Convolutional RNN

1. Introduction

2. Conclusion

3. Discussion

4. Abstract

5. Related work

6. Experiments

6.1. Emotion classification

6.2. Age and gender classification

6.3. Didn't use backpropogation through time (BPTT)

7. CLSTM Architecutre

7.1. CLSTM

7.2. CBLSTM

7.3. Extended CLSTM

7.4. Max, mean, last pooling

7.5. Use hidden state, cell state, output as extracted feature

8. Evaluation metric.

8.1. Unweighted average (UA) recall

9. Convolutional LSTM

10. Extracting features

11. LSTM & bi-directional LSTM

12. Convolution before RNN

13. Convolutional RNN