Pith. sign in

REVIEW 4 major objections 5 minor 2 cited by

Deep Learning Based Chatbot Models

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Open-domain chatbots are limited by training that averages over many valid replies, and speaker or persona priors can disambiguate response generation.

desk verdict A candid, self-aware student report whose survey and persona-conditioning argument are worth reading, but whose headline experimental claim rests on an uncontrolled baseline comparison. read the letter →

arxiv 1908.08835 v1 pith:4PNSMUVP submitted 2019-08-23 cs.CL

Richard Csaky This is my paper · ORCID
classification cs.CL
keywords open-domainchatbotssequence-to-sequencemodelsTransformerpersonaconditioningspeaker-addresseeembeddingsmaximumlikelihoodtrainingdialogueevaluationresponsediversity
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper surveys recent open-domain chatbot work and argues that the conversation domain is too ambiguous for the standard maximum-likelihood sequence-to-sequence recipe borrowed from machine translation. It supports this with experiments adapting the Transformer architecture to two dialogue corpora, finding that none of its trained models matched a published seq2seq baseline and that speaker-addressee annotations visibly change generated replies. The author's central proposal is that response generation should be conditioned on additional priors such as persona, mood, and conversation history, so that a source utterance no longer maps to one averaged reply.

What carries the argument

The load-bearing machinery is the encoder-decoder conditional language model with cross-entropy training, inherited from machine translation, together with the proposed fix of feeding prior embeddings (speaker, addressee, mood, persona) into the decoder. The paper's critique centers on the loss: maximizing $\log p(\text{reply}|\text{utterance})$ over a dataset with many valid replies makes the model average them, which the author identifies as the mechanism behind generic responses like "I don't know." The Transformer experiments are the test bed: a pure attention-based sequence-to-sequence model with no recurrence, trained on movie and subtitle dialogue, with and without speaker-addressee tokens prepended to the source.

What would settle it

Retrain the seq2seq baseline on the same Cornell and OpenSubtitles splits with matched preprocessing, vocabulary, and hyperparameter budget, then compare perplexity, BLEU, and human rankings; if the retrained baseline no longer beats the Transformer, the paper's main empirical claim fails. A second test: construct a dataset with several valid replies per utterance and check whether cross-entropy training produces a hidden-state average near generic answers; if not, the loss-averaging mechanism is not the cause.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that open-domain chatbots fail because they ignore the priors that make a reply determinate: for a given utterance there are many valid responses, and training with cross-entropy over utterance–reply pairs pushes the model to output an average that lands near generic, safe replies. The experiments are offered as preliminary evidence: a Transformer trained on Cornell Movie-Dialog and OpenSubtitles produced more diverse replies as it began to overfit, speaker and addressee tokens switched answers to the same question, and none of the trainings reached the quality of the published seq2seq baseline. The paper also argues that perplexity and BLEU reward the same averaging behavior and therefore do not measure conversational quality.

Load-bearing premise

The load-bearing premise is that the published seq2seq baseline is a fair comparison, even though it was not retrained on the same data splits, preprocessing, or vocabulary as the Transformer models.

Editorial extensions

If this is right

  • If conditioning on speaker and addressee embeddings is what disambiguates replies, persona-aware training should outperform ordinary seq2seq on human judgment even when perplexity and BLEU do not improve.
  • If perplexity and BLEU reward safe generic responses, the field's standard automatic metrics will keep mis-ranking conversational models, so better metrics must reward diversity and relevance to priors.
  • If overfitting improves response quality in the paper's experiments, validation-loss-based early stopping is the wrong criterion for open-domain dialogue.
  • If the Transformer needs domain-specific tuning, transferring NMT hyperparameters to dialogue is not a neutral starting point for comparing architectures.
  • Conditioning on mood or other priors would require datasets labeled with those priors, which currently do not exist for open-domain dialogue.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the speaker-addressee results generalize, any dialogue corpus with speaker IDs already contains a weak persona prior that can be learned without extra annotation.
  • If the loss-averaging account is correct, objectives that model a distribution over multiple valid replies rather than a single target, such as latent-variable or divergence-based training, should reduce generic outputs in the same experimental setup.
  • The temporal-delay and memory proposals from the future-work section could be turned into a concrete benchmark by recording response times from human-human chat and conditioning generation on predicted delay.
  • The argument that movie dialogue is not natural conversation also implies that results on Cornell and OpenSubtitles may understate what persona-conditioned models could do on real collected chat logs.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This manuscript, originally a 2017 scientific students' association report posted to arXiv in 2019, surveys recent work on neural chatbot models and then presents experiments adapting the Transformer architecture to open-domain dialogue. The survey portion reviews over 70 publications, grouping them into categories such as attention, pretraining, additional input features, knowledge bases, hierarchical models, task-oriented dialogue, and reinforcement learning. The paper argues that open-domain conversation is inherently one-to-many and that standard maximum-likelihood training therefore averages over multiple valid replies, producing generic responses such as "I don't know"; it proposes conditioning on priors such as speaker, addressee, and mood. Experimentally, the author trains Transformer models on Cornell Movie-Dialog and OpenSubtitles data, including speaker-addressee augmented variants, and reports perplexity and BLEU scores in Table 1 and qualitative response samples in Tables 2 through 7. The paper concludes that none of the Transformer trainings matched the performance of a baseline seq2seq model, while also acknowledging that the comparison is imperfect and that further experiments are needed.

