noun
squashing function
Technical term in machine learning and neural networks; refers to a function that compresses an input into a bounded range, such as a sigmoid or tanh function.
ニューラルネットワークでは、活性化関数としてスカッシング関数がよく使われる。
In neural networks, squashing functions are often used as activation functions.
活性化関数 is the broader term for activation functions, while スカッシング関数 specifically refers to those that squash input values into a limited range.
From English 'squashing' + 関数 (function). The term is a direct loan translation used in technical contexts.