Album cover art image generation with Generative Adversarial Networks
Authors:
Felipe Perez Stoppa,
Ester Vidaña-Vila,
Joan Navarro
Abstract:
Generative Adversarial Networks (GANs) were introduced by Goodfellow in 2014, and since then have become popular for constructing generative artificial intelligence models. However, the drawbacks of such networks are numerous, like their longer training times, their sensitivity to hyperparameter tuning, several types of loss and optimization functions and other difficulties like mode collapse. Cur…
▽ More
Generative Adversarial Networks (GANs) were introduced by Goodfellow in 2014, and since then have become popular for constructing generative artificial intelligence models. However, the drawbacks of such networks are numerous, like their longer training times, their sensitivity to hyperparameter tuning, several types of loss and optimization functions and other difficulties like mode collapse. Current applications of GANs include generating photo-realistic human faces, animals and objects. However, I wanted to explore the artistic ability of GANs in more detail, by using existing models and learning from them. This dissertation covers the basics of neural networks and works its way up to the particular aspects of GANs, together with experimentation and modification of existing available models, from least complex to most. The intention is to see if state of the art GANs (specifically StyleGAN2) can generate album art covers and if it is possible to tailor them by genre. This was attempted by first familiarizing myself with 3 existing GANs architectures, including the state of the art StyleGAN2. The StyleGAN2 code was used to train a model with a dataset containing 80K album cover images, then used to style images by picking curated images and mixing their styles.
△ Less
Submitted 9 December, 2022;
originally announced December 2022.
High-resolution millimeter-wave spectroscopy of CH$_2$DCl: paving the way for future astronomical observations of chloromethane isotopologues
Authors:
M. Melosso,
A. Achilli,
F. Tamassia,
E. Canè,
A. Pietropolli Charmet,
P. Stoppa,
L. Dore
Abstract:
Chloromethane is the only organochloride detected in space to date. Its recent observation towards the low-mass protostar IRAS 16293-2422 with ALMA offers a prompt for new laboratory studies of CH$_3$Cl and its isotopologues. Here, we report the investigation of the rotational spectrum of monodeuterated chloromethane CH$_2$DCl in the frequency region between 90 and 300 GHz. The measurements have b…
▽ More
Chloromethane is the only organochloride detected in space to date. Its recent observation towards the low-mass protostar IRAS 16293-2422 with ALMA offers a prompt for new laboratory studies of CH$_3$Cl and its isotopologues. Here, we report the investigation of the rotational spectrum of monodeuterated chloromethane CH$_2$DCl in the frequency region between 90 and 300 GHz. The measurements have been carried out with a frequency-modulation millimeter-wave spectrometer, arranged to perform saturation spectroscopy. From the analysis of hyperfine-resolved spectra of the two chlorine isotopologues CH$_2$D$^{35}$Cl and CH$_2$D$^{37}$Cl, consistent sets of accurate spectroscopic parameters have been obtained. This work provides reliable spectral predictions which can be used to guide radio-astronomical searches of CH$_2$DCl in the interstellar medium and represents a solid base for future analyses of high-resolution infrared spectra of monodeuterated chloromethane.
△ Less
Submitted 24 June, 2020;
originally announced June 2020.