Significance. If the paper's central claim were established, it would provide evidence that the same maximum-likelihood encoder-decoder framework that works for machine translation is insufficient for open-domain dialogue, and that architectural change or additional conditioning priors are needed. The survey portion is a useful structured overview of the 2014--2017 dialogue literature, and the author has released the training code, which supports reproducibility. The speaker-addressee analysis in Table 7 is a genuinely interesting demonstration that explicit speaker tokens can change generated responses in a measurable way. However, the experimental support for the headline claim is weak: the baseline seq2seq numbers are taken from a different paper with different preprocessing and data splits, the checkpoints used for qualitative examples are not the checkpoints used for quantitative scores, and no human evaluation or statistical uncertainty is reported. Those issues are partly acknowledged in the manuscript itself, which is to the author's credit, but they still mean the central comparison is not currently established.

major comments (4)
  1. [Section 5, Table 1 and Section 5.2] The central empirical claim that "none of the trainings achieved the performance of the baseline seq2seq model" is not supported by the evidence as presented. The baseline perplexity of 17 and the baseline sample responses are taken from Vinyals and Le (2015), but the Transformer models are trained with different preprocessing, different vocabulary sizes (32K, 40K, 100K), different data subsets, and different validation splits. The OpenSubtitles comparison is described as the most direct, but even there the data subset is merely "similar in size" to the baseline, not identical. A rigorous comparison would require retraining the seq2seq baseline on the same source-target pairs, tokenization, vocabulary, and evaluation protocol, and then comparing models at comparably selected checkpoints. As written, the comparison is confounded by dataset construction, model capacity, and training duration, so the conclusion in Section 5.2 cannot bear the weight of the paper's abstract claim that the general conversation domain demands different architectures.
  2. [Sections 5.1 and 5.2] There is a checkpoint-selection inconsistency between the quantitative and qualitative results. Section 5.1 states that the Cornell and Cornell S perplexity scores were computed at 20K and 16K training steps respectively, before overfitting; Section 5.2 then reports that the example responses for these models were generated from checkpoints at 219K and 238K steps, after deliberate overfitting. The OpenSubtitles F examples come from yet another checkpoint at 370K steps of finetuning. Tables 2 through 6 therefore do not correspond to the models whose scores appear in Table 1, and the qualitative conclusion that overfitting improves response quality is based on a post hoc selection of checkpoints that were not held out by any principled criterion. The manuscript should either report quantitative scores at the same checkpoints used for qualitative analysis or clearly separate the two analyses and justify the checkpoint choice with a pre-specified rule.
  3. [Section 3.4.2] The loss-function criticism is load-bearing for the paper's overall argument, but it is presented as an assertion rather than a derived or tested mechanism. The claim that cross-entropy training "makes the model learn an average of these responses" and that this average lies near generic safe answers is plausible but is not supported by any direct evidence in the paper. The overfitting observations in Section 5.2 are offered as circumstantial support, but they are confounded by the checkpoint-selection issue noted above and by the uncontrolled baseline comparison. To make this argument convincing, the author would need a direct test, for example comparing the embedding-space average of multiple valid references to the model's generated output, or constructing a dataset where ambiguity is controlled and showing that the model's outputs move toward the centroid of the reference set.
  4. [Section 5, Tables 1--7] The paper relies on perplexity, BLEU, and selected qualitative examples to compare models, yet Section 3.4.4 of the same manuscript correctly notes that these automatic metrics do not correlate with human judgment for dialogue. The qualitative examples in Tables 2 through 7 are selected after inspecting the outputs, and Section 5.2 explicitly says the source utterances were originally chosen for evaluating the seq2seq baseline, not the Transformer models. There is no human evaluation, no inter-annotator agreement, no multiple random seeds, and no error bars. Given that the conclusion is about the quality of conversational responses, the author should run a human evaluation on a random set of source utterances, with multiple evaluators, and report agreement and significance; at minimum, the current claims should be explicitly downgraded to anecdotal.
minor comments (5)
  1. [Abstract and title page] There are several typos and stylistic issues, including "Tranformer" in Section 4.1 and "state-of-of-the-art" in the abstract; these should be corrected.
  2. [Figure/table captions, Table 7] The table caption for Table 7 does not define the labels "Cornell S matching R" and "Cornell S different R"; the R suffix is not explained anywhere and should be defined, and the speaker/addressee direction should be stated explicitly in the caption.
  3. [Section 5.2] The phrase "the best responses from the two Cornell trainings were not those outputted before the models started to overfit" is a post hoc observation; if kept, it needs to be accompanied by a clear statement that these examples were selected to illustrate a phenomenon, not to measure average performance.
  4. [Section 4.2.2] The description of the OpenSubtitles subset as "62M sentences" is ambiguous because the unit of counting could be lines or sentence tokens; please clarify the exact construction and size of the training and validation splits used.
  5. [Section 6] The paper dates from 2017, and the discussion of the Transformer as "currently state-of-the-art in NMT" is anachronistic for the 2019 arXiv version; the author should update or frame such statements with a clear date or remove the implication that this is current as of publication.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's claims rest on external literature and independent experiments, and its stated hypotheses are not disguised definitions or fitted predictions.

