1 paper · 1 filter
Yingpeng Du, Zhu Sun, Ziyan Wang +3
Large language models (LLMs) provide a promising way for accurate session-based recommendation (SBR), but they demand substantial computational time and memory. Knowledge distillat…