Feature Wise
Feature-wise research explores how individual features within data contribute to model performance and interpretability across diverse machine learning tasks. Current efforts focus on developing methods for feature selection, extraction, and fusion, employing techniques like sparse autoencoders, attention mechanisms, and graph convolutional networks to optimize feature utilization and enhance model accuracy and explainability. This work is significant for improving model efficiency, robustness, and trustworthiness, with applications ranging from medical image analysis and malware detection to natural language processing and financial forecasting.
Papers
Bridging Pre-trained Language Models and Hand-crafted Features for Unsupervised POS Tagging
Houquan Zhou, Yang Li, Zhenghua Li, Min Zhang
Exploiting Cross Domain Acoustic-to-articulatory Inverted Features For Disordered Speech Recognition
Shujie Hu, Shansong Liu, Xurong Xie, Mengzhe Geng, Tianzi Wang, Shoukang Hu, Mingyu Cui, Xunying Liu, Helen Meng
A Closer Look at Knowledge Distillation with Features, Logits, and Gradients
Yen-Chang Hsu, James Smith, Yilin Shen, Zsolt Kira, Hongxia Jin
Selection of entropy based features for the analysis of the Archimedes' spiral applied to essential tremor
Karmele López-De-Ipiña, Alberto Bergareche, Patricia De La Riva, Jordi Sole-Casals, Marcos Faundez-Zanuy, Jose Felix Marti-Masso, Mikel Iturrate, Blanca Beitia, Pilar Calvo, Enric Sesa-Nogueras, Josep Roure, Itziar Gurrutxaga, Joseba Garcia-Melero
Vertical Federated Principal Component Analysis and Its Kernel Extension on Feature-wise Distributed Data
Yiu-ming Cheung, Juyong Jiang, Feng Yu, Jian Lou
CAFE: Learning to Condense Dataset by Aligning Features
Kai Wang, Bo Zhao, Xiangyu Peng, Zheng Zhu, Shuo Yang, Shuo Wang, Guan Huang, Hakan Bilen, Xinchao Wang, Yang You