full rationale

The paper's central claims are a literature survey plus preliminary Transformer experiments, not a derivation whose outputs are built into its inputs. The loss-function averaging argument in Section 3.4.2 is explicitly labeled a presumption ('My presumption is that it is precisely because of the loss function that the models learn to output safe and generic responses') and is offered as a causal hypothesis, not as a theorem restating its own conclusion. The experimental comparison in Section 5 imports the seq2seq baseline perplexity and sample responses from Vinyals and Le (2015); this is an external benchmark, not a parameter fitted inside the paper, so the conclusion that the Transformer trainings did not match the baseline is not circular, even though the comparison is uncontrolled and therefore a correctness risk. The qualitative checkpoint selection in Section 5.2 is selective reporting rather than a fitted-input-called-prediction step: the paper does not define a metric in terms of the responses it then claims as evidence. The speaker-addressee results in Table 7 follow from feeding speaker tokens as inputs and observing changed outputs, which is an empirical effect, not a self-definitional identity. The only self-reference found is the author's GitHub code link in Section 4.3.1, which is not load-bearing to any argument. No equation in the paper is defined in terms of the quantity it is supposed to predict, and no load-bearing claim reduces to a self-citation chain. Accordingly, the correct circularity finding is no significant circularity.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new physical or formal entities and derives no equations from fitted parameters. The load-bearing choices are the un-tuned Transformer hyperparameters, which confound the architecture comparison, and several domain assumptions about loss function behavior and evaluation metrics. These are stated in the paper, not hidden, but they are assumptions nonetheless.

free parameters (1)
  • Transformer base hyperparameters from NMT = base variant, batch 4096/2048 tokens, steps 350k/238k/1M/675k
    Adopted unchanged from Vaswani et al. (2017); author states in Section 6 that hyperparameters might explain the poor OpenSubtitles results, so the conclusion that Transformer underperforms is partially dependent on this un-tuned choice.
assumptions (3)
  • domain assumption Cross-entropy training pushes the model to output an average of all plausible replies, which lies near generic responses.
    Stated in Section 3.4.2 as the author's presumption; no formal proof or direct experiment is provided, and it underlies the paper's main criticism.
  • domain assumption Perplexity and BLEU do not correlate with human judgment for dialogue.
    The paper relies on Liu et al. (2016) and Tao et al. (2017) for this; it is an input assumption used to justify the qualitative analysis, not derived in the paper.
  • domain assumption Movie characters have learnable, consistent personas that speaker embeddings can capture.
    This is assumed in the speaker-addressee training setup (Section 4.3.3) and in the interpretation of Table 7; the paper itself acknowledges Cornell Movie-Dialog may be too small for this to generalize (Section 6.1).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deep Learning Based Chatbot Models." pith.science (2026). https://pith.science/paper/4PNSMUVP

@misc{pith2026190808835,
  author       = {Pith},
  title        = {Pith review of: Deep Learning Based Chatbot Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4PNSMUVP}},
  note         = {Machine review of arXiv:1908.08835}
}
read the original abstract

A conversational agent (chatbot) is a piece of software that is able to communicate with humans using natural language. Modeling conversation is an important task in natural language processing and artificial intelligence. While chatbots can be used for various tasks, in general they have to understand users' utterances and provide responses that are relevant to the problem at hand. In my work, I conduct an in-depth survey of recent literature, examining over 70 publications related to chatbots published in the last 3 years. Then, I proceed to make the argument that the very nature of the general conversation domain demands approaches that are different from current state-of-of-the-art architectures. Based on several examples from the literature I show why current chatbot models fail to take into account enough priors when generating responses and how this affects the quality of the conversation. In the case of chatbots, these priors can be outside sources of information that the conversation is conditioned on like the persona or mood of the conversers. In addition to presenting the reasons behind this problem, I propose several ideas on how it could be remedied. The next section focuses on adapting the very recent Transformer model to the chatbot domain, which is currently state-of-the-art in neural machine translation. I first present experiments with the vanilla model, using conversations extracted from the Cornell Movie-Dialog Corpus. Secondly, I augment the model with some of my ideas regarding the issues of encoder-decoder architectures. More specifically, I feed additional features into the model like mood or persona together with the raw conversation data. Finally, I conduct a detailed analysis of how the vanilla model performs on conversational data by comparing it to previous chatbot models and how the additional features affect the quality of the generated responses.

Figures

Figures reproduced from arXiv: 1908.08835 by the authors.

