| { |
| "paper_id": "D17-1039", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T16:16:46.338970Z" |
| }, |
| "title": "Unsupervised Pretraining for Sequence to Sequence Learning", |
| "authors": [ |
| { |
| "first": "Prajit", |
| "middle": [], |
| "last": "Ramachandran", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "prajit@google.com" |
| }, |
| { |
| "first": "Peter", |
| "middle": [ |
| "J" |
| ], |
| "last": "Liu", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "peterjliu@google.com" |
| }, |
| { |
| "first": "Quoc", |
| "middle": [ |
| "V" |
| ], |
| "last": "Le Google Brain", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "This work presents a general unsupervised learning method to improve the accuracy of sequence to sequence (seq2seq) models. In our method, the weights of the encoder and decoder of a seq2seq model are initialized with the pretrained weights of two language models and then fine-tuned with labeled data. We apply this method to challenging benchmarks in machine translation and abstractive summarization and find that it significantly improves the subsequent supervised models. Our main result is that pretraining improves the generalization of seq2seq models. We achieve state-of-theart results on the WMT English\u2192German task, surpassing a range of methods using both phrase-based machine translation and neural machine translation. Our method achieves a significant improvement of 1.3 BLEU from the previous best models on both WMT'14 and WMT'15 English\u2192German. We also conduct human evaluations on abstractive summarization and find that our method outperforms a purely supervised learning baseline in a statistically significant manner.", |
| "pdf_parse": { |
| "paper_id": "D17-1039", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "This work presents a general unsupervised learning method to improve the accuracy of sequence to sequence (seq2seq) models. In our method, the weights of the encoder and decoder of a seq2seq model are initialized with the pretrained weights of two language models and then fine-tuned with labeled data. We apply this method to challenging benchmarks in machine translation and abstractive summarization and find that it significantly improves the subsequent supervised models. Our main result is that pretraining improves the generalization of seq2seq models. We achieve state-of-theart results on the WMT English\u2192German task, surpassing a range of methods using both phrase-based machine translation and neural machine translation. Our method achieves a significant improvement of 1.3 BLEU from the previous best models on both WMT'14 and WMT'15 English\u2192German. We also conduct human evaluations on abstractive summarization and find that our method outperforms a purely supervised learning baseline in a statistically significant manner.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "Sequence to sequence (seq2seq) models Cho et al., 2014; Kalchbrenner and Blunsom, 2013; Allen, 1987; Neco and Forcada, 1997) are extremely effective on a variety of tasks that require a mapping between a variable-length input sequence to a variable-length output sequence. The main weakness of sequence to sequence models, and deep networks in general, lies in the fact that they can easily overfit when the amount of supervised training data is small.", |
| "cite_spans": [ |
| { |
| "start": 38, |
| "end": 55, |
| "text": "Cho et al., 2014;", |
| "ref_id": "BIBREF6" |
| }, |
| { |
| "start": 56, |
| "end": 87, |
| "text": "Kalchbrenner and Blunsom, 2013;", |
| "ref_id": "BIBREF15" |
| }, |
| { |
| "start": 88, |
| "end": 100, |
| "text": "Allen, 1987;", |
| "ref_id": "BIBREF0" |
| }, |
| { |
| "start": 101, |
| "end": 124, |
| "text": "Neco and Forcada, 1997)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "In this work, we propose a simple and effective technique for using unsupervised pretraining to improve seq2seq models. Our proposal is to initialize both encoder and decoder networks with pretrained weights of two language models. These pretrained weights are then fine-tuned with the labeled corpus. During the fine-tuning phase, we jointly train the seq2seq objective with the language modeling objectives to prevent overfitting.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "We benchmark this method on machine translation for English\u2192German and abstractive summarization on CNN and Daily Mail articles. Our main result is that a seq2seq model, with pretraining, exceeds the strongest possible baseline in both neural machine translation and phrasebased machine translation. Our model obtains an improvement of 1.3 BLEU from the previous best models on both WMT'14 and WMT'15 English\u2192German. On human evaluations for abstractive summarization, we find that our model outperforms a purely supervised baseline, both in terms of correctness and in avoiding unwanted repetition.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "We also perform ablation studies to understand the behaviors of the pretraining method. Our study confirms that among many other possible choices of using a language model in seq2seq with attention, the above proposal works best. Our study also shows that, for translation, the main gains come from the improved generalization due to the pretrained features. For summarization, pretraining the encoder gives large improvements, suggesting that the gains come from the improved optimization of the encoder that has been unrolled for hundreds of timesteps. On both tasks, our proposed method always improves generalization on the test sets. Figure 1 : Pretrained sequence to sequence model. The red parameters are the encoder and the blue parameters are the decoder. All parameters in a shaded box are pretrained, either from the source side (light red) or target side (light blue) language model. Otherwise, they are randomly initialized.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 639, |
| "end": 647, |
| "text": "Figure 1", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "In the following section, we will describe our basic unsupervised pretraining procedure for sequence to sequence learning and how to modify sequence to sequence learning to effectively make use of the pretrained weights. We then show several extensions to improve the basic model.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Methods", |
| "sec_num": "2" |
| }, |
| { |
| "text": "Given an input sequence x 1 , x 2 , ..., x m and an output sequence y n , y n\u22121 , ..., y 1 , the objective of sequence to sequence learning is to maximize the likelihood p(y n , y n\u22121 , ..., y 1 |x 1 , x 2 , ..., x m ). Common sequence to sequence learning methods decompose this objective as", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Basic Procedure", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "p(y n , y n\u22121 , ..., y 1 |x 1 , x 2 , ..., x m ) = n t=1 p(y t |y t\u22121 , ..., y 1 ; x 1 , x 2 , ..., x m ).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Basic Procedure", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "In sequence to sequence learning, an RNN encoder is used to represent x 1 , ..., x m as a hidden vector, which is given to an RNN decoder to produce the output sequence. Our method is based on the observation that without the encoder, the decoder essentially acts like a language model on y's. Similarly, the encoder with an additional output layer also acts like a language model. Thus it is natural to use trained languages models to initialize the encoder and decoder.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Basic Procedure", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "Therefore, the basic procedure of our approach is to pretrain both the seq2seq encoder and decoder networks with language models, which can be trained on large amounts of unlabeled text data. This can be seen in Figure 1 , where the parameters in the shaded boxes are pretrained. In the following we will describe the method in detail using machine translation as an example application.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 212, |
| "end": 220, |
| "text": "Figure 1", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Basic Procedure", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "First, two monolingual datasets are collected, one for the source side language, and one for the target side language. A language model (LM) is trained on each dataset independently, giving an LM trained on the source side corpus and an LM trained on the target side corpus.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Basic Procedure", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "After two language models are trained, a multilayer seq2seq model M is constructed. The embedding and first LSTM layers of the encoder and decoder are initialized with the pretrained weights. To be even more efficient, the softmax of the decoder is initialized with the softmax of the pretrained target side LM.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Basic Procedure", |
