Variational Autoencoders for Text generation has been developed lately and still in progress. On of the problem with RNN based decoders is the posterior collapse problem. To solve this, we take a unique combination of auto and non-autoregressive generation. Read more here...
Efforts have been put for improving image generation especially on conditional generation, where we have control over the generated images. To highlight this, we worked on generation of facial images from the Celeb A dataset. Read more here...
This work was done while I was interning at Deakin university under supervision of Prof. Jinho Choi. The Concept of transferring knowledge between machine learning models came into limelight in 2015 (Hinton et. al). The distillation scenario often consists of a teacher (well trained and generally complex model) and student (less complex and limitedly trained). There can be multiple of them. Hinton et. al focused on improving the student model by a combination of knowledge transfer between the ground truths and predictions of the teacher. However, in reality there might be cases when ground truths (labels) may not be available. In such cases we solely need to rely on the predictions of the teacher. Till date limited research has been done specific to this. Read more about this here...