Figure 1
Figure 1. A sample conversation with Cleverbot [Carpenter, 2017] [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Unfolding of an RNN over 3 time-steps. Here [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. A general seq2seq model, where (A, B, C) is the input sequence, < EOS > is a symbol used to delimit the end of the sentence and (W, X, Y, Z) is the output sequence [Sutskever et al., 2014] The simplest and initial form of the model is based on two RNNs, visualized in [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (19 more)
Figure 4
Figure 4. Figure 4: A 3 layer seq2seq model [Tensorflow, 2017]. The lines pointing from the last decoder [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: A bidirectional multilayer LSTM encoder and unidirectional LSTM decoder [Yin et al., [PITH_FULL_IMAGE:figures/full_fig_p015_5.png]
Figure 6
Figure 6. Figure 6: Original Attention Mechanism [Bahdanau et al., 2014] [PITH_FULL_IMAGE:figures/full_fig_p019_6.png]
Figure 7
Figure 7. Figure 7: Pixels represent the weights aij between inputs and outputs [Bahdanau et al., 2014] Attention can also be computed using only one sequence of symbols, called self- or intra￾attention. Self-attention has been successfully applied to a variety of tasks including reading …
Figure 8
Figure 8. Figure 8: Visualizing self-attention, lines represent the weights [PITH_FULL_IMAGE:figures/full_fig_p020_8.png]
Figure 9
Figure 9. Figure 9: A seq2seq model augmented with speaker embeddings [Li et al., 2016a]. [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]
Figure 10
Figure 10. Figure 10: A HRED with attention, described in detail in Section 3.3.1, upgraded with a KB [PITH_FULL_IMAGE:figures/full_fig_p024_10.png]
Figure 11
Figure 11. Figure 11: The Hierarchical Recurrent Encoder-Decoder Model [Serban et al., 2016]. [PITH_FULL_IMAGE:figures/full_fig_p026_11.png]
Figure 12
Figure 12. Figure 12: The Hierarchical Recurrent Attention Network [Xing et al., 2017b] [PITH_FULL_IMAGE:figures/full_fig_p028_12.png]
Figure 13
Figure 13. Figure 13: The 5 bAbi tasks for restaurant reservations. The dialog agent’s utterances are in blue, [PITH_FULL_IMAGE:figures/full_fig_p029_13.png]
Figure 14
Figure 14. Figure 14: The reinforcement learning framework [Sutton and Barto, 1998]. [PITH_FULL_IMAGE:figures/full_fig_p031_14.png]
Figure 15
Figure 15. Figure 15: A high-level visualization of the MultiModel [Kaiser et al., 2017b] [PITH_FULL_IMAGE:figures/full_fig_p033_15.png]
Figure 16
Figure 16. Figure 16: Bleu score correlation with human scores [Liu et al., 2016] on a Twitter corpus (left) [PITH_FULL_IMAGE:figures/full_fig_p036_16.png]
Figure 17
Figure 17. Figure 17: Correlation of ADEM (left) [Lowe et al., 2017] and RUBER (right) [Tao et al., 2017] [PITH_FULL_IMAGE:figures/full_fig_p037_17.png]
Figure 18
Figure 18. Figure 18: The architecture of the Transformer model [Vaswani et al., 2017]. The encoder network [PITH_FULL_IMAGE:figures/full_fig_p038_18.png]
Figure 19
Figure 19. Figure 19: A diagram of the scaled dot-product attention can be seen on the left, and of the multi [PITH_FULL_IMAGE:figures/full_fig_p040_19.png]
Figure 20
Figure 20. Figure 20: The diagram shows the validation loss over the entire training of the Cornell model. [PITH_FULL_IMAGE:figures/full_fig_p046_20.png]
Figure 21
Figure 21. Figure 21: The diagram shows the validation loss over the entire training of the Cornell S model. [PITH_FULL_IMAGE:figures/full_fig_p046_21.png]
Figure 22
Figure 22. Figure 22: The diagram shows the validation loss over the entire training of the OpenSubtitles and [PITH_FULL_IMAGE:figures/full_fig_p046_22.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Mock Deep Testing: Toward Separate Development of Data and Models for Deep Learning

    cs.SE 2025-02 conditional novelty 6.0 of 10

    KUnit's mock-based unit testing identified 63 issues in 50 DL programs and supported developers in resolving 63 issues in a user study with 36 participants.

  2. Leveraging Data Characteristics for Bug Localization in Deep Learning Programs

    cs.SE 2024-12 reject novelty 6.0 of 10

    Theia localizes structural bugs in Keras and PyTorch deep learning programs by combining dataset characteristics with static analysis of the model code, and reports 57/75 bugs on a 40-program Stack Overflow benchmark.

Reference graph

Works this paper leans on

127 extracted references · 22 canonical work pages · cited by 2 Pith papers

  1. [1]

    Chat Detection in an Intelligent Assistant: Combining Task-oriented and Non-task-oriented Spoken Dialogue Systems

    Akasaki, S. and Kaji, N. (2017). Chat detection in an intelligent assistant: Combining task-oriented and non-task-oriented spoken dialogue systems. arXiv preprint arXiv:1705.00746

    work page Pith review arXiv 2017

  2. [2]

    Apple (2017). Siri. https://www.apple.com/ios/siri/. Accessed: 2017-10-04

    2017

  3. [3]

    L., Kiros, J

    Ba, J. L., Kiros, J. R., and Hinton, G. E. (2016). Layer normalization. arXiv preprint arXiv:1607.06450

    arXiv 2016

  4. [4]

    Bahdanau, D., Cho, K., and Bengio, Y. (2014). Neural machine translation by jointly learning to align and translate. arXiv preprint arXiv:1409.0473

    arXiv 2014

  5. [5]

    Barone, A. V. M. and Sennrich, R. (2017). A parallel corpus of python functions and documentation strings for automated code documentation and code generation. arXiv preprint arXiv:1707.02275

    arXiv 2017

  6. [6]

    Bengio, Y., Ducharme, R., Vincent, P., and Jauvin, C. (2003). A neural probabilistic language model. Journal of machine learning research , 3(Feb):1137--1155

    2003

  7. [7]

