Modern metrics for generative learning like Fréchet Inception Distance (FID) and DINOv2-Fréchet Distance (FD-DINOv2) demonstrate impressive performance. However, they suffer from various shortcomings, like a bias towards specific generators and datasets. To address this problem, we propose the Fréchet Wavelet Distance (FWD) as a domain-agnostic metric based on the Wavelet Packet Transform ($\mathcal{W}_p$). FWD provides a sight across a broad spectrum of frequencies in images with a high resolution, preserving both spatial and textural aspects. Specifically, we use $\mathcal{W}_p$ to project generated and real images to the packet coefficient space. We then compute the Fréchet distance with the resultant coefficients to evaluate the quality of a generator. This metric is general-purpose and dataset-domain agnostic, as it does not rely on any pre-trained network, while being more interpretable due to its ability to compute Fréchet distance per packet, enhancing transparency. We conclude with an extensive evaluation of a wide variety of generators across various datasets that the proposed FWD can generalize and improve robustness to domain shifts and various corruptions compared to other metrics.

Modern methods for fine-tuning a Vision Transformer (ViT) like Low-Rank Adaptation (LoRA) and its variants demonstrate impressive performance. However, these methods ignore the high-dimensional nature of Multi-Head Attention (MHA) weight tensors. To address this limitation, we propose Canonical Rank Adaptation (CaRA). CaRA leverages tensor mathematics, first by tensorising the transformer into two different tensors; one for projection layers in MHA and the other for feed-forward layers. Second, the tensorised formulation is fine-tuned using the low-rank adaptation in Canonical-Polyadic Decomposition (CPD) form. Employing CaRA efficiently minimizes the number of trainable parameters. Experimentally, CaRA outperforms existing Parameter-Efficient Fine-Tuning (PEFT) methods in visual classification benchmarks such as Visual Task Adaptation Benchmark (VTAB)-1k and Fine-Grained Visual Categorization (FGVC).

Neural networks promise automated prostate segmentation for the development of precise and quantifiable image-based biomarkers in modern personalized oncology.Before clinical translation, however, their stability must be ensured.In this study, we train three-dimensional Ushaped convolutional neural networks to segment prostate magnetic resonance imaging (MRI) scans and evaluate different loss formulations to improve their performance.To evaluate generalizability and reproducibility of our networks, we compare their performance in a clinically acquired test/re-test MRI data set of 26 prostate cancer patients that was previously not seen by the networks.We find our networks to be generalizable with good reproducibility with a mean Intersection over Union of 0.88.While initial results are promising, anatomical accuracy remains to be evaluated in larger, multi-center data sets.To facilitate clinical applicability, we provide an easy to use toolbox online.* This work was partially funded by the German Federal Ministry of Education and Research (BMBF) within the "BNTrAInee" (16DHBK1022) and WestAI (01IS22094E WEST

Counting repetitive actions in long untrimmed videos is a challenging task that has many applications such as rehabilitation. State-of-the-art methods predict action counts by first generating a temporal self-similarity matrix (TSM) from the sampled frames and then feeding the matrix to a predictor network. The self-similarity matrix, however, is not an optimal input to a network since it discards too much information from the frame-wise embeddings. We thus rethink how a TSM can be utilized for counting repetitive actions and propose a framework that learns embeddings and predicts action start probabilities at full temporal resolution. The number of repeated actions is then inferred from the action start probabilities. In contrast to current approaches that have the TSM as an intermediate representation, we propose a novel loss based on a generated reference TSM, which enforces that the self-similarity of the learned frame-wise embeddings is consistent with the self-similarity of repeated actions. The proposed framework achieves state-of-the-art results on three datasets, i.e., RepCount, UCFRep, and Countix.

The fast wavelet transform is an essential workhorse in signal processing. Wavelets are local in the spatial- or temporal- and the frequency-domain. This property enables frequency domain analysis while preserving some spatiotemporal information. Until recently, wavelets rarely appeared in the machine learning literature. We provide the PyTorch Wavelet Toolbox to make wavelet methods more accessible to the deep learning community. Our PyTorch Wavelet Toolbox is well documented. A pip package is installable with `pip install ptwt`.

Abstract Recent advances in practical quantum computing have led to a variety of cloud-based quantum computing platforms that allow researchers to evaluate their algorithms on noisy intermediate-scale quantum devices. A common property of quantum computers is that they can exhibit instances of true randomness as opposed to pseudo-randomness obtained from classical systems. Investigating the effects of such true quantum randomness in the context of machine learning is appealing, and recent results vaguely suggest that benefits can indeed be achieved from the use of quantum random numbers. To shed some more light on this topic, we empirically study the effects of hardware-biased quantum random numbers on the initialization of artificial neural network weights in numerical experiments. We find no statistically significant difference in comparison with unbiased quantum random numbers as well as biased and unbiased random numbers from a classical pseudo-random number generator. The quantum random numbers for our experiments are obtained from real quantum hardware.