| "sec_num": "2.1" |
| }, |
| { |
| "text": "After the seq2seq model M is initialized with the two LMs, it is fine-tuned with a labeled dataset. However, this procedure may lead to catastrophic forgetting, where the model's performance on the language modeling tasks falls dramatically after fine-tuning (Goodfellow et al., 2013) . This may hamper the model's ability to generalize, especially when trained on small labeled datasets.", |
| "cite_spans": [ |
| { |
| "start": 259, |
| "end": 284, |
| "text": "(Goodfellow et al., 2013)", |
| "ref_id": "BIBREF10" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Monolingual language modeling losses", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "To ensure that the model does not overfit the labeled data, we regularize the parameters that were pretrained by continuing to train with the monolingual language modeling losses. The seq2seq and language modeling losses are weighted equally.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Monolingual language modeling losses", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "In our ablation study, we find that this technique is complementary to pretraining and is important in achieving high performance.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Monolingual language modeling losses", |
| "sec_num": "2.2" |
| }, |
| { |
| "text": "Pretraining and the monolingual language modeling losses provide the vast majority of improvements to the model. However in early experimentation, we found minor but consistent improvements with two additional techniques: a) residual connections and b) multi-layer attention (see Figure 2). Residual connections: As described, the input vector to the decoder softmax layer is a random vector because the high level (non-first) layers of the LSTM are randomly initialized. This introduces random gradients to the pretrained parameters. To avoid this, we use a residual connection from the output of the first LSTM layer directly to the input of the softmax (see Figure 2 -a).", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 280, |
| "end": 286, |
| "text": "Figure", |
| "ref_id": null |
| }, |
| { |
| "start": 661, |
| "end": 669, |
| "text": "Figure 2", |
| "ref_id": "FIGREF0" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Other improvements to the model", |
| "sec_num": "2.3" |
| }, |
| { |
| "text": "Multi-layer attention: In all our models, we use an attention mechanism (Bahdanau et al., 2015) , where the model attends over both top and first layer (see Figure 2 -b). More concretely, given a query vector q t from the decoder, encoder states from the first layer h 1 1 , . . . , h 1 T , and encoder states from the last layer h L 1 , . . . , h L T , we compute the attention context vector c t as follows:", |
| "cite_spans": [ |
| { |
| "start": 72, |
| "end": 95, |
| "text": "(Bahdanau et al., 2015)", |
| "ref_id": "BIBREF1" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 157, |
| "end": 165, |
| "text": "Figure 2", |
| "ref_id": "FIGREF0" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Other improvements to the model", |
| "sec_num": "2.3" |
| }, |
| { |
| "text": "\u03b1 i = exp(q t \u2022 h N i ) T j=1 exp(q t \u2022 h N j ) c 1 t = T i=1 \u03b1 i h 1 i c N t = T i=1 \u03b1 i h N i c t = [c 1 t ; c N t ]", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Other improvements to the model", |
| "sec_num": "2.3" |
| }, |
| { |
| "text": "3 Experiments", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Other improvements to the model", |
| "sec_num": "2.3" |
| }, |
| { |
| "text": "In the following section, we apply our approach to two important tasks in seq2seq learning: ma-chine translation and abstractive summarization. On each task, we compare against the previous best systems. We also perform ablation experiments to understand the behavior of each component of our method.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Other improvements to the model", |
| "sec_num": "2.3" |
| }, |
| { |
| "text": "Dataset and Evaluation: For machine translation, we evaluate our method on the WMT English\u2192German task (Bojar et al., 2015) . We used the WMT 14 training dataset, which is slightly smaller than the WMT 15 dataset. Because the dataset has some noisy examples, we used a language detection system to filter the training examples. Sentences pairs where either the source was not English or the target was not German were thrown away. This resulted in around 4 million training examples. Following Sennrich et al. 2015a, we use subword units (Sennrich et al., 2015b) with 89500 merge operations, giving a vocabulary size around 90000. The validation set is the concatenated new-stest2012 and newstest2013, and our test sets are newstest2014 and newstest2015. Evaluation on the validation set was with case-sensitive BLEU (Papineni et al., 2002) on tokenized text using multi-bleu.perl. Evaluation on the test sets was with case-sensitive BLEU on detokenized text using mteval-v13a.pl. The monolingual training datasets are the News Crawl English and German corpora, each of which has more than a billion tokens.", |
| "cite_spans": [ |
| { |
| "start": 103, |
| "end": 123, |
| "text": "(Bojar et al., 2015)", |
| "ref_id": "BIBREF2" |
| }, |
| { |
| "start": 538, |
| "end": 562, |
| "text": "(Sennrich et al., 2015b)", |
| "ref_id": "BIBREF28" |
| }, |
| { |
| "start": 817, |
| "end": 840, |
| "text": "(Papineni et al., 2002)", |
| "ref_id": "BIBREF22" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "Experimental settings: The language models were trained in the same fashion as (Jozefowicz et al., 2016) We used a 1 layer 4096 dimensional LSTM with the hidden state projected down to 1024 units (Sak et al., 2014) and trained for one week on 32 Tesla K40 GPUs. Our seq2seq model was a 3 layer model, where the second and third layers each have 1000 hidden units. The monolingual objectives, residual connection, and the modified attention were all used. We used the Adam optimizer (Kingma and Ba, 2015) and train with asynchronous SGD on 16 GPUs for speed. We used a learning rate of 5e-5 which is multiplied by 0.8 every 50K steps after an initial 400K steps, gradient clipping with norm 5.0 (Pascanu et al., 2013) , and dropout of 0.2 on non-recurrent connections (Zaremba et al., 2014) . We used early stopping on validation set perplexity. A beam size of 10 was used for decoding. Our ensemble is con-BLEU System ensemble? newstest2014 newstest2015 Phrase Based MT (Williams et al., 2016) -21.9 23.7 Supervised NMT (Jean et al., 2015) single -22.4 Edit Distance Transducer NMT (Stahlberg et al., 2016) single 21.7 24.1 Edit Distance Transducer NMT (Stahlberg et al., 2016) ensemble 8 22.9 25.7 Backtranslation (Sennrich et al., 2015a) single 22.7 25.7 Backtranslation (Sennrich et al., 2015a) ensemble 4 23.8 26.5 Backtranslation (Sennrich et al., 2015a) structed with the 5 best performing models on the validation set, which are trained with different hyperparameters.", |
| "cite_spans": [ |
| { |
| "start": 79, |
| "end": 104, |
| "text": "(Jozefowicz et al., 2016)", |
| "ref_id": "BIBREF14" |
| }, |
| { |
| "start": 196, |
| "end": 214, |
| "text": "(Sak et al., 2014)", |
| "ref_id": "BIBREF26" |
| }, |
| { |
| "start": 482, |
| "end": 503, |
| "text": "(Kingma and Ba, 2015)", |
| "ref_id": "BIBREF16" |
| }, |
| { |
| "start": 694, |
| "end": 716, |
| "text": "(Pascanu et al., 2013)", |
| "ref_id": "BIBREF23" |
| }, |
| { |
| "start": 767, |
| "end": 789, |
| "text": "(Zaremba et al., 2014)", |
| "ref_id": "BIBREF34" |
| }, |
| { |
| "start": 970, |
| "end": 993, |
| "text": "(Williams et al., 2016)", |
| "ref_id": "BIBREF32" |
| }, |
| { |
| "start": 1020, |
| "end": 1039, |
| "text": "(Jean et al., 2015)", |
| "ref_id": "BIBREF13" |
| }, |
| { |
| "start": 1082, |
| "end": 1106, |
| "text": "(Stahlberg et al., 2016)", |
| "ref_id": "BIBREF29" |
| }, |
| { |
| "start": 1153, |
| "end": 1177, |
| "text": "(Stahlberg et al., 2016)", |
| "ref_id": "BIBREF29" |
| }, |
| { |
| "start": 1215, |
| "end": 1239, |
| "text": "(Sennrich et al., 2015a)", |
| "ref_id": "BIBREF27" |
| }, |
| { |
| "start": 1273, |
| "end": 1297, |
| "text": "(Sennrich et al., 2015a)", |
| "ref_id": "BIBREF27" |
| }, |
| { |
| "start": 1335, |
| "end": 1359, |
| "text": "(Sennrich et al., 2015a)", |
