolution-…

We describe a method for generating minimal hard prime surface-link diagrams. We extend the known examples of minimal hard prime classical unknot and unlink diagrams up to three components and generate figures of all minimal hard prime surface-unknot and surface-unlink diagrams with prime base surface components up to …

Hardness proven for neural networks with natural weights.

ELF improves long-tailed classification by focusing on hard examples.

We analyze the semi-hard triplet loss using Edgeworth expansion for better understanding of its behavior.

Noise makes learning linear thresholds hard, but algorithms can still learn near-optimal thresholds.

SANS uses graph structure to find meaningful negatives for entity and relation embeddings.

It is well known that in stochastic multi-armed bandits (MAB), the sample mean of an arm is typically not an unbiased estimator of its true mean. In this paper, we decouple three different sources of this selection bias: adaptive \emph{sampling} of arms, adaptive \emph{stopping} of the experiment, and adaptively \emph{…

Single sample estimation for hard-constrained models like SAT and coloring problems.

Paper shows examples of almost Kähler manifolds satisfying Hard Lefschetz but not Betti-Hodge equality.

New findings show learning deeper neural networks is hard even with Gaussian inputs and non-degenerate weights.

Most speech recognition tasks pertain to mapping words across two modalities: acoustic and orthographic. In this work, we suggest learning encoders that map variable-length, acoustic or phonetic, sequences that represent words into fixed-dimensional vectors in a shared latent space; such that the distance between two w…

Study categorizes knots and links as rigid or shaky based on Reidemeister moves.

Machine learning has become one of the main components for task automation in many application domains. Despite the advancements and impressive achievements of machine learning, it has been shown that learning algorithms can be compromised by attackers both at training and test time. Machine learning systems are especi…

Efficient method for generating adversarial examples with limited query budget.

Proposes TFCL to mitigate negative transfer in MTL by collaborating across features and tasks.

The paper proves a generalized Lefschetz duality for a specific type of manifold.

The study investigates the impact of negative examples in contrastive learning.

There is a growing body of literature showing that deep neural networks are vulnerable to adversarial input modification. Recently this work has been extended from image classification to malware classification over boolean features. In this paper we present several new methods for training restricted networks in this …

Paper tackles hard shape constraints in kernel machines.

We empirically investigate the (negative) expected accuracy as an alternative loss function to cross entropy (negative log likelihood) for classification tasks. Coupled with softmax activation, it has small derivatives over most of its domain, and is therefore hard to optimize. A modified, leaky version is evaluated on…

Gradient descent struggles to learn equivariant neural networks, even with symmetries.

In this paper, we provide new complexity results for algorithms that learn discrete-variable Bayesian networks from data. Our results apply whenever the learning algorithm uses a scoring criterion that favors the simplest model able to represent the generative distribution exactly. Our results therefore hold whenever t…

Recent convolutional neural networks (CNNs) have led to impressive performance but often suffer from poor calibration. They tend to be overconfident, with the model confidence not always reflecting the underlying true ambiguity and hardness. In this paper, we propose angular visual hardness (AVH), a score given by the …

In this article supervised learning problems are solved using soft rule ensembles. We first review the importance sampling learning ensembles (ISLE) approach that is useful for generating hard rules. The soft rules are then obtained with logistic regression from the corresponding hard rules. In order to deal with the p…

New symplectic structures found on complex manifolds without Kähler structures.

Random forest is widely exploited as an ensemble learning method. In many practical applications, however, there is still a significant challenge to learn from imbalanced data. To alleviate this limitation, we propose a deep dynamic boosted forest (DDBF), a novel ensemble algorithm that incorporates the notion of hard …

For a Lie group $G=\R^{n}\ltimes_φ\R^{m}$ with the semi-simple action $φ:\R^{n}\to {\rm Aut}(\R^{m})$, we show that if $Γ$ is a finite extension of a lattice of $G$ then $K(Γ, 1)$ is formal. Moreover we show that a compact symplectic aspherical manifold with the fundamental group $Γ$ satisfies the hard Lefschetz proper…

Hard instances, which require a long time for a specific algorithm to solve, help (1) analyze the algorithm for accelerating it and (2) build a good benchmark for evaluating the performance of algorithms. There exist several efforts for automatic generation of hard instances. For example, evolutionary algorithms have b…

Using the Hard Lefschetz Theorem for Sasakian manifolds, we find two examples of compact K-contact nilmanifolds with no compatible Sasakian metric in dimensions five and seven, respectively

This work tackles learning Markov games with adversarial opponents and achieves both average reward and exploitation.

It is becoming increasingly important to understand the vulnerability of machine learning models to adversarial attacks. In this paper we study the feasibility of robust learning from the perspective of computational learning theory, considering both sample and computational complexity. In particular, our definition of…

Floer homology is a good example of homological invariants living in the infinite dimension. We suggest a way to construct this kind of invariants using only soft essentially finite-dimensional tools; no hard analysis or PDE is involved. This work is partially inspired by the M. Gromov's survey ``Soft and hard symplect…

Optimizes classifiers for varying levels of automation.

Zero-Shot learning has been shown to be an efficient strategy for domain adaptation. In this context, this paper builds on the recent work of Bucher et al. [1], which proposed an approach to solve Zero-Shot classification problems (ZSC) by introducing a novel metric learning based objective function. This objective fun…

Study reward-free RL in non-linear settings, improving efficiency and removing assumptions.

Study online control of unknown time-varying systems with negative and positive results.