2023-03-16 更新
Vector Quantized Time Series Generation with a Bidirectional Prior Model
Authors:Daesoo Lee, Sara Malacarne, Erlend Aune
Time series generation (TSG) studies have mainly focused on the use of Generative Adversarial Networks (GANs) combined with recurrent neural network (RNN) variants. However, the fundamental limitations and challenges of training GANs still remain. In addition, the RNN-family typically has difficulties with temporal consistency between distant timesteps. Motivated by the successes in the image generation (IMG) domain, we propose TimeVQVAE, the first work, to our knowledge, that uses vector quantization (VQ) techniques to address the TSG problem. Moreover, the priors of the discrete latent spaces are learned with bidirectional transformer models that can better capture global temporal consistency. We also propose VQ modeling in a time-frequency domain, separated into low-frequency (LF) and high-frequency (HF). This allows us to retain important characteristics of the time series and, in turn, generate new synthetic signals that are of better quality, with sharper changes in modularity, than its competing TSG methods. Our experimental evaluation is conducted on all datasets from the UCR archive, using well-established metrics in the IMG literature, such as Fr\’echet inception distance and inception scores. Our implementation on GitHub: \url{https://github.com/ML4ITS/TimeVQVAE}.
PDF accepted at AISTATS 2023
点此查看论文截图
Investigating GANsformer: A Replication Study of a State-of-the-Art Image Generation Model
Authors:Giorgia Adorni, Felix Boelter, Stefano Carlo Lambertenghi
The field of image generation through generative modelling is abundantly discussed nowadays. It can be used for various applications, such as up-scaling existing images, creating non-existing objects, such as interior design scenes, products or even human faces, and achieving transfer-learning processes. In this context, Generative Adversarial Networks (GANs) are a class of widely studied machine learning frameworks first appearing in the paper “Generative adversarial nets” by Goodfellow et al. that achieve the goal above. In our work, we reproduce and evaluate a novel variation of the original GAN network, the GANformer, proposed in “Generative Adversarial Transformers” by Hudson and Zitnick. This project aimed to recreate the methods presented in this paper to reproduce the original results and comment on the authors’ claims. Due to resources and time limitations, we had to constrain the network’s training times, dataset types, and sizes. Our research successfully recreated both variations of the proposed GANformer model and found differences between the authors’ and our results. Moreover, discrepancies between the publication methodology and the one implemented, made available in the code, allowed us to study two undisclosed variations of the presented procedures.
PDF
点此查看论文截图
Fashion-model pose recommendation and generation using Machine Learning
Authors:Vijitha Kannumuru, Santhosh Kannan S P, Krithiga Shankar, Joy Larnyoh, Rohith Mahadevan, Raja CSP Raman
Fashion-model pose is an important attribute in the fashion industry. Creative directors, modeling production houses, and top photographers always look for professional models able to pose. without the skill to correctly pose, their chances of landing professional modeling employment are regrettably quite little. There are occasions when models and photographers are unsure of the best pose to strike while taking photographs. This research concentrates on suggesting the fashion personnel a series of similar images based on the input image. The image is segmented into different parts and similar images are suggested for the user. This was achieved by calculating the color histogram of the input image and applying the same for all the images in the dataset and comparing the histograms. Synthetic images have become popular to avoid privacy concerns and to overcome the high cost of photoshoots. Hence, this paper also extends the work of generating synthetic images from the recommendation engine using styleGAN to an extent.
PDF
点此查看论文截图