1 paper
Chen Wang, Fangxin Wang, Ruocheng Guo +2
In Sequential Recommendation Systems (SRecsys), traditional training approaches that rely on Cross-Entropy (CE) loss often prioritize accuracy but fail to align well with user sati…