1 paper · 1 filter
Yihong Dong, Ge Li, Yongding Tao +7
Despite the remarkable successes of general-purpose neural networks, such as MLPs and Transformers, we find that they exhibit notable shortcomings in modeling and reasoning about p…