| "ref_id": "BIBREF27" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "Results: Table 1 shows the results of our method in comparison with other baselines. Our method achieves a new state-of-the-art for single model performance on both newstest2014 and newstest2015, significantly outperforming the competitive semi-supervised backtranslation technique (Sennrich et al., 2015a) . Equally impressive is the fact that our best single model outperforms the previous state of the art ensemble of 4 models. Our ensemble of 5 models matches or exceeds the previous best ensemble of 12 models.", |
| "cite_spans": [ |
| { |
| "start": 282, |
| "end": 306, |
| "text": "(Sennrich et al., 2015a)", |
| "ref_id": "BIBREF27" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 9, |
| "end": 16, |
| "text": "Table 1", |
| "ref_id": "TABREF2" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "Ablation study: In order to better understand the effects of pretraining, we conducted an ablation study by modifying the pretraining scheme. We were primarily interested in varying the pretraining scheme and the monolingual language modeling objectives because these two techniques produce the largest gains in the model. Figure 3 shows the drop in validation BLEU of various ablations compared with the full model. The full model uses LMs trained with monolingual data to initialize the encoder and decoder, in addition to the language modeling objective. In the follow-ing, we interpret the findings of the study. Note that some findings are specific to the translation task. Given the results from the ablation study, we can make the following observations:", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 323, |
| "end": 332, |
| "text": "Figure 3", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "\u2022 Only pretraining the decoder is better than only pretraining the encoder: Only pretraining the encoder leads to a 1.6 BLEU point drop while only pretraining the decoder leads to a 1.0 BLEU point drop.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "\u2022 Pretrain as much as possible because the benefits compound: given the drops of no pretraining at all (\u22122.0) and only pretraining the encoder (\u22121.6), the additive estimate of the drop of only pretraining the decoder side is \u22122.0 \u2212 (\u22121.6) = \u22120.4; however the actual drop is \u22121.0 which is a much larger drop than the additive estimate.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "\u2022 Pretraining the softmax is important: Pretraining only the embeddings and first LSTM layer gives a large drop of 1.6 BLEU points.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "\u2022 The language modeling objective is a strong regularizer: The drop in BLEU points of pretraining the entire model and not using the LM objective is as bad as using the LM objective without pretraining.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "\u2022 Pretraining on a lot of unlabeled data is essential for learning to extract powerful features: If the model is initialized with LMs that are pretrained on the source part and target part of the parallel corpus, the drop in performance is as large as not pretraining at all. However, performance remains strong when pretrained on the large, nonnews Wikipedia corpus.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "To understand the contributions of unsupervised pretraining vs. supervised training, we track the performance of pretraining as a function of dataset size. For this, we trained a a model with and without pretraining on random subsets of the English\u2192German corpus. Both models use the additional LM objective. The results are summarized in Figure 4 . When a 100% of the labeled data is used, the gap between the pretrained and no pretrain model is 2.0 BLEU points. However, that gap grows when less data is available. When trained on 20% of the labeled data, the gap becomes 3.8 BLEU points. This demonstrates that the pretrained models degrade less as the labeled dataset becomes smaller. ", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 339, |
| "end": 347, |
| "text": "Figure 4", |
| "ref_id": "FIGREF2" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Machine Translation", |
| "sec_num": "3.1" |
| }, |
| { |
| "text": "Dataset and Evaluation: For a low-resource abstractive summarization task, we use the CNN/Daily Mail corpus from (Hermann et al., 2015) . Following Nallapati et al. 2016, we modify the data collection scripts to restore the bullet point summaries. The task is to predict the bullet point summaries from a news article. The dataset has fewer than 300K document-summary pairs. To compare against Nallapati et al. (2016) , we used the anonymized corpus. However, for our ablation study, we used the non-anonymized corpus. 1 We evaluate our system using full length ROUGE (Lin, 2004) . For the anonymized corpus in particular, we considered each highlight as a separate sentence following Nallapati et al. (2016) . In this setting, we used the English Gigaword corpus (Napoles et al., 2012) as our larger, unlabeled \"monolingual\" corpus, although all data used in this task is in English.", |
| "cite_spans": [ |
| { |
| "start": 113, |
| "end": 135, |
| "text": "(Hermann et al., 2015)", |
| "ref_id": "BIBREF12" |
| }, |
| { |
| "start": 394, |
| "end": 417, |
| "text": "Nallapati et al. (2016)", |
| "ref_id": "BIBREF20" |
| }, |
| { |
| "start": 519, |
| "end": 520, |
| "text": "1", |
| "ref_id": null |
| }, |
| { |
| "start": 568, |
| "end": 579, |
| "text": "(Lin, 2004)", |
| "ref_id": "BIBREF17" |
| }, |
| { |
| "start": 685, |
| "end": 708, |
| "text": "Nallapati et al. (2016)", |
| "ref_id": "BIBREF20" |
| }, |
| { |
| "start": 764, |
| "end": 786, |
| "text": "(Napoles et al., 2012)", |
| "ref_id": "BIBREF21" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstractive Summarization", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "Experimental settings: We use subword units (Sennrich et al., 2015b) with 31500 merges, resulting in a vocabulary size of about 32000. We use up to the first 600 tokens of the document and System ROUGE-1 ROUGE-2 ROUGE-L Seq2seq + pretrained embeddings (Nallapati et al., 2016) 32.49 11.84 29.47 + temporal attention (Nallapati et al., 2016) 35 predict the entire summary. Only one language model is trained and it is used to initialize both the encoder and decoder, since the source and target languages are the same. However, the encoder and decoder are not tied. The LM is a one-layer LSTM of size 1024 trained in a similar fashion to Jozefowicz et al. (2016) . For the seq2seq model, we use the same settings as the machine translation experiments. The only differences are that we use a 2 layer model with the second layer having 1024 hidden units, and that the learning rate is multiplied by 0.8 every 30K steps after an initial 100K steps.", |
| "cite_spans": [ |
| { |
| "start": 44, |
| "end": 68, |
| "text": "(Sennrich et al., 2015b)", |
| "ref_id": "BIBREF28" |
| }, |
| { |
| "start": 252, |
| "end": 276, |
| "text": "(Nallapati et al., 2016)", |
| "ref_id": "BIBREF20" |
| }, |
| { |
| "start": 316, |
| "end": 340, |
| "text": "(Nallapati et al., 2016)", |
| "ref_id": "BIBREF20" |
| }, |
| { |
| "start": 637, |
| "end": 661, |
| "text": "Jozefowicz et al. (2016)", |
| "ref_id": "BIBREF14" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstractive Summarization", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "Results: Table 2 summarizes our results on the anonymized version of the corpus. Our pretrained model is only able to match the previous baseline seq2seq of Nallapati et al. (2016) . Interestingly, they use pretrained word2vec vectors to initialize their word em-beddings. As we show in our ablation study, just pretraining the embeddings itself gives a large improvement. Furthermore, our model is a unidirectional LSTM while they use a bidirectional LSTM. They also use a longer context of 800 tokens, whereas we used a context of 600 tokens due to GPU memory issues.", |
| "cite_spans": [ |
| { |
| "start": 157, |
| "end": 180, |
| "text": "Nallapati et al. (2016)", |
| "ref_id": "BIBREF20" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 9, |
| "end": 16, |
| "text": "Table 2", |
| "ref_id": "TABREF4" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Abstractive Summarization", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "Ablation study: We performed an ablation study similar to the one performed on the machine translation model. The results are reported in Figure 5 . Here we report the drops on ROUGE-1, ROUGE-2, and ROUGE-L on the nonanonymized validation set. Given the results from our ablation study, we can make the following observations:", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 138, |
| "end": 146, |
| "text": "Figure 5", |
