What is your opinion on the current state of basic research in China? China's scientific output has increased dramatically since 2015, when the ResNet paper was published.No, Chinese firms simply underinvest in pure NN research, they mostly focus on applied stuff and short-term results. For example, Microsoft Research Asia is based in Beijing yet produced many ground-breaking works - ResNet (most cited DL paper), Swin Transformers, etc. They also created LightGBM which is one of the most popular gradient boosting libraries now. So yeah, this shows that with enough investment Chinese researchers can and will create strong research papers.
One day before the US elections, MSS uses it to release a wild video of Biden confessing about doing all sorts of bad things....People are so going to use this for fake news.
If we are talking strictly about breakthrough/fundamental papers, then I think it is more or less the same. Primarily because there're very few such papers in total, regardless of the country. However, the overall quality of AI research in China has definitely increased - there're far more real innovations in terms of model architectures, training methods, etc. instead of just re-applying existing models to some new dataset.What is your opinion on the current state of basic research in China? China's scientific output has increased dramatically since 2015, when the ResNet paper was published.