In a matter that is analog to the study of natural image statistics, we study the natural statistics of the deep neural network activations at various layers. As we show, these statistics, similar to image statistics, follow a power law. We also show, both analytically and empirically, that with depth the exponent of this power law increases at a linear rate. As a direct implication of our discoveries, we present a method for performing Knowledge Distillation (KD). While classical KD methods consider the logits of the teacher network, more recent methods obtain a leap in performance by considering the activation maps. This, however, uses metrics that are suitable for comparing images. We propose to employ two additional loss terms that are based on the spectral properties of the intermediate activation maps. The proposed method obtains state of the art results on multiple image recognition KD benchmarks.
We consider the task of upscaling a low resolution thumbnail image of a person, to a higher resolution image, which preserves the person's identity and other attributes. Since the thumbnail image is of low resolution, many higher resolution versions exist. Previous approaches produce solutions where the person's identity is not preserved, or biased solutions, such as predominantly Caucasian faces. We address the existing ambiguity by first augmenting the feature extractor to better capture facial identity, facial attributes (such as smiling or not) and race, and second, use this feature extractor to generate high-resolution images which are identity preserving as well as conditioned on race and facial attributes. Our results indicate an improvement in face similarity recognition and lookalike generation as well as in the ability to generate higher resolution images which preserve an input thumbnail identity and whose race and attributes are maintained.
We present deep learning methods for the design of arrays and single instances of small antennas. Each design instance is conditioned on a target radiation pattern and is required to conform to specific spatial dimensions and to include, as part of its metallic structure, a set of predetermined locations. The solution, in the case of a single antenna, is based on a composite neural network that combines a simulation network, a hypernetwork, and a refinement network. In the design of the antenna array, we add an additional design level and employ a hypernetwork within a hypernetwork. The learning objective is based on measuring the similarity of the obtained radiation pattern to the desired one. Our experiments demonstrate that our approach is able to design novel antennas and antenna arrays that are compliant with the design requirements, considerably better than the baseline methods. We compare the solutions obtained by our method to existing designs and demonstrate a high level of overlap. When designing the antenna array of a cellular phone, the obtained solution displays improved properties over the existing one.
Anomaly detection, the task of identifying unusual samples in data, often relies on a large set of training samples. In this work, we consider the setting of few-shot anomaly detection in images, where only a few images are given at training. We devise a hierarchical generative model that captures the multi-scale patch distribution of each training image. We further enhance the representation of our model by using image transformations and optimize scale-specific patch-discriminators to distinguish between real and fake patches of the image, as well as between different transformations applied to those patches. The anomaly score is obtained by aggregating the patch-based votes of the correct transformation across scales and image regions. We demonstrate the superiority of our method on both the one-shot and few-shot settings, on the datasets of Paris, CIFAR10, MNIST and FashionMNIST as well as in the setting of defect detection on MVTec. In all cases, our method outperforms the recent baseline methods.
Single channel speech separation has experienced great progress in the last few years. However, training neural speech separation for a large number of speakers (e.g., more than 10 speakers) is out of reach for the current methods, which rely on the Permutation Invariant Loss (PIT). In this work, we present a permutation invariant training that employs the Hungarian algorithm in order to train with an $O(C^3)$ time complexity, where $C$ is the number of speakers, in comparison to $O(C!)$ of PIT based methods. Furthermore, we present a modified architecture that can handle the increased number of speakers. Our approach separates up to $20$ speakers and improves the previous results for large $C$ by a wide margin.
Generative diffusion models have emerged as leading models in speech and image generation. However, in order to perform well with a small number of denoising steps, a costly tuning of the set of noise parameters is needed. In this work, we present a simple and versatile learning scheme that can step-by-step adjust those noise parameters, for any given number of steps, while the previous work needs to retune for each number separately. Furthermore, without modifying the weights of the diffusion model, we are able to significantly improve the synthesis results, for a small number of steps. Our approach comes at a negligible computation cost.
Recent accelerated MRI reconstruction models have used Deep Neural Networks (DNNs) to reconstruct relatively high-quality images from highly undersampled k-space data, enabling much faster MRI scanning. However, these techniques sometimes struggle to reconstruct sharp images that preserve fine detail while maintaining a natural appearance. In this work, we enhance the image quality by using a Conditional Wasserstein Generative Adversarial Network combined with a novel Adaptive Gradient Balancing (AGB) technique that automates the process of combining the adversarial and pixel-wise terms and streamlines hyperparameter tuning. In addition, we introduce a Densely Connected Iterative Network, which is an undersampled MRI reconstruction network that utilizes dense connections. In MRI, our method minimizes artifacts, while maintaining a high-quality reconstruction that produces sharper images than other techniques. To demonstrate the general nature of our method, it is further evaluated on a battery of image-to-image translation experiments, demonstrating an ability to recover from sub-optimal weighting in multi-term adversarial training.
Transformers are increasingly dominating multi-modal reasoning tasks, such as visual question answering, achieving state-of-the-art results thanks to their ability to contextualize information using the self-attention and co-attention mechanisms. These attention modules also play a role in other computer vision tasks including object detection and image segmentation. Unlike Transformers that only use self-attention, Transformers with co-attention require to consider multiple attention maps in parallel in order to highlight the information that is relevant to the prediction in the model's input. In this work, we propose the first method to explain prediction by any Transformer-based architecture, including bi-modal Transformers and Transformers with co-attentions. We provide generic solutions and apply these to the three most commonly used of these architectures: (i) pure self-attention, (ii) self-attention combined with co-attention, and (iii) encoder-decoder attention. We show that our method is superior to all existing methods which are adapted from single modality explainability.
Radiologist examination of chest CT is an effective way for screening COVID-19 cases. In this work, we overcome three challenges in the automation of this process: (i) the limited number of supervised positive cases, (ii) the lack of region-based supervision, and (iii) the variability across acquisition sites. These challenges are met by incorporating a recent augmentation solution called SnapMix, by a new patch embedding technique, and by performing a test-time stability analysis. The three techniques are complementary and are all based on utilizing the heatmaps produced by the Class Activation Mapping (CAM) explainability method. Compared to the current state of the art, we obtain an increase of five percent in the F1 score on a site with a relatively high number of cases, and a gap twice as large for a site with much fewer training images.
We consider the problem of supervised classification, such that the features that the network extracts match an unseen set of semantic attributes, without any additional supervision. For example, when learning to classify images of birds into species, we would like to observe the emergence of features that zoologists use to classify birds. We propose training a neural network with discrete top-level activations, which is followed by a multi-layered perceptron (MLP) and a parallel decision tree. We present a theoretical analysis as well as a practical method for learning in the intersection of two hypothesis classes. Since real-world features are often sparse, a randomized sparsity regularization is also applied. Our results on multiple benchmarks show an improved ability to extract a set of features that are highly correlated with the set of unseen attributes.