| "ref_id": "FIGREF3" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Abstractive Summarization", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "\u2022 Pretraining appears to improve optimization: in contrast with the machine translation model, it is more beneficial to only pretrain the encoder than only the decoder of the sum-marization model. One interpretation is that pretraining enables the gradient to flow much further back in time than randomly initialized weights. This may also explain why pretraining on the parallel corpus is no worse than pretraining on a larger monolingual corpus.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstractive Summarization", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "\u2022 The language modeling objective is a strong regularizer: A model without the LM objective has a significant drop in ROUGE scores.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstractive Summarization", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "Human evaluation: As ROUGE may not be able to capture the quality of summarization, we also performed a small qualitative study to understand the human impression of the summaries produced by different models. We took 200 random documents and compared the performance of a pretrained and non-pretrained system. The document, gold summary, and the two system outputs were presented to a human evaluator who was asked to rate each system output on a scale of 1-5 with 5 being the best score. The system outputs were presented in random order and the evaluator did not know the identity of either output. The evaluator noted if there were repetitive phrases or sentences in either system outputs. Unwanted repetition was also noticed by Nallapati et al. (2016) . Table 3 and 4 show the results of the study. In both cases, the pretrained system outperforms the system without pretraining in a statistically significant manner. The better optimization enabled by pretraining improves the generated summaries and decreases unwanted repetition in the output. NP > P NP = P NP < P 29 88 83 Table 3 : The count of how often the no pretrain system (NP) achieves a higher, equal, and lower score than the pretrained system (P) in the side-byside study where the human evaluator gave each system a score from 1-5. The sign statistical test gives a p-value of < 0.0001 for rejecting the null hypothesis that there is no difference in the score obtained by either system.", |
| "cite_spans": [ |
| { |
| "start": 734, |
| "end": 757, |
| "text": "Nallapati et al. (2016)", |
| "ref_id": "BIBREF20" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 760, |
| "end": 767, |
| "text": "Table 3", |
| "ref_id": null |
| }, |
| { |
| "start": 1083, |
| "end": 1090, |
| "text": "Table 3", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Abstractive Summarization", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "Unsupervised pretraining has been intensively studied in the past years, most notably is the work by Dahl et al. (2012) Table 4 : The count of how often the pretrain and no pretrain systems contain repeated phrases or sentences in their outputs in the side-by-side study.", |
| "cite_spans": [ |
| { |
| "start": 101, |
| "end": 119, |
| "text": "Dahl et al. (2012)", |
| "ref_id": "BIBREF7" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 120, |
| "end": 127, |
| "text": "Table 4", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Related Work", |
| "sec_num": "4" |
| }, |
| { |
| "text": "McNemar's test gives a p-value of < 0.0001 for rejecting the null hypothesis that the two systems repeat the same proportion of times. The pretrained system clearly repeats less than the system without pretraining.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Related Work", |
| "sec_num": "4" |
| }, |
| { |
| "text": "acoustic models. More recent acoustic models have found pretraining unnecessary (Xiong et al., 2016; Chan et al., 2015) , probably because the reconstruction objective of deep belief networks is too easy. In contrast, we find that pretraining language models by next step prediction significantly improves seq2seq on challenging real world datasets. Despite its appeal, unsupervised learning has not been widely used to improve supervised training. Dai and Le (2015); Radford et al. (2017) are amongst the rare studies which showed the benefits of pretraining in a semi-supervised learning setting. Their methods are similar to ours except that they did not have a decoder network and thus could not apply to seq2seq learning. Similarly, Zhang and Zong (2016) found it useful to add an additional task of sentence reordering of sourceside monolingual data for neural machine translation. Various forms of transfer or multitask learning with seq2seq framework also have the flavors of our algorithm (Zoph et al., 2016; Luong et al., 2015; Firat et al., 2016) .", |
| "cite_spans": [ |
| { |
| "start": 80, |
| "end": 100, |
| "text": "(Xiong et al., 2016;", |
| "ref_id": "BIBREF33" |
| }, |
| { |
| "start": 101, |
| "end": 119, |
| "text": "Chan et al., 2015)", |
| "ref_id": "BIBREF3" |
| }, |
| { |
| "start": 468, |
| "end": 489, |
| "text": "Radford et al. (2017)", |
| "ref_id": "BIBREF25" |
| }, |
| { |
| "start": 738, |
| "end": 759, |
| "text": "Zhang and Zong (2016)", |
| "ref_id": "BIBREF35" |
| }, |
| { |
| "start": 998, |
| "end": 1017, |
| "text": "(Zoph et al., 2016;", |
| "ref_id": "BIBREF37" |
| }, |
| { |
| "start": 1018, |
| "end": 1037, |
| "text": "Luong et al., 2015;", |
| "ref_id": "BIBREF18" |
| }, |
| { |
| "start": 1038, |
| "end": 1057, |
| "text": "Firat et al., 2016)", |
| "ref_id": "BIBREF9" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Related Work", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Perhaps most closely related to our method is the work by Gulcehre et al. (2015) , who combined a language model with an already trained seq2seq model by fine-tuning additional deep output layers. Empirically, their method produces small improvements over the supervised baseline. We suspect that their method does not produce significant gains because (i) the models are trained independently of each other and are not fine-tuned (ii) the LM is combined with the seq2seq model after the last layer, wasting the benefit of the low level LM features, and (iii) only using the LM on the decoder side. Venugopalan et al. (2016) addressed (i) but still experienced minor improvements. Using pretrained GloVe embedding vectors (Pennington et al., 2014) had more impact.", |
| "cite_spans": [ |
| { |
| "start": 58, |
| "end": 80, |
| "text": "Gulcehre et al. (2015)", |
| "ref_id": "BIBREF11" |
| }, |
| { |
| "start": 599, |
| "end": 624, |
| "text": "Venugopalan et al. (2016)", |
| "ref_id": "BIBREF31" |
| }, |
| { |
| "start": 722, |
| "end": 747, |
| "text": "(Pennington et al., 2014)", |
| "ref_id": "BIBREF24" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Related Work", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Related to our approach in principle is the work by Chen et al. (2016) who proposed a two-term, theoretically motivated unsupervised objective for unpaired input-output samples. Though they did not apply their method to seq2seq learning, their framework can be modified to do so. In that case, the first term pushes the output to be highly probable under some scoring model, and the second term ensures that the output depends on the input. In the seq2seq setting, we interpret the first term as a pretrained language model scoring the output sequence. In our work, we fold the pretrained language model into the decoder. We believe that using the pretrained language model only for scoring is less efficient that using all the pretrained weights. Our use of labeled examples satisfies the second term. These connections provide a theoretical grounding for our work.", |
| "cite_spans": [ |
| { |
| "start": 52, |
| "end": 70, |
| "text": "Chen et al. (2016)", |
| "ref_id": "BIBREF4" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Related Work", |
| "sec_num": "4" |
| }, |
| { |
| "text": "In our experiments, we benchmark our method on machine translation, where other unsupervised methods are shown to give promising results (Sennrich et al., 2015a; Cheng et al., 2016) . In backtranslation (Sennrich et al., 2015a) , the trained model is used to decode unlabeled data to yield extra labeled data. One can argue that this method may not have a natural analogue to other tasks such as summarization. We note that their technique is complementary to ours, and may lead to additional gains in machine translation. The method of using autoencoders in Cheng et al. (2016) is promising, though it can be argued that autoencoding is an easy objective and language modeling may force the unsupervised models to learn better features.", |
| "cite_spans": [ |
| { |
| "start": 137, |
| "end": 161, |
| "text": "(Sennrich et al., 2015a;", |
| "ref_id": "BIBREF27" |
| }, |
| { |
| "start": 162, |
| "end": 181, |
| "text": "Cheng et al., 2016)", |
| "ref_id": "BIBREF5" |
| }, |
| { |
| "start": 203, |
