Naoya Muramatsu's Highlights on 'Model compression and optimization: Why think bigger when you can think smaller?' | Glasp