Variational quantum circuits build the foundation for various classes of quantum algorithms. In a nutshell, the weights of a parametrized quantum circuit are varied until the empirical sampling distribution of the circuit is sufficiently close to a desired outcome. Numerical first-order methods are applied frequently to fit the parameters of the circuit, but most of the time, the circuit itself, that is, the actual composition of gates, is fixed. Methods for optimizing the circuit design jointly with the weights have been proposed, but empirical results are rather scarce. Here, we consider a simple evolutionary strategy that addresses the trade-off between finding appropriate circuit ar-chitectures and parameter tuning. We evaluate our method both via simulation and on actual quantum hardware. Our benchmark problems include the transverse field Ising Hamiltonian and the Sherrington-Kirkpatrick spin model. Despite the shortcomings of current noisy intermediate-scale quantum hardware, we find only a minor slowdown on actual quantum machines compared to simulations. Moreover, we investigate which mutation operations most significantly contribute to the optimization. The results provide intuition on how randomized search heuristics behave on actual quantum hardware and layout a path for further refinement of evolutionary quantum gate circuits.

We introduce canonical weight normalization for convolutional neural networks. Inspired by the canonical tensor decomposition, we express the weight tensors in so-called canonical networks as scaled sums of outer vector products. In particular, we train network weights in the decomposed form, where scale weights are optimized separately for each mode. Additionally, similarly to weight normalization, we include a global scaling parameter. We study the initialization of the canonical form by running the power method and by drawing randomly from Gaussian or uniform distributions. Our results indicate that we can replace the power method with cheaper initializations drawn from standard distributions. The canonical re-parametrization leads to competitive normalization performance on the MNIST, CIFAR10, and SVHN data sets. Moreover, the formulation simplifies network compression. Once training has converged, the canonical form allows convenient model-compression by truncating the parameter sums.

Abstract As neural networks become able to generate realistic artificial images, they have the potential to improve movies, music, video games and make the internet an even more creative and inspiring place. Yet, the latest technology potentially enables new digital ways to lie. In response, the need for a diverse and reliable method toolbox arises to identify artificial images and other content. Previous work primarily relies on pixel-space convolutional neural networks or the Fourier transform. To the best of our knowledge, synthesized fake image analysis and detection methods based on a multi-scale wavelet-packet representation, localized in both space and frequency, have been absent thus far. The wavelet transform conserves spatial information to a degree, allowing us to present a new analysis. Comparing the wavelet coefficients of real and fake images allows interpretation. Significant differences are identified. Additionally, this paper proposes to learn a model for the detection of synthetic images based on the wavelet-packet representation of natural and generated images. Our forensic classifiers exhibit competitive or improved performance at small network sizes, as we demonstrate on the Flickr Faces High Quality, Large-scale Celeb Faces Attributes and Large-scale Scene UNderstanding source identification problems. Furthermore, we study the binary Face Forensics++ (ff++) fake-detection problem.

Convolutional neural networks (CNN)s have become the go-to choice for most image and video processing tasks. Most CNN architectures rely on pooling layers to reduce the resolution along spatial dimensions. The reduction allows subsequent deep convolution layers to operate with greater efficiency. This paper introduces adaptive wavelet pooling layers, which employ fast wavelet transforms (FWT) to reduce the feature resolution. The FWT decomposes the input features into multiple scales reducing the feature dimensions by removing the fine-scale subbands. Our approach adds extra flexibility through wavelet-basis function optimization and coefficient weighting at different scales. The adaptive wavelet layers integrate directly into well-known CNNs like the LeNet, Alexnet, or Densenet architectures. Using these networks, we validate our approach and find competitive performance on the MNIST, CIFAR10, and SVHN (street view house numbers) data-sets.

As service robots are entering more and more homes it gets evermore important to find behavior strategies that ensure a harmonic coexistence between those systems and their users. In this paper, we present a novel approach to enable a mobile robot to provide timely assistance to a user moving in its environment, while simultaneously avoiding unnecessary movements as well as interferences with the user. We developed a framework that uses information about the last object interaction to predict possible future movement destinations of the user and infer where they might need assistance based on prior knowledge. Given this prediction, the robot chooses the best position for itself that minimizes the time until assistance can be provided as well as avoids interferences with other activities of the user. We evaluated our approach in comparison to state-of-the-art methods in simulated environments and performed a user study in a virtual reality environment. Our evaluation demonstrates that our approach is able to decrease both the time until assistance is provided and the travel distance of the robot as well as increases the average distance between the user and the robot in comparison to state-of-the-art systems. Additionally, the robot behavior generated by our method is rated as more pleasant by our study participants than comparable literature approaches.

Complex numbers have long been favoured for digital signal processing, yet complex representations rarely appear in deep learning architectures. RNNs, widely used to process time series and sequence information, could greatly benefit from complex representations. We present a novel complex gated recurrent cell, which is a hybrid cell combining complex-valued and norm-preserving state transitions with a gating mechanism. The resulting RNN exhibits excellent stability and convergence properties and performs competitively on the synthetic memory and adding task, as well as on the real-world tasks of human motion prediction.