| "end": 227, |
| "text": "(Sennrich et al., 2015a)", |
| "ref_id": "BIBREF27" |
| }, |
| { |
| "start": 559, |
| "end": 578, |
| "text": "Cheng et al. (2016)", |
| "ref_id": "BIBREF5" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Related Work", |
| "sec_num": "4" |
| }, |
| { |
| "text": "We presented a novel unsupervised pretraining method to improve sequence to sequence learning. The method can aid in both generalization and optimization. Our scheme involves pretraining two language models in the source and target domain, and initializing the embeddings, first LSTM layers, and softmax of a sequence to sequence model with the weights of the language models. Using our method, we achieved state-of-the-art machine translation results on both WMT'14 and WMT'15 English to German. A key advantage of this technique is that it is flexible and can be applied to a large variety of tasks.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Conclusion", |
| "sec_num": "5" |
| }, |
| { |
| "text": "We encourage future researchers to use the nonanonymized version because it is a more realistic summarization setting with a larger vocabulary. Our numbers on the non-anonymized test set are 35.56 ROUGE-1, 14.60 ROUGE-2, and 25.08 ROUGE-L. We did not consider highlights as separate sentences.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Several studies on natural language and back-propagation", |
| "authors": [ |
| { |
| "first": "Robert", |
| "middle": [ |
| "B" |
| ], |
| "last": "Allen", |
| "suffix": "" |
| } |
| ], |
| "year": 1987, |
| "venue": "IEEE First International Conference on Neural Networks", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Robert B. Allen. 1987. Several studies on natural lan- guage and back-propagation. IEEE First Interna- tional Conference on Neural Networks.", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Neural machine translation by jointly learning to align and translate", |
| "authors": [ |
| { |
| "first": "Dzmitry", |
| "middle": [], |
| "last": "Bahdanau", |
| "suffix": "" |
| }, |
| { |
| "first": "Kyunghyun", |
| "middle": [], |
| "last": "Cho", |
| "suffix": "" |
| }, |
| { |
| "first": "Yoshua", |
| "middle": [], |
| "last": "Bengio", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "ICLR", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Ben- gio. 2015. Neural machine translation by jointly learning to align and translate. In ICLR.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "Findings of the 2015 workshop on statistical machine translation", |
| "authors": [ |
| { |
| "first": "Ond\u0159ej", |
| "middle": [], |
| "last": "Bojar", |
| "suffix": "" |
| }, |
| { |
| "first": "Rajen", |
| "middle": [], |
| "last": "Chatterjee", |
| "suffix": "" |
| }, |
| { |
| "first": "Christian", |
| "middle": [], |
| "last": "Federmann", |
| "suffix": "" |
| }, |
| { |
| "first": "Barry", |
| "middle": [], |
| "last": "Haddow", |
| "suffix": "" |
| }, |
| { |
| "first": "Matthias", |
| "middle": [], |
| "last": "Huck", |
| "suffix": "" |
| }, |
| { |
| "first": "Chris", |
| "middle": [], |
| "last": "Hokamp", |
| "suffix": "" |
| }, |
| { |
| "first": "Philipp", |
| "middle": [], |
| "last": "Koehn", |
| "suffix": "" |
| }, |
| { |
| "first": "Varvara", |
| "middle": [], |
| "last": "Logacheva", |
| "suffix": "" |
| }, |
| { |
| "first": "Christof", |
| "middle": [], |
| "last": "Monz", |
| "suffix": "" |
| }, |
| { |
| "first": "Matteo", |
| "middle": [], |
| "last": "Negri", |
| "suffix": "" |
| }, |
| { |
| "first": "Matt", |
| "middle": [], |
| "last": "Post", |
| "suffix": "" |
| }, |
| { |
| "first": "Carolina", |
| "middle": [], |
| "last": "Scarton", |
| "suffix": "" |
| }, |
| { |
| "first": "Lucia", |
| "middle": [], |
| "last": "Specia", |
| "suffix": "" |
| }, |
| { |
| "first": "Marco", |
| "middle": [], |
| "last": "Turchi", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "Proceedings of the Tenth Workshop on Statistical Machine Translation", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ond\u0159ej Bojar, Rajen Chatterjee, Christian Federmann, Barry Haddow, Matthias Huck, Chris Hokamp, Philipp Koehn, Varvara Logacheva, Christof Monz, Matteo Negri, Matt Post, Carolina Scarton, Lucia Specia, and Marco Turchi. 2015. Findings of the 2015 workshop on statistical machine translation. In Proceedings of the Tenth Workshop on Statistical Machine Translation.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Listen, attend and spell", |
| "authors": [ |
| { |
| "first": "William", |
| "middle": [], |
| "last": "Chan", |
| "suffix": "" |
| }, |
| { |
| "first": "Navdeep", |
| "middle": [], |
| "last": "Jaitly", |
| "suffix": "" |
| }, |
| { |
| "first": "Quoc", |
| "middle": [ |
| "V" |
| ], |
| "last": "Le", |
| "suffix": "" |
| }, |
| { |
| "first": "Oriol", |
| "middle": [], |
| "last": "Vinyals", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1508.01211" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "William Chan, Navdeep Jaitly, Quoc V. Le, and Oriol Vinyals. 2015. Listen, attend and spell. arXiv preprint arXiv:1508.01211.", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "Unsupervised learning of predictors from unpaired input-output samples", |
| "authors": [ |
| { |
| "first": "Jianshu", |
| "middle": [], |
| "last": "Chen", |
| "suffix": "" |
| }, |
| { |
| "first": "Po-Sen", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| }, |
| { |
| "first": "Xiaodong", |
| "middle": [], |
| "last": "He", |
| "suffix": "" |
| }, |
| { |
| "first": "Jianfeng", |
| "middle": [], |
| "last": "Gao", |
| "suffix": "" |
| }, |
| { |
| "first": "Li", |
| "middle": [], |
| "last": "Deng", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jianshu Chen, Po-Sen Huang, Xiaodong He, Jianfeng Gao, and Li Deng. 2016. Unsupervised learning of predictors from unpaired input-output samples. abs/1606.04646.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Semisupervised learning for neural machine translation", |
| "authors": [ |
| { |
| "first": "Yong", |
| "middle": [], |
| "last": "Cheng", |
| "suffix": "" |
| }, |
| { |
| "first": "Wei", |
| "middle": [], |
| "last": "Xu", |
| "suffix": "" |
| }, |
| { |
| "first": "Zhongjun", |
| "middle": [], |
| "last": "He", |
| "suffix": "" |
| }, |
| { |
| "first": "Wei", |
| "middle": [], |
| "last": "He", |
| "suffix": "" |
| }, |
| { |
| "first": "Hua", |
| "middle": [], |
| "last": "Wu", |
| "suffix": "" |
| }, |
| { |
| "first": "Maosong", |
| "middle": [], |
| "last": "Sun", |
| "suffix": "" |
| }, |
| { |
| "first": "Yang", |
| "middle": [], |
| "last": "Liu", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1606.04596" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Yong Cheng, Wei Xu, Zhongjun He, Wei He, Hua Wu, Maosong Sun, and Yang Liu. 2016. Semi- supervised learning for neural machine translation. arXiv preprint arXiv:1606.04596.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Learning phrase representations using RNN encoder-decoder for statistical machine translation", |
| "authors": [ |
| { |
| "first": "Kyunghyun", |
| "middle": [], |
| "last": "Cho", |
| "suffix": "" |
| }, |
| { |
| "first": "Bart", |
| "middle": [], |
| "last": "Van Merri\u00ebnboer", |
| "suffix": "" |
| }, |
| { |
| "first": "Caglar", |
| "middle": [], |
| "last": "Gulcehre", |
| "suffix": "" |
| }, |
| { |
| "first": "Dzmitry", |
| "middle": [], |
| "last": "Bahdanau", |
| "suffix": "" |
| }, |
| { |
| "first": "Fethi", |
| "middle": [], |
| "last": "Bougares", |
| "suffix": "" |
| }, |
| { |
| "first": "Holger", |
| "middle": [], |
| "last": "Schwenk", |
| "suffix": "" |
| }, |
| { |
| "first": "Yoshua", |
| "middle": [], |
| "last": "Bengio", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "EMNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Kyunghyun Cho, Bart Van Merri\u00ebnboer, Caglar Gul- cehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning phrase representations using RNN encoder-decoder for statistical machine translation. In EMNLP.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "Context-dependent pre-trained deep neural networks for large-vocabulary speech recognition", |
| "authors": [ |
| { |
| "first": "G", |
| "middle": [ |
| "E" |
| ], |
| "last": "Dahl", |
| "suffix": "" |
| }, |
| { |
| "first": "D", |
| "middle": [], |
| "last": "Yu", |
| "suffix": "" |
| }, |
| { |
| "first": "L", |
| "middle": [], |
| "last": "Deng", |
| "suffix": "" |
| }, |
| { |
| "first": "A", |
| "middle": [], |
| "last": "Acero", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "IEEE Transactions on Audio, Speech, and Language Processing", |
| "volume": "20", |
| "issue": "1", |
