Generative Neural Fields by Mixtures of Neural Implicit Functions
This work addresses the need for efficient and scalable generative models across multiple data modalities, though it appears incremental as it builds on existing neural implicit and diffusion techniques.
The paper tackles the problem of efficiently generating diverse data types by proposing a method that uses mixtures of neural implicit functions to create generative neural fields, achieving competitive performance on benchmarks for images, voxels, and NeRF scenes without domain-specific designs.
We propose a novel approach to learning the generative neural fields represented by linear combinations of implicit basis networks. Our algorithm learns basis networks in the form of implicit neural representations and their coefficients in a latent space by either conducting meta-learning or adopting auto-decoding paradigms. The proposed method easily enlarges the capacity of generative neural fields by increasing the number of basis networks while maintaining the size of a network for inference to be small through their weighted model averaging. Consequently, sampling instances using the model is efficient in terms of latency and memory footprint. Moreover, we customize denoising diffusion probabilistic model for a target task to sample latent mixture coefficients, which allows our final model to generate unseen data effectively. Experiments show that our approach achieves competitive generation performance on diverse benchmarks for images, voxel data, and NeRF scenes without sophisticated designs for specific modalities and domains.