arXiv research

Link condition for simplicial complexes to be CUB spaces.

Novel loss function improves few-shot classification performance.

A new DR formulation improves metric learning for faster and more stable performance.

We address the problem of learning fine-grained cross-modal representations. We propose an instance-based deep metric learning approach in joint visual and textual space. The key novelty of this paper is that it shows that using per-image semantic supervision leads to substantial improvement in zero-shot performance ov…

We introduce the isoperimetric loss as a regularization criterion for learning the map from a visual representation to a semantic embedding, to be used to transfer knowledge to unknown classes in a zero-shot learning setting. We use a pre-trained deep neural network model as a visual representation of image data, a Wor…

LDA improves image classification accuracy with fewer features.

Generative Latent Implicit Conditional Optimization (GLICO) learns from small samples.

Generating an image from its description is a challenging task worth solving because of its numerous practical applications ranging from image editing to virtual reality. All existing methods use one single caption to generate a plausible image. A single caption by itself, can be limited, and may not be able to capture…

Two things seem to be indisputable in the contemporary deep learning discourse: 1. The categorical cross-entropy loss after softmax activation is the method of choice for classification. 2. Training a CNN classifier from scratch on small datasets does not work well. In contrast to this, we show that the cosine loss fun…

New method prevents forgetting in continual learning tasks.

A proxy network for few-shot learning improves model generalization.

Incremental class learning, a scenario in continual learning context where classes and their training data are sequentially and disjointedly observed, challenges a problem widely known as catastrophic forgetting. In this work, we propose a novel incremental class learning method that can significantly reduce memory ove…

Synthesizing images or texts automatically is a useful research area in the artificial intelligence nowadays. Generative adversarial networks (GANs), which are proposed by Goodfellow in 2014, make this task to be done more efficiently by using deep neural networks. We consider generating corresponding images from an in…

Simple fine-tuning of GANs with frozen discriminator outperforms existing methods.

Zero-shot learning (ZSL) is concerned with the recognition of previously unseen classes. It relies on additional semantic knowledge for which a mapping can be learned with training examples of seen classes. While classical ZSL considers the recognition performance on unseen classes only, generalized zero-shot learning …

In recent years, deep metric learning has achieved promising results in learning high dimensional semantic feature embeddings where the spatial relationships of the feature vectors match the visual similarities of the images. Similarity search for images is performed by determining the vectors with the smallest distanc…

Humans are able to explain their reasoning. On the contrary, deep neural networks are not. This paper attempts to bridge this gap by introducing a new way to design interpretable neural networks for classification, inspired by physiological evidence of the human visual system's inner-workings. This paper proposes a neu…

Paper proposes a method to select base classes for few-shot learning.

New algorithm TURTLE outperforms MAML and meta-learner LSTM.

A key component of most neural network architectures is the use of normalization layers, such as Batch Normalization. Despite its common use and large utility in optimizing deep architectures, it has been challenging both to generically improve upon Batch Normalization and to understand the circumstances that lend them…

Paper tackles target shift in zero-shot learning using adversarial learning.

Few-shot learning algorithms aim to learn model parameters capable of adapting to unseen classes with the help of only a few labeled examples. A recent regularization technique - Manifold Mixup focuses on learning a general-purpose representation, robust to small changes in the data distribution. Since the goal of few-…

When we are faced with challenging image classification tasks, we often explain our reasoning by dissecting the image, and pointing out prototypical aspects of one class or another. The mounting evidence for each of the classes helps us make our final decision. In this work, we introduce a deep network architecture -- …

New Convolutional Unit improves Batch Whitening performance.

Gaussian Processes (GPs) are known to provide accurate predictions and uncertainty estimates even with small amounts of labeled data by capturing similarity between data points through their kernel function. However traditional GP kernels are not very effective at capturing similarity between high dimensional data poin…

Faster ZSL with continual learning and self-gating.

DRAGON improves learning for rare classes in unbalanced datasets using class descriptions.

TransMatch uses transfer learning to improve few-shot learning accuracy.

MAIN network learns attributes without unseen class attributes for faster, more adaptable ZSL.

Transformer-based method discovers objects from images without labels.

Paper proposes a new pipeline for few-shot classification using forget-update module and channel vector sequence.

Learning to classify unseen class samples at test time is popularly referred to as zero-shot learning (ZSL). If test samples can be from training (seen) as well as unseen classes, it is a more challenging problem due to the existence of strong bias towards seen classes. This problem is generally known as \emph{generali…

NACs learn modular neural architectures without domain knowledge.

AdamP optimizes momentum-based optimizers for scale-invariant weights, improving model performance.

Study of tangent spaces in diffeological spaces under Lie group actions.

(1,1) non-L-space knots are foliar in 3D space.

Universal spaces for finite topological spaces simplify shape descriptions.

The paper extends Stone duality to topological convexity spaces.

No Einstein hypersurfaces found in Damek-Ricci spaces.

The distance function $\varrho(p,q)$ (or $d(p,q)$) of a distance space (general metric space) is not differentiable in general. We investigate such distance spaces over $\mathbb R^n$, whose distance functions are differentiable like in case of Finsler spaces. These spaces have several good properties, yet they are no F…

New kernels defined for various spaces, including measures.

Metric spaces uniquely split into Hilbert and non-line-split parts.

In a paper (math.DG/0403528) we obtained explicit examples of Moishezon twistor spaces of some compact self-dual four-manifolds admitting a non-trivial Killing field, and also determined their moduli space. In this note we investigate minitwistor spaces associated to these twistor spaces. We determine their structure, …

Study geometry of tetrahedra in complex hyperbolic space and Hilbert spaces.

New curvature positivity helps classify spherical spaces and complex projective spaces.

Study of complete space-like self-expanders in Minkovski space.

The abstract discusses the linear and smooth structures of mapping spaces.

Study on convergence of transformed metric spaces as dimensions grow.