| "pages": "30--42", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "G. E. Dahl, D. Yu, L. Deng, and A. Acero. 2012. Context-dependent pre-trained deep neural networks for large-vocabulary speech recognition. IEEE Transactions on Audio, Speech, and Language Pro- cessing, 20(1):30-42.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Semisupervised sequence learning", |
| "authors": [ |
| { |
| "first": "M", |
| "middle": [], |
| "last": "Andrew", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Dai", |
| "suffix": "" |
| }, |
| { |
| "first": "V", |
| "middle": [], |
| "last": "Quoc", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Le", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "NIPS", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Andrew M. Dai and Quoc V. Le. 2015. Semi- supervised sequence learning. In NIPS.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "Zero-resource translation with multilingual neural machine translation", |
| "authors": [ |
| { |
| "first": "Orhan", |
| "middle": [], |
| "last": "Firat", |
| "suffix": "" |
| }, |
| { |
| "first": "Baskaran", |
| "middle": [], |
| "last": "Sankaran", |
| "suffix": "" |
| }, |
| { |
| "first": "Yaser", |
| "middle": [], |
| "last": "Al-Onaizan", |
| "suffix": "" |
| }, |
| { |
| "first": "T", |
| "middle": [], |
| "last": "Fatos", |
| "suffix": "" |
| }, |
| { |
| "first": "Kyunghyun", |
| "middle": [], |
| "last": "Yarman-Vural", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Cho", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1606.04164" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Orhan Firat, Baskaran Sankaran, Yaser Al-Onaizan, Fatos T. Yarman-Vural, and Kyunghyun Cho. 2016. Zero-resource translation with multi- lingual neural machine translation. arXiv preprint arXiv:1606.04164.", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "An empirical investigation of catastrophic forgetting in gradient-based neural networks", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Ian", |
| "suffix": "" |
| }, |
| { |
| "first": "Mehdi", |
| "middle": [], |
| "last": "Goodfellow", |
| "suffix": "" |
| }, |
| { |
| "first": "Da", |
| "middle": [], |
| "last": "Mirza", |
| "suffix": "" |
| }, |
| { |
| "first": "Aaron", |
| "middle": [], |
| "last": "Xiao", |
| "suffix": "" |
| }, |
| { |
| "first": "Yoshua", |
| "middle": [], |
| "last": "Courville", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Bengio", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1312.6211" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ian J Goodfellow, Mehdi Mirza, Da Xiao, Aaron Courville, and Yoshua Bengio. 2013. An em- pirical investigation of catastrophic forgetting in gradient-based neural networks. arXiv preprint arXiv:1312.6211.", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "On using monolingual corpora in neural machine translation", |
| "authors": [ |
| { |
| "first": "Caglar", |
| "middle": [], |
| "last": "Gulcehre", |
| "suffix": "" |
| }, |
| { |
| "first": "Orhan", |
| "middle": [], |
| "last": "Firat", |
| "suffix": "" |
| }, |
| { |
| "first": "Kelvin", |
| "middle": [], |
| "last": "Xu", |
| "suffix": "" |
| }, |
| { |
| "first": "Kyunghyun", |
| "middle": [], |
| "last": "Cho", |
| "suffix": "" |
| }, |
| { |
| "first": "Loic", |
| "middle": [], |
| "last": "Barrault", |
| "suffix": "" |
| }, |
| { |
| "first": "Huei-Chi", |
| "middle": [], |
| "last": "Lin", |
| "suffix": "" |
| }, |
| { |
| "first": "Fethi", |
| "middle": [], |
| "last": "Bougares", |
| "suffix": "" |
| }, |
| { |
| "first": "Holger", |
| "middle": [], |
| "last": "Schwenk", |
| "suffix": "" |
| }, |
| { |
| "first": "Yoshua", |
| "middle": [], |
| "last": "Bengio", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1503.03535" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Caglar Gulcehre, Orhan Firat, Kelvin Xu, Kyunghyun Cho, Loic Barrault, Huei-Chi Lin, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2015. On us- ing monolingual corpora in neural machine transla- tion. arXiv preprint arXiv:1503.03535.", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "Teaching machines to read and comprehend", |
| "authors": [ |
| { |
| "first": "Karl", |
| "middle": [], |
| "last": "Moritz Hermann", |
| "suffix": "" |
| }, |
| { |
| "first": "Tomas", |
| "middle": [], |
| "last": "Kocisky", |
| "suffix": "" |
| }, |
| { |
| "first": "Edward", |
| "middle": [], |
| "last": "Grefenstette", |
| "suffix": "" |
| }, |
| { |
| "first": "Lasse", |
| "middle": [], |
| "last": "Espeholt", |
| "suffix": "" |
| }, |
| { |
| "first": "Will", |
| "middle": [], |
| "last": "Kay", |
| "suffix": "" |
| }, |
| { |
| "first": "Mustafa", |
| "middle": [], |
| "last": "Suleyman", |
| "suffix": "" |
| }, |
| { |
| "first": "Phil", |
| "middle": [], |
| "last": "Blunsom", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "NIPS", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Karl Moritz Hermann, Tomas Kocisky, Edward Grefenstette, Lasse Espeholt, Will Kay, Mustafa Su- leyman, and Phil Blunsom. 2015. Teaching ma- chines to read and comprehend. In NIPS.", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "Montreal neural machine translation systems for WMT'15", |
| "authors": [ |
| { |
| "first": "S\u00e9bastien", |
| "middle": [], |
| "last": "Jean", |
| "suffix": "" |
| }, |
| { |
| "first": "Orhan", |
| "middle": [], |
| "last": "Firat", |
| "suffix": "" |
| }, |
| { |
| "first": "Kyunghyun", |
| "middle": [], |
| "last": "Cho", |
| "suffix": "" |
| }, |
| { |
| "first": "Roland", |
| "middle": [], |
| "last": "Memisevic", |
| "suffix": "" |
| }, |
| { |
| "first": "Yoshua", |
| "middle": [], |
| "last": "Bengio", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "Proceedings of the Tenth Workshop on Statistical Machine Translation", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "S\u00e9bastien Jean, Orhan Firat, Kyunghyun Cho, Roland Memisevic, and Yoshua Bengio. 2015. Montreal neural machine translation systems for WMT'15. In Proceedings of the Tenth Workshop on Statistical Machine Translation.", |
| "links": null |
| }, |
| "BIBREF14": { |
| "ref_id": "b14", |
| "title": "Exploring the limits of language modeling", |
| "authors": [ |
| { |
| "first": "Rafal", |
| "middle": [], |
| "last": "Jozefowicz", |
| "suffix": "" |
| }, |
| { |
| "first": "Oriol", |
| "middle": [], |
| "last": "Vinyals", |
| "suffix": "" |
| }, |
| { |
| "first": "Mike", |
| "middle": [], |
| "last": "Schuster", |
| "suffix": "" |
| }, |
| { |
| "first": "Noam", |
| "middle": [], |
| "last": "Shazeer", |
| "suffix": "" |
| }, |
| { |
| "first": "Yonghui", |
| "middle": [], |
| "last": "Wu", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1602.02410" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rafal Jozefowicz, Oriol Vinyals, Mike Schuster, Noam Shazeer, and Yonghui Wu. 2016. Exploring the limits of language modeling. arXiv preprint arXiv:1602.02410.", |
| "links": null |
| }, |
| "BIBREF15": { |
| "ref_id": "b15", |
| "title": "Recurrent continuous translation models", |
| "authors": [ |
| { |
| "first": "Nal", |
| "middle": [], |
| "last": "Kalchbrenner", |
| "suffix": "" |
| }, |
| { |
| "first": "Phil", |
| "middle": [], |
| "last": "Blunsom", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "EMNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Nal Kalchbrenner and Phil Blunsom. 2013. Recurrent continuous translation models. In EMNLP.", |
| "links": null |
| }, |
| "BIBREF16": { |
| "ref_id": "b16", |
| "title": "Adam: A method for stochastic optimization", |
| "authors": [ |
| { |
| "first": "Diederik", |
| "middle": [], |
| "last": "Kingma", |
| "suffix": "" |
| }, |
| { |
| "first": "Jimmy", |
| "middle": [], |
| "last": "Ba", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "ICLR", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Diederik Kingma and Jimmy Ba. 2015. Adam: A method for stochastic optimization. In ICLR.", |
| "links": null |
| }, |
| "BIBREF17": { |
| "ref_id": "b17", |
| "title": "ROUGE: a package for automatic evaluation of summaries", |
| "authors": [ |
| { |
| "first": "Chin-Yew", |
| "middle": [], |
| "last": "Lin", |
| "suffix": "" |
| } |
| ], |
| "year": 2004, |
| "venue": "Proceedings of the Workshop on Text Summarization Branches Out", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Chin-Yew Lin. 2004. ROUGE: a package for auto- matic evaluation of summaries. In Proceedings of the Workshop on Text Summarization Branches Out (WAS 2004).", |
| "links": null |
| }, |
| "BIBREF18": { |
| "ref_id": "b18", |
| "title": "Multi-task sequence to sequence learning", |
| "authors": [ |
| { |
| "first": "Minh-Thang", |
| "middle": [], |
| "last": "Luong", |
| "suffix": "" |
| }, |
| { |
| "first": "Quoc", |
| "middle": [ |
| "V" |
| ], |
| "last": "Le", |
| "suffix": "" |
| }, |
| { |
| "first": "Ilya", |
| "middle": [], |
| "last": "Sutskever", |
| "suffix": "" |
| }, |
| { |
| "first": "Oriol", |
| "middle": [], |
| "last": "Vinyals", |
| "suffix": "" |
| }, |
| { |
