Semi-supervised Learning by Latent Space Energy-Based Model of Symbol-Vector Coupling
This addresses semi-supervised learning problems for researchers and practitioners, but appears incremental as it builds on existing energy-based and latent space approaches.
The paper tackles semi-supervised learning by proposing a latent space energy-based prior model that couples latent and symbolic vectors, enabling classification from inferred latent vectors. The method performs well across image, text, and tabular data domains.
This paper proposes a latent space energy-based prior model for semi-supervised learning. The model stands on a generator network that maps a latent vector to the observed example. The energy term of the prior model couples the latent vector and a symbolic one-hot vector, so that classification can be based on the latent vector inferred from the observed example. In our learning method, the symbol-vector coupling, the generator network and the inference network are learned jointly. Our method is applicable to semi-supervised learning in various data domains such as image, text, and tabular data. Our experiments demonstrate that our method performs well on semi-supervised learning tasks.