Изменения

Перейти к: навигация, поиск

Мета-обучение

266 байт добавлено, 26 январь
Замена текстовых ссылок на сноски
Ограничения
- No free lunch teorem <ref>[https://www.researchgate.net/publication/221997149_No_Free_Lunch_Theorems_for_Search Wolpert and Macready, 1996; ]</ref><ref>[https://www.researchgate.net/publication/228671734_Toward_a_justification_of_meta-learning_Is_the_no_free_lunch_theorem_a_show-stopper Giraud-Carrier and Provost, 2005]</ref>
<h2>Simple view</h2>
Во время вычисления похожести задач важно нормализовать все мета-фичи [bardnet], использовать feature selection [todorovski] или использовать dimensionality reduction (PCA, например).
<h2>== Примечания==<references/h2>
* https://lilianweng.github.io/lil-log/2018/11/30/meta-learning.html#define-the-meta-learning-problem* https://arxiv.org/pdf/1810.03548.pdf* https://www.ml4aad.org/wp-content/uploads/2018/09/chapter2-metalearning.pdf* https://openreview.net/pdf?id=rJY0-Kcll* https://www.fruct.org/publications/ainl-fruct/files/Fil.pdf* Alexandros Kalousis and Melanie Hilario. Model selection via via meta-learning: a comparativestudy.Intl Journ. on Artificial Intelligence Tools, 10(4):525–554, 2001.* R ́emi Bardenet, M ́aty ́as Brendel, Bal ́azs K ́egl, and Michele Sebag. Collaborative hyperparameter tuning. In Proceedings of ICML 2013, pages 199–207, 2013* L Todorovski and S Dzeroski. Experiments in meta-level learning with ILP.Lecture Notes in Computer Science, 1704:98–106, 1999.
Анонимный участник

Навигация