Aug, 2023
释放相似性匹配的潜力:可扩展性、监督与预训练
Unlocking the Potential of Similarity Matching: Scalability, Supervision and Pre-training
Yanis Bahroun, Shagesh Sridharan, Atithi Acharya, Dmitri B. Chklovskii, Anirvan M. Sengupta
TL;DR本研究主要介绍了通过 Convolutional Nonnegative SM 和类似于 canonical correlation analysis 的方法来扩大 primarily unsupervised similarity matching(SM)框架,以便在大规模数据集上实现 SM,并通过与使用 backpropagation 算法训练的模型进行比较,评估所提出方法的特征表现。