sequence to sequence, B3TB2006, i

Size: px

Start display at page:

Download "sequence to sequence, B3TB2006, i"

かずまさてっちがわら
5 years ago
Views:

1 B3TB

2 sequence to sequence, B3TB2006, i

3 A Study on a Style Control for Dialogue Response Generation Reina Akama Abstract We propose a new dialogue response generation model combining sequence to sequence and metastasis learning for controling a STYLE, which impresses a character of a specific speaker. To generate consistent style responses, most of priviouse works for collected text pairs by handmade rules or human annotation. However, thier costs were too expensive to learn response generation models with a consistent style. In our methods, after learning pre-training model with a largescale copora without restriction styles, we learn a response generation model with a small-scale copora with a style restriction and the pre-training model. Compared with responses between our proposed model and baseline models without metastasis learning, our model generated consistent style and appropriate responses to input texts. Keywords: response generation, dialog, style control, newral network, transfer learning Graduation Thesis, Department of Information and Intelligent Systems, Tohoku University, B3TB2006, March 31, ii

4 Contents Sequence to sequence Transfer Transfer+freq Transfer+sim iii

5 List of Figures 1 seq2seq iv

6 List of Tables v

7 1 Twitter (SNS) 1 Recurrent Neural Network (RNN) sequence to sequence (seq2seq) [18] SNS seq2seq 1 2) 4) 6) 1 [19, 20] [10] Table 1: 1) User: 2) System: 3) User: 4) System: 5) User: 6) System: 1

8 seq2seq SNS 2

9 2 2.1 Sequence to sequence 1 seq2seq [4, 18] Sutskever seq2seq 1 / / / / / / Figure 1: seq2seq seq2seq RNN RNN (x 1,..., x T ) t (1) h t (2) y t T (y 1,..., y T ) W hx W hh W yh h t = sigm(w hx x t + W hh h t 1 ) (1) y t = W yh h t (2) RNN T RNN v () RNN T ( ) Seq2seq RNN Long Short-Term Memory (LSTM) [7] Sutskever seq2seq 2 LSTM 1 LSTM (3) 3

10 (x 1,..., x T ) (y 1,..., y T ) p(y 1,..., y T x 1,..., x T ) p(y 1,..., y T x 1,..., x T ) = T t=1 p(y t v, y 1,..., y t 1 ) (3) v LSTM 1 y 1 LSTM v seq2seq [5, 18] [14] [15] seq2seq seq2seq Walker [19, 11, 12, 13] Walker [20] 4

11 seq2seq [10] Twitter [1] [2] [3] [9, 6] [16] GloVe[17]. 5

12 3 3.1?? Sutskever seq2seq [18] seq2seq RNN Transfer ( ) N s Transfer seq2seq seq2seq 事前学習コーパスコーパス内の単語学習で扱う語彙スタイルコーパスコーパス内の単語 Figure 2: 6

13 3.2 Seq2seq ( unk ) 2 2 Transfer+freq Transfer+sim Transfer+freq Transfer+freq N s N p N s 2 N p = 25, 000, N s = 1, ,000 24,000 1, Transfer+sim Transfer+sim Glove [17] 2 7

14 8

15 4 Twitter TV % 5% Table 2: ,688, ,880 12,564 12,102 1,476 2, Twitter Twitter URL TV

16 0.3% 0.04% Table 3: (a) (b)??! 10

17 Transfer 3(c) Transfer+freq 3(d) Transfer+sim 3(e) 25,000 5 Transfer Transfer+freq 1, Transfer+sim Twitter GloVe seq2seq Base 3(a), Mixed 3(b) 2 Base 25,000 Base 3 Mixed Mixed 24,000 1,000 24,

18 seq2seq seq2seq seq2seq seq2seq 事前学習コーパスコーパス内の単語学習で扱う語彙スタイルコーパスコーパス内の単語事前学習コーパス + スタイルコーパスコーパス内の単語学習で扱う単語スタイルコーパスコーパス内の単語 (a) Base (b) Mixed seq2seq seq2seq seq2seq seq2seq 事前学習コーパスコーパス内の単語学習で扱う語彙スタイルコーパスコーパス内の単語事前学習コーパスコーパス内の単語学習で扱う語彙スタイルコーパスコーパス内の単語 (c) Transfer (d) Transfer+freq seq2seq seq2seq 事前学習コーパスコーパス内の単語学習で扱う語彙スタイルコーパスコーパス内の単語 (e) Transfer+sim Figure 3: 5 12

19 Seq2seq 1024 LSTM dropout rate 0.2 Adam [8] Adam ) / 2 2) / 2. Twitter Base 80% 3 80% Transfer+freq Mixed Mixed 5 Transfer Transfer+freq 90% 13

