appricot’s diary

日々の勉強のメモ

Transfer Learning

To tune or note to tune? Adapting Pretrained Representations to Diverse tasks ==> 読後メモ

論文 こちらの論文の内容を、要約。 arxiv.org 概要 Tranfer Learningは、2つのStageから構成 Pre-training: 汎用的な入力データのRepresentationを学習 (例:単語のベクトル表現を学習) Adaptation: そのRepresentationを新たな問題に提供 (例:そのベクト…

Adapted Tree Boosting for Transfer Learning ==> 読後メモ

論文 arxiv.org 概要 Alipayでの不正検知に、Transfer Learning(転移学習)を適用。 通常、Transfer Learningといえば、ニューラルネットワークを用いたものが多いが、この論文では、XGBを利用。 Transfer Learningの分類 Instance based methods Source Doma…

Building a Credit Risk Model using Transfer Learning and Domain Adaptation ==> 読後メモ

Paper Title: Building a Credit Risk Model using Transfer Learning and Domain Adaptation Frontiers | Credit Risk Modeling Using Transfer Learning and Domain Adaptation Summary 概要 Transfer learningをCredit Scoringに応用 前提: Source Domai…

TransBoost: A Boosting-Tree Kernel Transfer Learning Algorithm for Improving Financial Inclusion ==> 読後メモ

Paper Title: TransBoost: A Boosting-Tree Kernel Transfer Learning Algorithm for Improving Financial Inclusion [2112.02365] TransBoost: A Boosting-Tree Kernel Transfer Learning Algorithm for Improving Financial Inclusion Summary 概要 Gradie…