| "first": "Lukasz", |
| "middle": [], |
| "last": "Kaiser", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "ICLR", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Minh-Thang Luong, Quoc V. Le, Ilya Sutskever, Oriol Vinyals, and Lukasz Kaiser. 2015. Multi-task se- quence to sequence learning. In ICLR.", |
| "links": null |
| }, |
| "BIBREF19": { |
| "ref_id": "b19", |
| "title": "Distributed representations of words and phrases and their compositionality", |
| "authors": [ |
| { |
| "first": "Tomas", |
| "middle": [], |
| "last": "Mikolov", |
| "suffix": "" |
| }, |
| { |
| "first": "Ilya", |
| "middle": [], |
| "last": "Sutskever", |
| "suffix": "" |
| }, |
| { |
| "first": "Kai", |
| "middle": [], |
| "last": "Chen", |
| "suffix": "" |
| }, |
| { |
| "first": "Greg", |
| "middle": [ |
| "S" |
| ], |
| "last": "Corrado", |
| "suffix": "" |
| }, |
| { |
| "first": "Jeff", |
| "middle": [], |
| "last": "Dean", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "NIPS", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S. Cor- rado, and Jeff Dean. 2013. Distributed representa- tions of words and phrases and their compositional- ity. In NIPS.", |
| "links": null |
| }, |
| "BIBREF20": { |
| "ref_id": "b20", |
| "title": "Sequence-to-sequence RNNs for text summarization", |
| "authors": [ |
| { |
| "first": "Ramesh", |
| "middle": [], |
| "last": "Nallapati", |
| "suffix": "" |
| }, |
| { |
| "first": "Bing", |
| "middle": [], |
| "last": "Xiang", |
| "suffix": "" |
| }, |
| { |
| "first": "Bowen", |
| "middle": [], |
| "last": "Zhou", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1602.06023" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ramesh Nallapati, Bing Xiang, and Bowen Zhou. 2016. Sequence-to-sequence RNNs for text summa- rization. arXiv preprint arXiv:1602.06023.", |
| "links": null |
| }, |
| "BIBREF21": { |
| "ref_id": "b21", |
| "title": "Annotated gigaword", |
| "authors": [ |
| { |
| "first": "Courtney", |
| "middle": [], |
| "last": "Napoles", |
| "suffix": "" |
| }, |
| { |
| "first": "Matthew", |
| "middle": [], |
| "last": "Gormley", |
| "suffix": "" |
| }, |
| { |
| "first": "Benjamin", |
| "middle": [], |
| "last": "Van Durme", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "Proceedings of the Joint Workshop on Automatic Knowledge Base Construction and Web-scale Knowledge Extraction. ACL", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Courtney Napoles, Matthew Gormley, and Benjamin Van Durme. 2012. Annotated gigaword. In Pro- ceedings of the Joint Workshop on Automatic Knowl- edge Base Construction and Web-scale Knowledge Extraction. ACL.", |
| "links": null |
| }, |
| "BIBREF22": { |
| "ref_id": "b22", |
| "title": "BLEU: A method for automatic evaluation of machine translation", |
| "authors": [ |
| { |
| "first": "Kishore", |
| "middle": [], |
| "last": "Papineni", |
| "suffix": "" |
| }, |
| { |
| "first": "Salim", |
| "middle": [], |
| "last": "Roukos", |
| "suffix": "" |
| }, |
| { |
| "first": "Todd", |
| "middle": [], |
| "last": "Ward", |
| "suffix": "" |
| }, |
| { |
| "first": "Wei-Jing", |
| "middle": [], |
| "last": "Zhu", |
| "suffix": "" |
| } |
| ], |
| "year": 2002, |
| "venue": "ACL", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Kishore Papineni, Salim Roukos, Todd Ward, and Wei- Jing Zhu. 2002. BLEU: A method for automatic evaluation of machine translation. In ACL.", |
| "links": null |
| }, |
| "BIBREF23": { |
| "ref_id": "b23", |
| "title": "On the difficulty of training recurrent neural networks. ICML", |
| "authors": [ |
| { |
| "first": "Razvan", |
| "middle": [], |
| "last": "Pascanu", |
| "suffix": "" |
| }, |
| { |
| "first": "Tomas", |
| "middle": [], |
| "last": "Mikolov", |
| "suffix": "" |
| }, |
| { |
| "first": "Yoshua", |
| "middle": [], |
| "last": "Bengio", |
| "suffix": "" |
| } |
| ], |
| "year": 2013, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Razvan Pascanu, Tomas Mikolov, and Yoshua Bengio. 2013. On the difficulty of training recurrent neural networks. ICML.", |
| "links": null |
| }, |
| "BIBREF24": { |
| "ref_id": "b24", |
| "title": "Glove: Global vectors for word representation", |
| "authors": [ |
| { |
| "first": "Jeffrey", |
| "middle": [], |
| "last": "Pennington", |
| "suffix": "" |
| }, |
| { |
| "first": "Richard", |
| "middle": [], |
| "last": "Socher", |
| "suffix": "" |
| }, |
| { |
| "first": "Christopher", |
| "middle": [ |
| "D" |
| ], |
| "last": "Manning", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "EMNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jeffrey Pennington, Richard Socher, and Christo- pher D. Manning. 2014. Glove: Global vectors for word representation. In EMNLP.", |
| "links": null |
| }, |
| "BIBREF25": { |
| "ref_id": "b25", |
| "title": "Learning to generate reviews and discovering sentiment", |
| "authors": [ |
| { |
| "first": "Alec", |
| "middle": [], |
| "last": "Radford", |
| "suffix": "" |
| }, |
| { |
| "first": "Rafal", |
| "middle": [], |
| "last": "Jozefowicz", |
| "suffix": "" |
| }, |
| { |
| "first": "Ilya", |
| "middle": [], |
| "last": "Sutskever", |
| "suffix": "" |
| } |
| ], |
| "year": 2017, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1704.01444" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Alec Radford, Rafal Jozefowicz, and Ilya Sutskever. 2017. Learning to generate reviews and discovering sentiment. arXiv preprint arXiv:1704.01444.", |
| "links": null |
| }, |
| "BIBREF26": { |
| "ref_id": "b26", |
| "title": "Long short-term memory recurrent neural network architectures for large scale acoustic modeling", |
| "authors": [ |
| { |
| "first": "Hasim", |
| "middle": [], |
| "last": "Sak", |
| "suffix": "" |
| }, |
| { |
| "first": "Andrew", |
| "middle": [ |
| "W" |
| ], |
| "last": "Senior", |
| "suffix": "" |
| }, |
| { |
| "first": "Fran\u00e7oise", |
| "middle": [], |
| "last": "Beaufays", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "INTERSPEECH", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Hasim Sak, Andrew W. Senior, and Fran\u00e7oise Bea- ufays. 2014. Long short-term memory recurrent neural network architectures for large scale acoustic modeling. In INTERSPEECH.", |
| "links": null |
| }, |
| "BIBREF27": { |
| "ref_id": "b27", |
| "title": "Improving neural machine translation models with monolingual data", |
| "authors": [ |
| { |
| "first": "Rico", |
| "middle": [], |
| "last": "Sennrich", |
| "suffix": "" |
| }, |
| { |
| "first": "Barry", |
| "middle": [], |
| "last": "Haddow", |
| "suffix": "" |
| }, |
| { |
| "first": "Alexandra", |
| "middle": [], |
| "last": "Birch", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1511.06709" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rico Sennrich, Barry Haddow, and Alexandra Birch. 2015a. Improving neural machine translation models with monolingual data. arXiv preprint arXiv:1511.06709.", |
| "links": null |
| }, |
| "BIBREF28": { |
| "ref_id": "b28", |
| "title": "Neural machine translation of rare words with subword units", |
| "authors": [ |
| { |
| "first": "Rico", |
| "middle": [], |
| "last": "Sennrich", |
| "suffix": "" |
| }, |
| { |
| "first": "Barry", |
| "middle": [], |
| "last": "Haddow", |
| "suffix": "" |
| }, |
| { |
| "first": "Alexandra", |
| "middle": [], |
| "last": "Birch", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1508.07909" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rico Sennrich, Barry Haddow, and Alexandra Birch. 2015b. Neural machine translation of rare words with subword units. arXiv preprint arXiv:1508.07909.", |
| "links": null |
| }, |
| "BIBREF29": { |
| "ref_id": "b29", |
| "title": "The edit distance transducer in action: The university of cambridge english-german system at wmt16", |
| "authors": [ |
| { |
| "first": "Felix", |
| "middle": [], |
| "last": "Stahlberg", |
| "suffix": "" |
| }, |
| { |
| "first": "Eva", |
| "middle": [], |
| "last": "Hasler", |
| "suffix": "" |
| }, |
| { |
| "first": "Bill", |
| "middle": [], |
| "last": "Byrne", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of the First Conference on Machine Translation", |
| "volume": "", |
| "issue": "", |
| "pages": "377--384", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Felix Stahlberg, Eva Hasler, and Bill Byrne. 2016. The edit distance transducer in action: The univer- sity of cambridge english-german system at wmt16. In Proceedings of the First Conference on Machine Translation, pages 377-384, Berlin, Germany. As- sociation for Computational Linguistics.", |
| "links": null |
| }, |
| "BIBREF30": { |
| "ref_id": "b30", |
| "title": "Sequence to sequence learning with neural networks", |
| "authors": [ |
| { |
| "first": "Ilya", |
| "middle": [], |
| "last": "Sutskever", |
| "suffix": "" |
| }, |
| { |
| "first": "Oriol", |
| "middle": [], |
| "last": "Vinyals", |
| "suffix": "" |
| }, |
| { |
