├── Deep_Recomender_RL_System.ipynb ├── Embedding_model.ipynb ├── README.md ├── RL.ipynb ├── Reinforcement Learning in Recommender Systems.pptx ├── actor_5000_fixed.h5 ├── critic_5000_fixed.h5 ├── training_precision_%_top_5 (1).png ├── training_precision_%_top_5 (17).png ├── training_precision_%_top_5 (21).png └── training_precision_%_top_5.png /README.md: -------------------------------------------------------------------------------- 1 | # Deep RL-based Recommender System 2 | 3 | ## Paper and code - Deep Reinforcement Learning based Recommendation with Explicit User-Item Interactions Modeling 4 | https://arxiv.org/pdf/1810.12027.pdf 5 | 6 | Code available at: https://github.com/backgom2357/Recommender_system_via_deep_RL 7 | 8 | To check our training results, use [the actor model](https://github.com/UnibucProjects/DeepRLRecommenderSystem/blob/main/actor_5000_fixed.h5) and [the critic model](https://github.com/UnibucProjects/DeepRLRecommenderSystem/blob/main/critic_5000_fixed.h5) after 5000 epochs of training. 9 | 10 | Dataset used available on [Kaggle](https://www.kaggle.com/odedgolden/movielens-1m-dataset) 11 | 12 | #### Precision first 50 epochs 13 | ![training precision first 50](https://github.com/UnibucProjects/DeepRLRecommenderSystem/blob/main/training_precision_%25_top_5.png?raw=true) 14 | #### Precision first 100 epochs 15 | ![training precision first 100](https://github.com/UnibucProjects/DeepRLRecommenderSystem/blob/main/training_precision_%25_top_5%20(1).png?raw=true) 16 | #### Precision first 1500 epochs 17 | ![training precision first 1500](https://github.com/UnibucProjects/DeepRLRecommenderSystem/blob/main/training_precision_%25_top_5%20(17).png?raw=true) 18 | #### Precision first 5000 epochs 19 | ![training precision first 5000](https://github.com/UnibucProjects/DeepRLRecommenderSystem/blob/main/training_precision_%25_top_5%20(21).png?raw=true) 20 | 21 | For the embedding model, consult visualisations [here](https://github.com/UnibucProjects/DeepRLRecommenderSystem/blob/main/Embedding_model.ipynb) 22 | 23 | ### Related articles: 24 | 25 | * [28] X. Zhao, L. Zhang, Z. Ding, D. Yin, Y. Zhao, and J. Tang, “Deep reinforcement learning for list-wise recommendations,” CoRR, vol. abs/1801.00209, 2018. Available [here](https://arxiv.org/abs/1801.00209) 26 | 27 | * [32] Y. Hu, Q. Da, A. Zeng, Y. Yu, and Y. Xu, “Reinforcement learning to rank in e-commerce search engine: Formalization, analysis, and application,” CoRR, vol. abs/1803.00710, 2018 28 | Available [here](https://arxiv.org/pdf/1803.00710.pdf) 29 | 30 | * [19] G. Zheng, F. Zhang, Z. Zheng, Y. Xiang, N. J. Yuan, X. Xie, and Z. Li, “DRN: A deep reinforcement learning framework for news recommendation,” in WWW 2018, Lyon, France, April 23-27, 2018, 2018, pp. 167–176. 31 | Available [here](https://www.personal.psu.edu/~gjz5038/paper/www2018_reinforceRec/www2018_reinforceRec.pdf) 32 | 33 | * [29] X. Zhao, L. Zhang, Z. Ding, L. Xia, J. Tang, and D. Yin, “Recommendations with negative feedback via pairwise deep reinforcement learning,” CoRR, vol. abs/1802.06501, 2018. 34 | Available [here](https://arxiv.org/abs/1802.06501) 35 | 36 | ### Team discussion - summary of all the above articles 37 | * https://docs.google.com/document/d/1bRsCVpxZx7NLvZJg_JApn6GUzTF9AThOlWKX8gprEdY/edit?usp=sharing 38 | 39 | ### Check our presentation [here](https://github.com/UnibucProjects/DeepRLRecommenderSystem/blob/main/Reinforcement%20Learning%20in%20Recommender%20Systems.pptx) 40 | -------------------------------------------------------------------------------- /Reinforcement Learning in Recommender Systems.pptx: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/UnibucProjects/DeepRLRecommenderSystem/d79afdb7f20ff06c3feaa5b4ed9a4bcad3d6edd5/Reinforcement Learning in Recommender Systems.pptx -------------------------------------------------------------------------------- /actor_5000_fixed.h5: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/UnibucProjects/DeepRLRecommenderSystem/d79afdb7f20ff06c3feaa5b4ed9a4bcad3d6edd5/actor_5000_fixed.h5 -------------------------------------------------------------------------------- /critic_5000_fixed.h5: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/UnibucProjects/DeepRLRecommenderSystem/d79afdb7f20ff06c3feaa5b4ed9a4bcad3d6edd5/critic_5000_fixed.h5 -------------------------------------------------------------------------------- /training_precision_%_top_5 (1).png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/UnibucProjects/DeepRLRecommenderSystem/d79afdb7f20ff06c3feaa5b4ed9a4bcad3d6edd5/training_precision_%_top_5 (1).png -------------------------------------------------------------------------------- /training_precision_%_top_5 (17).png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/UnibucProjects/DeepRLRecommenderSystem/d79afdb7f20ff06c3feaa5b4ed9a4bcad3d6edd5/training_precision_%_top_5 (17).png -------------------------------------------------------------------------------- /training_precision_%_top_5 (21).png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/UnibucProjects/DeepRLRecommenderSystem/d79afdb7f20ff06c3feaa5b4ed9a4bcad3d6edd5/training_precision_%_top_5 (21).png -------------------------------------------------------------------------------- /training_precision_%_top_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/UnibucProjects/DeepRLRecommenderSystem/d79afdb7f20ff06c3feaa5b4ed9a4bcad3d6edd5/training_precision_%_top_5.png --------------------------------------------------------------------------------