Jie Fu

Latest

Defending Against Weight-Poisoning Backdoor Attacks for Parameter-Efficient Fine-Tuning
Exploring Clean Label Backdoor Attacks and Defense in Language Models
Beyond Fully-Connected Layers with Quaternions:Parameterization of Hypercomplex Multiplications with 1/n Parameters
Would you Rather? A New Benchmark for Learning Machine Alignment with Cultural Values and Social Preferences
Light and Efficient Neural Natural Language Processing with Quaternion Networks
Simple and Effective Curriculum Pointer-Generator Networks for Reading Comprehension over Long Narratives