| "first": "V", |
| "middle": [], |
| "last": "Quoc", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Le", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "NIPS", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ilya Sutskever, Oriol Vinyals, and Quoc V. Le. 2014. Sequence to sequence learning with neural net- works. In NIPS.", |
| "links": null |
| }, |
| "BIBREF31": { |
| "ref_id": "b31", |
| "title": "Improving LSTM-based video description with linguistic knowledge mined from text", |
| "authors": [ |
| { |
| "first": "Subhashini", |
| "middle": [], |
| "last": "Venugopalan", |
| "suffix": "" |
| }, |
| { |
| "first": "Lisa", |
| "middle": [ |
| "Anne" |
| ], |
| "last": "Hendricks", |
| "suffix": "" |
| }, |
| { |
| "first": "Raymond", |
| "middle": [], |
| "last": "Mooney", |
| "suffix": "" |
| }, |
| { |
| "first": "Kate", |
| "middle": [], |
| "last": "Saenko", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1604.01729" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Subhashini Venugopalan, Lisa Anne Hendricks, Ray- mond Mooney, and Kate Saenko. 2016. Improv- ing LSTM-based video description with linguis- tic knowledge mined from text. arXiv preprint arXiv:1604.01729.", |
| "links": null |
| }, |
| "BIBREF32": { |
| "ref_id": "b32", |
| "title": "Edinburgh's statistical machine translation systems for wmt16", |
| "authors": [ |
| { |
| "first": "Philip", |
| "middle": [], |
| "last": "Williams", |
| "suffix": "" |
| }, |
| { |
| "first": "Rico", |
| "middle": [], |
| "last": "Sennrich", |
| "suffix": "" |
| }, |
| { |
| "first": "Maria", |
| "middle": [], |
| "last": "Nadejde", |
| "suffix": "" |
| }, |
| { |
| "first": "Matthias", |
| "middle": [], |
| "last": "Huck", |
| "suffix": "" |
| }, |
| { |
| "first": "Barry", |
| "middle": [], |
| "last": "Haddow", |
| "suffix": "" |
| }, |
| { |
| "first": "Ond\u0159ej", |
| "middle": [], |
| "last": "Bojar", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "Proceedings of the First Conference on Machine Translation", |
| "volume": "", |
| "issue": "", |
| "pages": "399--410", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Philip Williams, Rico Sennrich, Maria Nadejde, Matthias Huck, Barry Haddow, and Ond\u0159ej Bojar. 2016. Edinburgh's statistical machine translation systems for wmt16. In Proceedings of the First Conference on Machine Translation, pages 399- 410, Berlin, Germany. Association for Computa- tional Linguistics.", |
| "links": null |
| }, |
| "BIBREF33": { |
| "ref_id": "b33", |
| "title": "Achieving human parity in conversational speech recognition", |
| "authors": [ |
| { |
| "first": "Wayne", |
| "middle": [], |
| "last": "Xiong", |
| "suffix": "" |
| }, |
| { |
| "first": "Jasha", |
| "middle": [], |
| "last": "Droppo", |
| "suffix": "" |
| }, |
| { |
| "first": "Xuedong", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| }, |
| { |
| "first": "Frank", |
| "middle": [], |
| "last": "Seide", |
| "suffix": "" |
| }, |
| { |
| "first": "Mike", |
| "middle": [], |
| "last": "Seltzer", |
| "suffix": "" |
| }, |
| { |
| "first": "Andreas", |
| "middle": [], |
| "last": "Stolcke", |
| "suffix": "" |
| }, |
| { |
| "first": "Dong", |
| "middle": [], |
| "last": "Yu", |
| "suffix": "" |
| }, |
| { |
| "first": "Geoffrey", |
| "middle": [], |
| "last": "Zweig", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Wayne Xiong, Jasha Droppo, Xuedong Huang, Frank Seide, Mike Seltzer, Andreas Stolcke, Dong Yu, and Geoffrey Zweig. 2016. Achieving human parity in conversational speech recognition. abs/1610.05256.", |
| "links": null |
| }, |
| "BIBREF34": { |
| "ref_id": "b34", |
| "title": "Recurrent neural network regularization", |
| "authors": [ |
| { |
| "first": "Wojciech", |
| "middle": [], |
| "last": "Zaremba", |
| "suffix": "" |
| }, |
| { |
| "first": "Ilya", |
| "middle": [], |
| "last": "Sutskever", |
| "suffix": "" |
| }, |
| { |
| "first": "Oriol", |
| "middle": [], |
| "last": "Vinyals", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "arXiv": [ |
| "arXiv:1409.2329" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Wojciech Zaremba, Ilya Sutskever, and Oriol Vinyals. 2014. Recurrent neural network regularization. arXiv preprint arXiv:1409.2329.", |
| "links": null |
| }, |
| "BIBREF35": { |
| "ref_id": "b35", |
| "title": "Exploiting source-side monolingual data in neural machine translation", |
| "authors": [ |
| { |
| "first": "Jiajun", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| }, |
| { |
| "first": "Chengqing", |
| "middle": [], |
| "last": "Zong", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "EMNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jiajun Zhang and Chengqing Zong. 2016. Exploit- ing source-side monolingual data in neural machine translation. In EMNLP.", |
| "links": null |
| }, |
| "BIBREF36": { |
| "ref_id": "b36", |
| "title": "Very deep convolutional networks for end-to-end speech recognition", |
| "authors": [ |
| { |
| "first": "Yu", |
| "middle": [], |
| "last": "Zhang", |
| "suffix": "" |
| }, |
| { |
| "first": "William", |
| "middle": [], |
| "last": "Chan", |
| "suffix": "" |
| }, |
| { |
| "first": "Navdeep", |
| "middle": [], |
| "last": "Jaitly", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Yu Zhang, William Chan, and Navdeep Jaitly. 2016. Very deep convolutional networks for end-to-end speech recognition. abs/1610.03022.", |
| "links": null |
| }, |
| "BIBREF37": { |
| "ref_id": "b37", |
| "title": "Transfer learning for low-resource neural machine translation", |
| "authors": [ |
| { |
| "first": "Barret", |
| "middle": [], |
| "last": "Zoph", |
| "suffix": "" |
| }, |
| { |
| "first": "Deniz", |
| "middle": [], |
| "last": "Yuret", |
| "suffix": "" |
| }, |
| { |
| "first": "Jonathan", |
| "middle": [], |
| "last": "May", |
| "suffix": "" |
| }, |
| { |
| "first": "Kevin", |
| "middle": [], |
| "last": "Knight", |
| "suffix": "" |
| } |
| ], |
| "year": 2016, |
| "venue": "EMNLP", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Barret Zoph, Deniz Yuret, Jonathan May, and Kevin Knight. 2016. Transfer learning for low-resource neural machine translation. In EMNLP.", |
| "links": null |
| }, |
| "BIBREF38": { |
| "ref_id": "b38", |
| "title": "Asynchronous translations with recurrent neural nets", |
| "authors": [ |
| { |
| "first": "P", |
| "middle": [], |
| "last": "Ramon", |
| "suffix": "" |
| }, |
| { |
| "first": "Mikel", |
| "middle": [ |
| "L" |
| ], |
| "last": "\u00d1eco", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Forcada", |
| "suffix": "" |
| } |
| ], |
| "year": 1997, |
| "venue": "Neural Networks", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ramon P.\u00d1eco and Mikel L. Forcada. 1997. Asyn- chronous translations with recurrent neural nets. Neural Networks.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "FIGREF0": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "Two small improvements to the baseline model: (a) residual connection, and (b) multi-layer attention." |
| }, |
| "FIGREF1": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "t r a in o n p a r a ll e l c o r p u s 2.0 N o p r e t r a in in g 2.0 O n ly p r e t r a in e m b e d d in g s t r a in o n W ik ip e d iaFigure 3: English\u2192German ablation study measuring the difference in validation BLEU between various ablations and the full model. More negative is worse. The full model uses LMs trained with monolingual data to initialize the encoder and decoder, plus the language modeling objective." |
| }, |
| "FIGREF2": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "Validation performance of pretraining vs. no pretraining when trained on a subset of the entire labeled dataset for English\u2192German translation." |
| }, |
| "FIGREF3": { |
| "type_str": "figure", |
| "uris": null, |
| "num": null, |
| "text": "r e t r a in in g O n ly p r e t r a in d e c o d e r N o L M o b je c t iv e O n ly p r e t r a in e m b e d d in g s O n ly p r e t r a in e m b e d d in g s & L S T M O n ly p r e t r a in e n c o d e r P r e t r a in o n p a r a ll e l c o r p u s ROUGE1 ROUGE2ROUGEL Summarization ablation study measuring the difference in validation ROUGE between various ablations and the full model. More negative is worse. The full model uses LMs trained with unlabeled data to initialize the encoder and decoder, plus the language modeling objective." |
| }, |
| "TABREF2": { |
| "html": null, |
| "num": null, |
| "type_str": "table", |
| "content": "<table/>", |
| "text": "English\u2192German performance on WMT test sets. Our pretrained model outperforms all other models. Note that the model without pretraining uses the LM objective." |
| }, |
| "TABREF4": { |
| "html": null, |
| "num": null, |
| "type_str": "table", |
| "content": "<table/>", |
| "text": "Results on the anonymized CNN/Daily Mail dataset." |
| } |
| } |
| } |
| } |