20 Table 4: Base 39 (78%) 18 (36%) Mixed 39 (78%) 23 (46%) Transfer 41 (82%) 39 (78%) Transfer+freq 38 (76%) 39 (78%) Transfer+sim 39 (78%) 38 (76%) * 4 Table 5: Base 43 (86%) 0 (0%) Mixed 40 (80%) 16 (32%) Transfer 29 (58%) 45 (90%) Transfer+freq 31 (62%) 47 (94%) Transfer+sim 32 (64%) 44 (88%) * % Base Transfer+freq 7 2) Transfer+freq

21 3 1 1 Base 3 1,400 - Table 6: 1) User: 2) System: 3) User: 4) System: 5) User: 6) System: Table 7: 1) User: 2) System: 3) User: 4) System: 5) User: 6) System: 15

22 6 seq2seq TV 16

23 Preferred Networks TV 17

24 References [1] Nips 2005 workshop inductive transfer: 10 years later [2] Andrew Arnold, Ramesh Nallapati, and William W Cohen. A comparative study of methods for transductive transfer learning. In Data Mining Workshops, ICDM Workshops Seventh IEEE International Conference on, pp IEEE, [3] John Blitzer, Mark Dredze, Fernando Pereira, et al. Biographies, bollywood, boom-boxes and blenders: Domain adaptation for sentiment classification. In ACL, Vol. 7, pp , [4] Kyunghyun Cho, Bart van Merrienboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. Learning phrase representations using rnn encoder-decoder for statistical machine translation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). Association for Computational Linguistics, [5] Bahdanau Dzmitry, Cho Kyunghyun, and Bengio Yoshua. Neural machine translation by jointly learning to align and translate. In The International Conference on Learning Representations (ICLR), [6] Almut Silja Hildebrand, Matthias Eck, Stephan Vogel, and Alex Waibel. Adaptation of the translation model for statistical machine translation based on information retrieval. In Proceedings of EAMT, Vol. 2005, pp , [7] Sepp Hochreiter and Jürgen Schmidhuber. Long short-term memory. Neural computation, pp , [8] Diederik Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In The International Conference on Learning Representations (ICLR),

25 [9] Philipp Koehn and Josh Schroeder. Experiments in domain adaptation for statistical machine translation. In Proceedings of the second workshop on statistical machine translation, pp Association for Computational Linguistics, [10] Jiwei Li, Michel Galley, Chris Brockett, Georgios Spithourakis, Jianfeng Gao, and Bill Dolan. A persona-based neural conversation model. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, pp , [11] Grace I Lin and Marilyn A Walker. All the world s a stage: Learning character models from film. In AIIDE, [12] François Mairesse and Marilyn A Walker. Towards personality-based user adaptation: psychologically informed stylistic language generation. User Modeling and User-Adapted Interaction, pp , [13] François Mairesse and Marilyn A Walker. Controlling user perceptions of linguistic style: Trainable generation of personality traits. Computational Linguistics, pp , [14] Ramesh Nallapati, Bowen Zhou, Cicero dos Santos, Caglar Gulcehre, and Bing Xiang. Abstractive text summarization using sequence-to-sequence rnns and beyond. In Proceedings of The 20th SIGNLL Conference on Computational Natural Language Learning, pp Association for Computational Linguistics, [15] Vinyals Oriol and Le Quoc. A neural conversational model. In International Conference on Machine Learning(ICML) Deep Learning Workshop 2015, [16] Sinno Jialin Pan and Qiang Yang. A survey on transfer learning. IEEE Transactions on knowledge and data engineering, Vol. 22, No. 10, pp , [17] Jeffrey Pennington, Richard Socher, and Christopher D. Manning. Glove: Global vectors for word representation. In Empirical Meth- 19

26 ods in Natural Language Processing (EMNLP), pp , [18] Ilya Sutskever, Oriol Vinyals, and Quoc V Le. Sequence to sequence learning with neural networks. In Advances in neural information processing systems, pp , [19] Marilyn A Walker, Grace I Lin, and Jennifer Sawyer. An annotated corpus of film dialogue for learning and characterizing character style. In LREC, pp , [20],,,,,.., Vol. 31, No. 1, pp. DSF E 1,

Haiku Generation Based on Motif Images Using Deep Learning Koki Yoneda 1 Soichiro Yokoyama 2 Tomohisa Yamashita 2 Hidenori Kawamura Scho

Haiku Generation Based on Motif Images Using Deep Learning 1 2 2 2 Koki Yoneda 1 Soichiro Yokoyama 2 Tomohisa Yamashita 2 Hidenori Kawamura 2 1 1 School of Engineering Hokkaido University 2 2 Graduate