Overview

Synthetic data generation studies how to create artificial examples that reproduce useful features of real data or provide exact, controllable labels. Common approaches include procedural simulators and domain randomization, generative models such as GANs, variational autoencoders and diffusion models, and large language models or agents that synthesize text, dialogues, and tool-use trajectories. These methods are used to address scarce or sensitive data, cover rare cases, and build controlled training and evaluation sets, with quality often measured by fidelity, privacy, downstream performance, and transfer to real benchmarks. A central open tension is that realism or strong benchmark scores do not guarantee representative, unbiased, causally valid data or reliable performance in real settings.

Latest papers

661 papers40 papers in the last four weeks, down 45% on the four weeks before. 0.3% of all new papers.

Dec 17, 2021

Generation of data on discontinuous manifolds via continuous stochastic non-invertible networks

The generation of discontinuous distributions is a difficult task for most known frameworks such as generative autoencoders and generative adversarial networks. Generative non-invertible models are unable to accurately generate such distributions, require long training and often are subject to mode collapse. Variational autoencoders (VAEs), which are based on the idea of keeping the latent space to be Gaussian for the sake of a simple sampling, allow an accurate reconstruction, while they experience significant limitations at generation task. In this work, instead of trying to keep the latent space to be Gaussian, we use a pre-trained contrastive encoder to obtain a clustered latent space. Then, for each cluster, representing a unimodal submanifold, we train a dedicated low complexity network to generate this submanifold from the Gaussian distribution. The proposed framework is based on the information-theoretic formulation of mutual information maximization between the input data and latent space representation. We derive a link between the cost functions and the information-theoretic formulation. We apply our approach to synthetic 2D distributions to demonstrate both reconstruction and generation of discontinuous distributions using continuous stochastic networks.