BabyLlama-2: Ensemble-Distilled Models Consistently Outperform Teachers With Limited Databy Jean-Loup Tastet, Inar TimiryasovFirst submitted to…
Personalized Federated Learning via Backbone Self-Distillationby Pengju Wang, Bochao Liu, Dan Zeng, Chenggang Yan, Shiming…
Dynamic Integration of Task-Specific Adapters for Class Incremental Learningby Jiashuo Li, Shaokun Wang, Bo Qian,…
Federated Learning with Label-Masking Distillationby Jianghu Lu, Shikun Li, Kexin Bao, Pengju Wang, Zhenxing Qian,…
VCAT: Vulnerability-aware and Curiosity-driven Adversarial Training for Enhancing Autonomous Vehicle Robustnessby Xuan Cai, Zhiyong Cui,…
Privacy-Preserving Student Learning with Differentially Private Data-Free Distillationby Bochao Liu, Jianghu Lu, Pengju Wang, Junjie…
Look Through Masks: Towards Masked Face Recognition with De-Occlusion Distillationby Chenyu Li, Shiming Ge, Daichi…
Multimodal Generalized Category Discoveryby Yuchang Su, Renping Zhou, Siyu Huang, Xingjian Li, Tianyang Wang, Ziyue…
Leveraging Distillation Techniques for Document Understanding: A Case Study with FLAN-T5by Marcel Lamott, Muhammad Armaghan…
Gradient-free Post-hoc Explainability Using Distillation Aided Learnable Approachby Debarpan Bhattacharya, Amir H. Poorjam, Deepak Mittal,…