PluGeN: Multi-Label Conditional Generation from Pre-trained Models
Authors: Maciej Wołczyk, Magdalena Proszewska, Łukasz Maziarka, Maciej Zieba, Patryk Wielopolski, Rafał Kurczab, Marek Smieja8647-8656
AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments demonstrate that Plu Ge N preserves the quality of backbone models while adding the ability to control the values of labeled attributes. Implementation is available at https://github.com/gmum/plugen. Extensive experiments performed on the domain of images and a dataset of chemical compounds demonstrate that Plu Ge N is a reusable plugin that can be applied to various architectures including GANs and VAEs. |
| Researcher Affiliation | Collaboration | Maciej Wołczyk1*, Magdalena Proszewska1 , Łukasz Maziarka1, Maciej Zieba2,4, Patryk Wielopolski2, Rafał Kurczab3, Marek Smieja1 1Jagiellonian University 2Wroclaw University of Science and Technology 3 Institute of Pharmacology PAS 4 Tooploox |
| Pseudocode | No | No pseudocode or algorithm block found. |
| Open Source Code | Yes | Implementation is available at https://github.com/gmum/plugen. |
| Open Datasets | Yes | First, we consider the state-of-the-art Style GAN architecture (Karras, Laine, and Aila 2019), which was trained on Flickr-Faces-HQ (FFHQ) containing 70 000 high-quality images of resolution 1024 1024. We use Celeb A database, where every image of the size 256 256 is annotated with 40 binary labels. As a backbone model, we use Char VAE (G omez-Bombarelli et al. 2018), which is a type of recurrent network used for processing SMILES (Weininger 1988), a textual representation of molecules. It was trained on ZINC 250k database (Sterling and Irwin 2015) of commercially available chemical compounds. |
| Dataset Splits | No | No explicit training/validation/test split percentages or counts are provided in the main text. |
| Hardware Specification | No | No specific hardware details (e.g., GPU/CPU models, memory) used for running experiments are provided in the paper. |
| Software Dependencies | No | The paper mentions 'RDKit package' but does not provide specific version numbers for it or any other software dependencies like Python or deep learning frameworks. |
| Experiment Setup | No | The paper mentions some parameters for the probabilistic model, such as 'm0 = -1, m1 = 1', but states that 'the selection of σ0, σ1 will be discussed is the supplementary materials'. It also refers to details taken from other papers for backbone models, but lacks specific hyperparameters like learning rate, batch size, or optimizer settings in the main text. |