Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[FEAT] LightGCN으로 학습한 유저 임베딩을 Autoencoder 계열에 사용 #18

Open
2 tasks
Glanceyes opened this issue Jun 24, 2022 · 1 comment
Open
2 tasks
Assignees
Labels
📈 enhancement New feature or request 🔖 inference Issue with inference 🧬 modeling Issue with training or validation

Comments

@Glanceyes
Copy link
Contributor

Glanceyes commented Jun 24, 2022

😎 Responsibility

@wh4044, @juk1329 , @Glanceyes

💡Issue

  • @juk1329님이 실험한 LightGCN으로 문제 추천 결과 Inference 시 유저의 성향에 과도하게 편향된 결과가 나온다는 보고가 있었습니다.
  • LightGCN의 학습된 유저 또는 태그 임베딩을 Multi-DAE와 Multi-VAE의 입력의 앞단에 붙이는 것을 적용해 볼 필요가 있습니다.

🔗 Reference

Yifan Chen, and Maarten de Rijke. 2017. A Collective Variational Autoencoder for Top-N Recommendation with Side Information. ACM

🧭 To Do

  • LightGCN 구현 코드 최적화
  • LightGCN로 학습된 유저 임베딩 저장하여 이를 Autoencoder 계열 모델에 사용하는 로직 구현
@Glanceyes Glanceyes added 📈 enhancement New feature or request 🧬 modeling Issue with training or validation 🔖 inference Issue with inference labels Jun 24, 2022
@Glanceyes
Copy link
Contributor Author

LightGCN 모델 학습 시 데이터 전처리 부분에서 1시간 이상이 소요되어 꽤 많은 시간이 소모됩니다.
먼저 LightGCN 모델에 사용할 데이터 전처리 부분을 최적화하는 것부터 진행하겠습니다.

for loop로 도는 것보다는 applygroupby를 이용해는 게 좋아 보입니다.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
📈 enhancement New feature or request 🔖 inference Issue with inference 🧬 modeling Issue with training or validation
Projects
None yet
Development

No branches or pull requests

3 participants