137 citations · 227 across the 23 of their papers we have counts for
23 papers
D3: Data Diversity Design for Systematic Generalization in Visual Question Answering
Amir Rahimi, Vanessa D'Amario, Moyuru Yamada +3
Systematic generalization is a crucial aspect of intelligence, which refers to the ability to generalize to novel tasks by combining known subtasks and concepts. One critical facto…
Modularity Trumps Invariance for Compositional Robustness
Ian Mason, Anirban Sarkar, Tomotake Sasaki +1
By default neural networks are not robust to changes in data distribution. This has been demonstrated with simple image corruptions, such as blurring or adding noise, degrading ima…
Deephys: Deep Electrophysiology, Debugging Neural Networks under Distribution Shifts
Anirban Sarkar, Matthew Groth, Ian Mason +2
Deep Neural Networks (DNNs) often fail in out-of-distribution scenarios. In this paper, we introduce a tool to visualize and understand such failures. We draw inspiration from conc…
Do Neural Networks for Segmentation Understand Insideness?
Kimberly Villalobos, Vilim Štih, Amineh Ahmadinejad +6
The insideness problem is an aspect of image segmentation that consists of determining which pixels are inside and outside a region. Deep Neural Networks (DNNs) excel in segmentati…
Symmetry Perception by Deep Networks: Inadequacy of Feed-Forward Architectures and Improvements with Recurrent Connections
Shobhita Sundaram, Darius Sinha, Matthew Groth +2
Symmetry is omnipresent in nature and perceived by the visual system of many species, as it facilitates detecting ecologically important classes of objects in our environment. Symm…
Transformer Module Networks for Systematic Generalization in Visual Question Answering
Moyuru Yamada, Vanessa D'Amario, Kentaro Takemoto +2
Transformers achieve great performance on Visual Question Answering (VQA). However, their systematic generalization capabilities, i.e., handling novel combinations of known concept…
Robust Upper Bounds for Adversarial Training
Dimitris Bertsimas, Xavier Boix, Kimberly Villalobos Carballo +1
Many state-of-the-art adversarial training methods for deep learning leverage upper bounds of the adversarial loss to provide security guarantees against adversarial attacks. Yet,…
Three approaches to facilitate DNN generalization to objects in out-of-distribution orientations and illuminations
Akira Sakai, Taro Sunagawa, Spandan Madan +7
The training data distribution is often biased towards objects in certain orientations and illumination conditions. While humans have a remarkable capability of recognizing objects…
Emergent Neural Network Mechanisms for Generalization to Objects in Novel Orientations
Avi Cooper, Xavier Boix, Daniel Harari +4
The capability of Deep Neural Networks (DNNs) to recognize objects in orientations outside the distribution of the training data is not well understood. We present evidence that DN…
The Foes of Neural Network's Data Efficiency Among Unnecessary Input Dimensions
Vanessa D'Amario, Sanjana Srivastava, Tomotake Sasaki +1
Datasets often contain input dimensions that are unnecessary to predict the output label, e.g. background in object recognition, which lead to more trainable parameters. Deep Neura…
In-distribution adversarial attacks on object recognition models using gradient-free search
Spandan Madan, Tomotake Sasaki, Hanspeter Pfister +2
Neural networks are susceptible to small perturbations in the form of 2D rotations and shifts, image crops, and even changes in object colors. Past works attribute these errors to…
How Modular Should Neural Module Networks Be for Systematic Generalization?
Vanessa D'Amario, Tomotake Sasaki, Xavier Boix
Neural Module Networks (NMNs) aim at Visual Question Answering (VQA) via composition of modules that tackle a sub-task. NMNs are a promising strategy to achieve systematic generali…
When and how CNNs generalize to out-of-distribution category-viewpoint combinations
Spandan Madan, Timothy Henry, Jamell Dozier +6
Object recognition and viewpoint estimation lie at the heart of visual understanding. Recent works suggest that convolutional neural networks (CNNs) fail to generalize to out-of-di…
Robustness to Transformations Across Categories: Is Robustness To Transformations Driven by Invariant Neural Representations?
Hojin Jang, Syed Suleman Abbas Zaidi, Xavier Boix +4
Deep Convolutional Neural Networks (DCNNs) have demonstrated impressive robustness to recognize objects under transformations (eg. blur or noise) when these transformations are inc…
Frivolous Units: Wider Networks Are Not Really That Wide
Stephen Casper, Xavier Boix, Vanessa D'Amario +4
A remarkable characteristic of overparameterized deep neural networks (DNNs) is that their accuracy does not degrade when the network's width is increased. Recent evidence suggests…
Minimal Images in Deep Neural Networks: Fragile Object Recognition in Natural Images
Sanjana Srivastava, Guy Ben-Yosef, Xavier Boix
The human ability to recognize objects is impaired when the object is not shown in full. "Minimal images" are the smallest regions of an image that remain recognizable for humans.…
Theory IIIb: Generalization in Deep Networks
Tomaso Poggio, Qianli Liao, Brando Miranda +3
A main puzzle of deep neural networks (DNNs) revolves around the apparent absence of "overfitting", defined in this paper as follows: the expected error does not get worse when inc…
Theory of Deep Learning III: explaining the non-overfitting puzzle
Tomaso Poggio, Kenji Kawaguchi, Qianli Liao +5
A main puzzle of deep networks revolves around the absence of overfitting despite large overparametrization and despite the large capacity demonstrated by zero training error on ra…
Herding Generalizes Diverse M -Best Solutions
Ece Ozkan, Gemma Roig, Orcun Goksel +1
We show that the algorithm to extract diverse M -solutions from a Conditional Random Field (called divMbest [1]) takes exactly the form of a Herding procedure [2], i.e. a determini…
Foveation-based Mechanisms Alleviate Adversarial Examples
Yan Luo, Xavier Boix, Gemma Roig +2
We show that adversarial examples, i.e., the visually imperceptible perturbations that result in Convolutional Neural Networks (CNNs) fail, can be alleviated with a mechanism based…
Comment on "Ensemble Projection for Semi-supervised Image Classification"
Xavier Boix, Gemma Roig, Luc Van Gool
In a series of papers by Dai and colleagues [1,2], a feature map (or kernel) was introduced for semi- and unsupervised learning. This feature map is build from the output of an ens…
SEEDS: Superpixels Extracted via Energy-Driven Sampling
Michael Van den Bergh, Xavier Boix, Gemma Roig +1
Superpixel algorithms aim to over-segment the image by grouping pixels that belong to the same object. Many state-of-the-art superpixel algorithms rely on minimizing objective func…
Random Binary Mappings for Kernel Learning and Efficient SVM
Gemma Roig, Xavier Boix, Luc Van Gool
Support Vector Machines (SVMs) are powerful learners that have led to state-of-the-art results in various computer vision problems. SVMs suffer from various drawbacks in terms of s…