PreGIP: Watermarking the Pretraining of Graph Neural Networks for Deep Intellectual Property Protectionby Enyan Dai,…
Attention with Markov: A Framework for Principled Analysis of Transformers via Markov Chainsby Ashok Vardhan…
Scaling Laws for Downstream Task Performance in Machine Translationby Berivan Isik, Natalia Ponomareva, Hussein Hazimeh,…
A Hard-to-Beat Baseline for Training-free CLIP-based Adaptationby Zhengbo Wang, Jian Liang, Lijun Sheng, Ran He,…
On the Emergence of Cross-Task Linearity in the Pretraining-Finetuning Paradigmby Zhanpeng Zhou, Zijun Chen, Yilan…
Evaluating the Factuality of Zero-shot Summarizers Across Varied Domainsby Sanjana Ramprasad, Kundan Krishna, Zachary C…
FROSTER: Frozen CLIP Is A Strong Teacher for Open-Vocabulary Action Recognitionby Xiaohu Huang, Hao Zhou,…
Shortened LLaMA: Depth Pruning for Large Language Models with Comparison of Retraining Methodsby Bo-Kyeong Kim,…
BECLR: Batch Enhanced Contrastive Few-Shot Learningby Stylianos Poulakakis-Daktylidis, Hadi Jamali-RadFirst submitted to arxiv on: 4…
Variance Alignment Score: A Simple But Tough-to-Beat Data Selection Method for Multimodal Contrastive Learningby Yiping…