Mitigating over-exploration in latent space optimization using LES
This addresses a specific bottleneck in LSO for black-box discrete optimization, offering an incremental improvement with practical benefits.
The paper tackles the problem of over-exploration in Latent Space Optimization (LSO), which leads to unrealistic solutions, by developing Latent Exploration Score (LES) to mitigate this issue, resulting in enhanced solution quality across five benchmark tasks and twenty-two VAE models while maintaining high objective values.
We develop Latent Exploration Score (LES) to mitigate over-exploration in Latent Space Optimization (LSO), a popular method for solving black-box discrete optimization problems. LSO utilizes continuous optimization within the latent space of a Variational Autoencoder (VAE) and is known to be susceptible to over-exploration, which manifests in unrealistic solutions that reduce its practicality. LES leverages the trained decoder's approximation of the data distribution, and can be employed with any VAE decoder - including pretrained ones - without additional training, architectural changes or access to the training data. Our evaluation across five LSO benchmark tasks and twenty-two VAE models demonstrates that LES always enhances the quality of the solutions while maintaining high objective values, leading to improvements over existing solutions in most cases. We believe that new avenues to LSO will be opened by LES' ability to identify out of distribution areas, differentiability, and computational tractability. Open source code for LES is available at https://github.com/OmerRonen/les.