Search Results for author:
Found 25 papers, 11 papers with code
On genuine invariance learning without weight-tying
We demonstrate that even when a network learns to correctly classify samples on a group orbit, the underlying decision-making in such a model does not attain genuine invariance.
Learning to Summarize Videos by Contrasting Clips
Video summarization aims at choosing parts of a video that narrate a story as close as possible to the original one.
LieGG: Studying Learned Lie Group Generators
Symmetries built into a neural network have appeared to be very beneficial for a wide range of tasks as it saves the data to learn them.
Contrasting quadratic assignments for set-based representation learning
The views are ordered in pairs, such that they are either positive, encoding different views of the same object, or negative, corresponding to views of different objects.
Human-Object Interaction Detection via Weak Supervision
iii) We evaluate Align-Former on HICO-DET [5] and V-COCO [13], and show that Align-Former outperforms existing image-level supervised HO-I detectors by a large margin (4. 71% mAP improvement from 16. 14% to 20. 85% on HICO-DET [5]).
Wiggling Weights to Improve the Robustness of Classifiers
To demonstrate that wiggling the weights consistently improves classification, we choose a standard network and modify it to a transform-augmented network.
PIE: Pseudo-Invertible Encoder
We consider the problem of information compression from high dimensional data.
How to Transform Kernels for Scale-Convolutions
We aim for accurate scale-equivariant convolutional neural networks (SE-CNNs) applicable for problems where high granularity of scale and small kernel sizes are required.
Built-in Elastic Transformations for Improved Robustness
We focus on building robustness in the convolutions of neural visual classifiers, especially against natural perturbations like elastic deformations, occlusions and Gaussian noise.
DISCO: accurate Discrete Scale Convolutions
In recent work scale equivariance was added to convolutional neural networks.
Natural Perturbed Training for General Robustness of Neural Network Classifiers
For Cifar-10 and STL-10 natural perturbed training even improves the accuracy for clean data and reaches the state of the art performance.
Self-Selective Context for Interaction Recognition
Our experiments show that SSC leads to an important increase in interaction recognition performance, while using much fewer parameters.
Adversarial and Natural Perturbations for General Robustness
In this paper we aim to explore the general robustness of neural network classifiers by utilizing adversarial as well as natural perturbations.
Scale Equivariance Improves Siamese Tracking
We develop the theory for scale-equivariant Siamese trackers, and provide a simple recipe for how to make a wide range of existing trackers scale-equivariant.
Ranked #1 on
Visual Object Tracking
on OTB-2013
Diagnosing Rarity in Human-Object Interaction Detection
To that end, in this paper, we propose to diagnose rarity in HOI detection.
Explaining with Counter Visual Attributes and Examples
Hence, inspired by the way of human explanations in this paper we provide attribute-based and example-based explanations.
Understanding Misclassifications by Attributes
We study the changes in attributes for clean as well as adversarial images in both standard and adversarially robust networks.
Scale-Equivariant Steerable Networks
The effectiveness of Convolutional Neural Networks (CNNs) has been substantially attributed to their built-in property of translation equivariance.
Ranked #33 on
Image Classification
on STL-10
Interpreting Adversarial Examples with Attributes
Deep computer vision systems being vulnerable to imperceptible and carefully crafted noise have raised questions regarding the robustness of their decisions.
Long-term Tracking in the Wild: A Benchmark
We introduce the OxUvA dataset and benchmark for evaluating single-object tracking algorithms.
i-RevNet: Deep Invertible Networks
An analysis of i-RevNets learned representations suggests an alternative explanation for the success of deep networks by a progressive contraction and linear separation with depth.
A Biologically Plausible Model for Rapid Natural Scene Identification
From these neural measurements and the contrast statistics of the natural image stimuli, we derive an across subject Weibull response model.
The Distribution Family of Similarity Distances
This fundamental insight opens new directions in the assessment of feature similarity, with projected improvements in object and scene recognition algorithms.
Predicting Brain States from fMRI Data: Incremental Functional Principal Component Regression
We propose a method for reconstruction of human brain states directly from functional neuroimaging data.
