Ep. 23 - October 31, 2023

arXiv research summaries for Computation Vision and Pattern Recognition from October 31, 2023. Today's Themes (AI Generated) Improving generalizability of models to new datasets through techniques like style transfer data augmentation. Leveraging vision-language models like CLIP for continual learning and adapting them with adapters or prompt tuning. Using transformers for tasks like whole slide image captioning and sign language motion generation. Diffusion models for tasks like image reconstruction, video generation, and image animation. Analysis and improvements to NeRF for more stable and artifact-free novel view synthesis.