    Bordes, A., Boureau, Y.-L., and Weston, J. (2016). Learning end-to-end goal-oriented dialog. arXiv preprint arXiv:1605.07683

    arXiv 2016

  8. [8]

    Bottou, L. (2010). Large-scale machine learning with stochastic gradient descent. In Proceedings of COMPSTAT'2010 , pages 177--186. Springer

    2010

Show all 127 references
  1. [9]

    Britz, D. (2015). Recurrent neural network tutorial. http://www.wildml.com/2015/09/recurrent-neural-networks-tutorial-part-1-introduction-to-rnns/. Accessed: 2017-10-09

    2015

  2. [10]

    Carpenter, R. (2017). Cleverbot. http://www.cleverbot.com/. Accessed: 2017-10-04

    2017

  3. [11]

    and Manning, C

    Chen, D. and Manning, C. (2014). A fast and accurate dependency parser using neural networks. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP) , pages 740--750

    2014

  4. [12]

    Cheng, J., Dong, L., and Lapata, M. (2016). Long short-term memory-networks for machine reading. arXiv preprint arXiv:1601.06733

    2016 arXiv

  5. [13]

    Chiu, C.-C., Lawson, D., Luo, Y., Tucker, G., Swersky, K., Sutskever, I., and Jaitly, N. (2017). An online sequence-to-sequence model for noisy speech recognition. arXiv preprint arXiv:1706.06428

    2017 arXiv

  6. [14]

    Cho, K., Van Merri \"e nboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., and Bengio, Y. (2014). Learning phrase representations using rnn encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078

    2014 arXiv

  7. [15]

    Choudhary, S., Srivastava, P., Ungar, L., and Sedoc, J. (2017). Domain aware neural dialog system. arXiv preprint arXiv:1708.00897

    2017 arXiv

  8. [16]

    and Lee, L

    Danescu-Niculescu-Mizil, C. and Lee, L. (2011). Chameleons in imagined conversations: A new approach to understanding coordination of linguistic style in dialogs. In Proceedings of the 2nd Workshop on Cognitive Modeling and Computational Linguistics , pages 76--87. Association...

    2011

  9. [17]

    M., Lee, S., and Batra, D

    Das, A., Kottur, S., Moura, J. M., Lee, S., and Batra, D. (2017). Learning cooperative visual dialog agents with deep reinforcement learning. arXiv preprint arXiv:1703.06585

    2017 arXiv

  10. [18]

    Dodge, J., Gane, A., Zhang, X., Bordes, A., Chopra, S., Miller, A., Szlam, A., and Weston, J. (2015). Evaluating prerequisite qualities for learning end-to-end dialog systems. arXiv preprint arXiv:1511.06931

    2015 arXiv

  11. [19]

    and Manning, C

    Eric, M. and Manning, C. D. (2017). A copy-augmented sequence-to-sequence architecture gives good performance on task-oriented dialogue. arXiv preprint arXiv:1701.04024

    2017 arXiv

  12. [20]

    The babi project

    Facebook (2017). The babi project. https://research.fb.com/downloads/babi/. Accessed: 2017-10-13

    2017

  13. [21]

    Feng, Y., Zhang, S., Zhang, A., Wang, D., and Abel, A. (2017). Memory-augmented neural machine translation. arXiv preprint arXiv:1708.02005

    2017 arXiv

  14. [22]

    Gehring, J., Auli, M., Grangier, D., Yarats, D., and Dauphin, Y. N. (2017). Convolutional sequence to sequence learning. arXiv preprint arXiv:1705.03122

    2017 arXiv

  15. [23]

    Ghazvininejad, M., Brockett, C., Chang, M.-W., Dolan, B., Gao, J., Yih, W.-t., and Galley, M. (2017). A knowledge-grounded neural conversation model. arXiv preprint arXiv:1702.01932

    2017 arXiv

  16. [24]

    Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., and Bengio, Y. (2014). Generative adversarial nets. In Advances in neural information processing systems , pages 2672--2680

    2014

  17. [25]

    Google assistant

    Google (2017). Google assistant. https://assistant.google.com/. Accessed: 2017-10-04

    2017

  18. [26]

    Goyal, K., Neubig, G., Dyer, C., and Berg-Kirkpatrick, T. (2017). A continuous relaxation of beam search for end-to-end training of neural sequence models. arXiv preprint arXiv:1708.00111

    2017 arXiv

  19. [27]

    Gu, J., Lu, Z., Li, H., and Li, V. O. (2016). Incorporating copying mechanism in sequence-to-sequence learning. arXiv preprint arXiv:1603.06393

    2016 arXiv

  20. [28]

    Harris, Z. S. (1954). Distributional structure. Word , 10(2-3):146--162

    1954

  21. [29]

    and Titov, I

    Havrylov, S. and Titov, I. (2017). Emergence of language with multi-agent games: Learning to communicate with sequences of symbols. arXiv preprint arXiv:1705.11192

    2017 arXiv

  22. [30]

    He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 770--778

    2016

  23. [31]

    Henderson, M. (2015). Machine learning for dialog state tracking: A review. In Machine Learning in Spoken Language Processing Workshop

    2015

  24. [32]

    Hochreiter, S. (1998). The vanishing gradient problem during learning recurrent neural nets and problem solutions. International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems , 6(02):107--116

