Hello everyone, In next week's reading discussion, we will talk about: Diffusion-LM Improves Controllable Text Generation https://arxiv.org/abs/2205.14217 This paper proposes to improve controllable text generation via a new language model based on continuous diffusions. I highly recommend you read the paper in advance and think about: 1. What's the diffusion model? 2. What method does Diffusion-LM use to control the generation of text? 3. Compared to other works on controllable text generation, what is the main advantage of this work? Any interesting idea we can derive from this work? 4. What are the limitations and remaining challenges of this work? 5. Any idea to address the remaining challenges? We will hold the discussion on Monday 4-5 pm in Torgersen 3160A. You may also join via Zoom https://virginiatech.zoom.us/j/5961845863. Hope everyone has a great weekend. Happy reading! Best, Ying