Speaker : YI-CHENG HUNG

Presentation on theme: "Speaker : YI-CHENG HUNG"— Presentation transcript:

1 Speaker : YI-CHENG HUNG
Curriculum Learning for Heterogeneous Star Network Embedding via Deep Reinforcement Learning Source : WSDM 2018 Advisor : JIA-LING KOH Speaker : YI-CHENG HUNG Date:2018/04/24

2 Outline Introduction Method Experiment Conclusion

3 Outline Introduction Method Experiment Conclusion

4 Introduction-motivation
訓練模型 預測類型: 奇幻類型 NN的INPUT是embedding vector，要如何訓練一個足夠代表input的embedding vector? Embedding vector

5 Introduction-goal 如何得知最好的修課順序? Goal:在考試時拿到最好的成績 修課順序

6 Outline Introduction Method Experiment Conclusion

7 Method-framework 假設已經一經上過(A,A,B,B,C)的課程了 如何決定下一次的課程?

8 Method-Planning Module

9 Q-Learn η是學習率，決定此次誤差有多少被學習 Γ是折扣率，決定未來策略的比重 現實 估計

10 Method-framework

11 Method-Learning Module

12 Method-Integrating

13 Outline Introduction Method Experiment Conclusion

14 Experiment-dataset 多標籤類型會進行排除

15 Quantitative results in the unsupervised setting

16 Quantitative results in the semi-supervised setting

17 Performance w.r.t. training steps on the DBLP dataset.

18 Performance v.s. the running time.

19 Performance w.r.t. #simulations on the IMDB dataset.

20 Outline Introduction Method Experiment Conclusion

21 Conclusion Proposed a deep reinforcement learning based approach
Integrates the learning and planning strategies Experimental results proved the effectiveness and efficiency of our approach.