    1998

  25. [33]

    and Schmidhuber, J

    Hochreiter, S. and Schmidhuber, J. (1997). Long short-term memory. Neural computation , 9(8):1735--1780

    1997

  26. [34]

    Jean, S., Cho, K., Memisevic, R., and Bengio, Y. (2014). On using very large target vocabulary for neural machine translation. arXiv preprint arXiv:1412.2007

    2014 arXiv

  27. [35]

    Jena, G., Vashisht, M., Basu, A., Ungar, L., and Sedoc, J. (2017). Enterprise to computer: Star trek chatbot. arXiv preprint arXiv:1708.00818

    2017 arXiv

  28. [36]

    K., Mi, F., and Faltings, B

    Joshi, C. K., Mi, F., and Faltings, B. (2017). Personalization in goal-oriented dialog. arXiv preprint arXiv:1706.07503

    2017 arXiv

  29. [37]

    and Bengio, S

    Kaiser, . and Bengio, S. (2016). Can active memory replace attention? In Advances in Neural Information Processing Systems , pages 3781--3789

    2016

  30. [38]

    N., and Chollet, F

    Kaiser, L., Gomez, A. N., and Chollet, F. (2017a). Depthwise separable convolutions for neural machine translation. arXiv preprint arXiv:1706.03059

    2017 arXiv

  31. [39]

    N., Shazeer, N., Vaswani, A., Parmar, N., Jones, L., and Uszkoreit, J

    Kaiser, L., Gomez, A. N., Shazeer, N., Vaswani, A., Parmar, N., Jones, L., and Uszkoreit, J. (2017b). One model to learn them all. arXiv preprint arXiv:1706.05137

    2017 arXiv

  32. [40]

    Kalchbrenner, N., Espeholt, L., Simonyan, K., Oord, A. v. d., Graves, A., and Kavukcuoglu, K. (2016). Neural machine translation in linear time. arXiv preprint arXiv:1610.10099

    2016 arXiv

  33. [41]

    Kandasamy, K., Bachrach, Y., Tomioka, R., Tarlow, D., and Carter, D. (2017). Batch policy gradient methods for improving neural conversation models. arXiv preprint arXiv:1702.03334

    2017 arXiv

  34. [42]

    Konstas, I., Iyer, S., Yatskar, M., Choi, Y., and Zettlemoyer, L. (2017). Neural amr: Sequence-to-sequence models for parsing and generation. arXiv preprint arXiv:1704.08381

    2017 arXiv

  35. [43]

    M., Lee, S., and Batra, D

    Kottur, S., Moura, J. M., Lee, S., and Batra, D. (2017). Natural language does not emerge'naturally'in multi-agent dialog. arXiv preprint arXiv:1706.08502

    2017 arXiv

  36. [44]

    Krizhevsky, A., Sutskever, I., and Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems , pages 1097--1105

    2012

  37. [45]

    Lample, G., Ballesteros, M., Subramanian, S., Kawakami, K., and Dyer, C. (2016). Neural architectures for named entity recognition. arXiv preprint arXiv:1603.01360

    2016 arXiv

  38. [46]

    Li, J., Galley, M., Brockett, C., Gao, J., and Dolan, B. (2015). A diversity-promoting objective function for neural conversation models. arXiv preprint arXiv:1510.03055

    2015 arXiv

  39. [47]

    P., Gao, J., and Dolan, B

    Li, J., Galley, M., Brockett, C., Spithourakis, G. P., Gao, J., and Dolan, B. (2016a). A persona-based neural conversation model. arXiv preprint arXiv:1603.06155

    2016 arXiv

  40. [48]

    H., Chopra, S., Ranzato, M., and Weston, J

    Li, J., Miller, A. H., Chopra, S., Ranzato, M., and Weston, J. (2016b). Dialogue learning with human-in-the-loop. arXiv preprint arXiv:1611.09823

    2016 arXiv

  41. [49]

    Li, J., Monroe, W., Ritter, A., Galley, M., Gao, J., and Jurafsky, D. (2016c). Deep reinforcement learning for dialogue generation. arXiv preprint arXiv:1606.01541

    2016 arXiv

  42. [50]

    Li, J., Monroe, W., Shi, T., Ritter, A., and Jurafsky, D. (2017). Adversarial learning for neural dialogue generation. arXiv preprint arXiv:1701.06547

    2017 arXiv

  43. [51]

    Li, X., Mou, L., Yan, R., and Zhang, M. (2016d). Stalematebreaker: A proactive content-introducing approach to automatic human-computer conversation. arXiv preprint arXiv:1604.04358

    2016 arXiv

  44. [52]

    Lin, Z., Feng, M., Santos, C. N. d., Yu, M., Xiang, B., Zhou, B., and Bengio, Y. (2017). A structured self-attentive sentence embedding. arXiv preprint arXiv:1703.03130

    2017 arXiv

  45. [53]

    and Bibauw, S

    Lison, P. and Bibauw, S. (2017). Not all dialogues are created equal: Instance weighting for neural conversational models. arXiv preprint arXiv:1704.08966

    2017 arXiv

  46. [54]

    and Tiedemann, J

    Lison, P. and Tiedemann, J. (2016). Opensubtitles2016: Extracting large parallel corpora from movie and tv subtitles. In LREC

    2016

  47. [55]

