フォロー
Momose Oyama
Momose Oyama
確認したメール アドレス: sys.i.kyoto-u.ac.jp - ホームページ
タイトル
引用先
引用先
Norm of Word Embedding Encodes Information Gain
M Oyama, S Yokoi, H Shimodaira
Proceedings of the 2023 Conference on Empirical Methods in Natural Language …, 2022
72022
Discovering Universal Geometry in Embeddings with ICA
H Yamagiwa*, M Oyama*, H Shimodaira
Proceedings of the 2023 Conference on Empirical Methods in Natural Language …, 2023
62023
Revisiting Cosine Similarity via Normalized ICA-transformed Embeddings
H Yamagiwa, M Oyama, H Shimodaira
arXiv preprint arXiv:2406.10984, 2024
2024
Predicting Drug-Gene Relations via Analogy Tasks with Word Embeddings
H Yamagiwa, R Hashimoto, K Arakane, K Murakami, S Soeda, M Oyama, ...
arXiv preprint arXiv:2406.00984, 2024
2024
Analogy Tasks in BioConceptVec using Biological Pathways
H Yamagiwa, R Hashimoto, K Arakane, K Murakami, M Oyama, ...
IEICE Technical Report; IEICE Tech. Rep. 123 (91), 113-120, 2023
2023
A Mechanism of Implicit Regularization in Deep Learning
M Kubo, G Sugiura, K Shinzato, M Oyama
現在システムで処理を実行できません。しばらくしてからもう一度お試しください。
論文 1–6