Multiobjective distribution matching

Authors: Xiaoyuan Zhang, Peijie Li, Yingying Yu, Yichi Zhang, Han Zhao, Qingfu Zhang

ICML 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on real-world images demonstrate that both algorithms can generate high-quality interpolated images across multiple distributions.
Researcher Affiliation Academia 1Department of Computer Science, City UHK. 2Department of Mathematics, HKU. 3Department of Statistics, IU. 4Department of Computer Science, UIUC. Correspondence to: Qingfu Zhang <EMAIL>.
Pseudocode Yes Algorithm 1 MODM on the Preference Simplex. Algorithm 2 Multiobjective VAE (MOVAE)
Open Source Code No The paper does not provide concrete access to source code. It describes algorithms and experimental results but does not include a repository link or an explicit statement about code release for the methodology described.
Open Datasets Yes We evaluate MOVAE on the Quick Draw dataset
Dataset Splits No The paper mentions "Number of training images is around 12K" but does not specify any splits for validation, test sets, or a detailed splitting methodology.
Hardware Specification No The paper does not provide specific hardware details. It mentions image size, network parameters, and optimizer, but no GPU/CPU models or other computer specifications used for experiments.
Software Dependencies No The paper mentions "The optimizer is Adam with a learning rate of 3e-5." but does not provide any specific software versions for libraries, frameworks, or programming languages.
Experiment Setup Yes Both the encoder and decoder networks have around 157K parameters. Number of training images is around 12K. The optimizer is Adam with a learning rate of 3e-5.