    V., Noseworthy, M., Charlin, L., and Pineau, J

    Liu, C.-W., Lowe, R., Serban, I. V., Noseworthy, M., Charlin, L., and Pineau, J. (2016). How not to evaluate your dialogue system: An empirical study of unsupervised evaluation metrics for dialogue response generation. arXiv preprint arXiv:1603.08023

    2016 arXiv

  48. [56]

    V., Angelard-Gontier, N., Bengio, Y., and Pineau, J

    Lowe, R., Noseworthy, M., Serban, I. V., Angelard-Gontier, N., Bengio, Y., and Pineau, J. (2017). Towards an automatic turing test: Learning to evaluate dialogue responses. arXiv preprint arXiv:1708.07149

    2017 arXiv

  49. [57]

    Lowe, R., Pow, N., Serban, I., and Pineau, J. (2015). The ubuntu dialogue corpus: A large dataset for research in unstructured multi-turn dialogue systems. arXiv preprint arXiv:1506.08909

    2015 arXiv

  50. [58]

    Luong, M.-T., Pham, H., and Manning, C. D. (2015). Effective approaches to attention-based neural machine translation. arXiv preprint arXiv:1508.04025

    2015 arXiv

  51. [59]

    V., Vinyals, O., and Zaremba, W

    Luong, M.-T., Sutskever, I., Le, Q. V., Vinyals, O., and Zaremba, W. (2014). Addressing the rare word problem in neural machine translation. arXiv preprint arXiv:1410.8206

    2014 arXiv

  52. [60]

    D., Sch \"u tze, H., et al

    Manning, C. D., Sch \"u tze, H., et al. (1999). Foundations of statistical natural language processing , volume 999. MIT Press

    1999

  53. [61]

    Marietto, M. d. G. B., de Aguiar, R. V., Barbosa, G. d. O., Botelho, W. T., Pimentel, E., Fran c a, R. d. S., and da Silva, V. L. (2013). Artificial intelligence markup language: A brief tutorial. arXiv preprint arXiv:1307.3091

    2013 arXiv

  54. [62]

    Microsoft (2017a). Cortana. https://www.microsoft.com/en-us/windows/cortana. Accessed: 2017-10-04

    2017

  55. [63]

    Microsoft bot framework

    Microsoft (2017b). Microsoft bot framework. https://dev.botframework.com/. Accessed: 2017-10-04

    2017

  56. [64]

    Mikolov, T., Chen, K., Corrado, G., and Dean, J. (2013a). Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781

    2013 arXiv

  57. [65]

    S., and Dean, J

    Mikolov, T., Sutskever, I., Chen, K., Corrado, G. S., and Dean, J. (2013b). Distributed representations of words and phrases and their compositionality. In Advances in neural information processing systems , pages 3111--3119

    2013

  58. [66]

    Miller, A., Fisch, A., Dodge, J., Karimi, A.-H., Bordes, A., and Weston, J. (2016). Key-value memory networks for directly reading documents. arXiv preprint arXiv:1606.03126

    2016 arXiv

  59. [67]

    H., Feng, W., Fisch, A., Lu, J., Batra, D., Bordes, A., Parikh, D., and Weston, J

    Miller, A. H., Feng, W., Fisch, A., Lu, J., Batra, D., Bordes, A., Parikh, D., and Weston, J. (2017). Parlai: A dialog research software platform. arXiv preprint arXiv:1705.06476

    2017 arXiv

  60. [68]

    Mitchell, M. (1998). An introduction to genetic algorithms . MIT press

    1998

  61. [69]

    Mnih, V., Kavukcuoglu, K., Silver, D., Graves, A., Antonoglou, I., Wierstra, D., and Riedmiller, M. (2013). Playing atari with deep reinforcement learning. arXiv preprint arXiv:1312.5602

    2013 arXiv

  62. [70]

    A., Veness, J., Bellemare, M

    Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., Graves, A., Riedmiller, M., Fidjeland, A. K., Ostrovski, G., et al. (2015). Human-level control through deep reinforcement learning. Nature , 518(7540):529--533

    2015

  63. [71]

    Nallapati, R., Zhou, B., Gulcehre, C., Xiang, B., et al. (2016). Abstractive text summarization using sequence-to-sequence rnns and beyond. arXiv preprint arXiv:1602.06023

    2016 arXiv

  64. [72]

    Olah, C. (2015). Understanding lstm networks. http://colah.github.io/posts/2015-08-Understanding-LSTMs/. Accessed: 2017-10-08

    2015

  65. [73]

    Opensubtitles

    opensubtitles.org (2017). Opensubtitles. https://www.opensubtitles.org/. Accessed: 2017-10-08

    2017

  66. [74]

    Papineni, K., Roukos, S., Ward, T., and Zhu, W.-J. (2002). Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting on association for computational linguistics , pages 311--318. Association for Computational Linguistics

    2002

  67. [75]

    a ckstr \

    Parikh, A. P., T \"a ckstr \"o m, O., Das, D., and Uszkoreit, J. (2016). A decomposable attention model for natural language inference. arXiv preprint arXiv:1606.01933

    2016 arXiv

  68. [76]

    Park, Y., Patwardhan, S., Visweswariah, K., and Gates, S. C. (2008). An empirical analysis of word error rate and keyword error rate. In INTERSPEECH , pages 2070--2073

    2008

  69. [77]

    Paulus, R., Xiong, C., and Socher, R. (2017). A deep reinforced model for abstractive summarization. arXiv preprint arXiv:1705.04304

    2017 arXiv

  70. [78]

    J., and Le, Q

    Ramachandran, P., Liu, P. J., and Le, Q. V. (2016). Unsupervised pretraining for sequence to sequence learning. arXiv preprint arXiv:1611.02683

    2016 arXiv

  71. [79]

    E., Hinton, G

    Rumelhart, D. E., Hinton, G. E., Williams, R. J., et al. (1988). Learning representations by back-propagating errors. Cognitive modeling , 5(3):1

    1988

  72. [80]

    and Buckley, C

    Salton, G. and Buckley, C. (1988). Term-weighting approaches in automatic text retrieval. Information processing & management , 24(5):513--523

    1988

  73. [81]

    and Paliwal, K

    Schuster, M. and Paliwal, K. K. (1997). Bidirectional recurrent neural networks. IEEE Transactions on Signal Processing , 45(11):2673--2681

    1997

  74. [82]

    Sennrich, R., Haddow, B., and Birch, A. (2015). Neural machine translation of rare words with subword units. arXiv preprint arXiv:1508.07909

    2015 arXiv

  75. [83]

    V., Klinger, T., Tesauro, G., Talamadupula, K., Zhou, B., Bengio, Y., and Courville, A

    Serban, I. V., Klinger, T., Tesauro, G., Talamadupula, K., Zhou, B., Bengio, Y., and Courville, A. C. (2017a). Multiresolution recurrent neural networks: An application to dialogue response generation. In AAAI , pages 3288--3294

    2017

  76. [84]

    V., Lowe, R., Charlin, L., and Pineau, J

    Serban, I. V., Lowe, R., Charlin, L., and Pineau, J. (2015). A survey of available corpora for building data-driven dialogue systems. arXiv preprint arXiv:1512.05742

    2015 arXiv

  77. [85]

    V., Sankar, C., Germain, M., Zhang, S., Lin, Z., Subramanian, S., Kim, T., Pieper, M., Chandar, S., Ke, N

    Serban, I. V., Sankar, C., Germain, M., Zhang, S., Lin, Z., Subramanian, S., Kim, T., Pieper, M., Chandar, S., Ke, N. R., et al. (2017b). A deep reinforcement learning chatbot. arXiv preprint arXiv:1709.02349

    2017 arXiv

  78. [86]

    V., Sordoni, A., Bengio, Y., Courville, A

    Serban, I. V., Sordoni, A., Bengio, Y., Courville, A. C., and Pineau, J. (2016). Building end-to-end dialogue systems using generative hierarchical neural network models. In AAAI , pages 3776--3784

    2016

  79. [87]

    V., Sordoni, A., Lowe, R., Charlin, L., Pineau, J., Courville, A

    Serban, I. V., Sordoni, A., Lowe, R., Charlin, L., Pineau, J., Courville, A. C., and Bengio, Y. (2017c). A hierarchical latent variable encoder-decoder model for generating dialogues. In AAAI , pages 3295--3301

    2017

  80. [88]

    Shang, L., Lu, Z., and Li, H. (2015). Neural responding machine for short-text conversation. arXiv preprint arXiv:1503.02364

    2015 arXiv

  81. [89]

    Shannon, M. (2017). Optimizing expected word error rate via sampling for speech recognition. arXiv preprint arXiv:1706.02776

    2017 arXiv

  82. [90]

    Shao, Y., Gouws, S., Britz, D., Goldie, A., Strope, B., and Kurzweil, R. (2017). Generating high-quality and informative conversation responses with sequence-to-sequence models. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing , pages 2...

    2017

  83. [91]

    Shazeer, N., Mirhoseini, A., Maziarz, K., Davis, A., Le, Q., Hinton, G., and Dean, J. (2017). Outrageously large neural networks: The sparsely-gated mixture-of-experts layer. arXiv preprint arXiv:1701.06538

    2017 arXiv

  84. [92]

    Shen, X., Su, H., Li, Y., Li, W., Niu, S., Zhao, Y., Aizawa, A., and Long, G. (2017). A conditional variational framework for dialog generation. arXiv preprint arXiv:1705.00316

    2017 arXiv

  85. [93]

    J., Guez, A., Sifre, L., Van Den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., et al

    Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., Van Den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., et al. (2016). Mastering the game of go with deep neural networks and tree search. Nature , 529(7587):484--489

    2016

  86. [94]

    Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., Hubert, T., Baker, L., Lai, M., Bolton, A., Chen, Y., Lillicrap, T., Hui, F., Sifre, L., van den Driessche, G., Graepel, T., and Hassabis, D. (2017). Mastering the game of go without human knowle...

    2017

  87. [95]

    Song, Y., Yan, R., Li, X., Zhao, D., and Zhang, M. (2016). Two are better than one: An ensemble of retrieval-and generation-based dialog systems. arXiv preprint arXiv:1610.07149

    2016 arXiv

  88. [96]

    Sordoni, A., Galley, M., Auli, M., Brockett, C., Ji, Y., Mitchell, M., Nie, J.-Y., Gao, J., and Dolan, B. (2015). A neural network approach to context-sensitive generation of conversational responses. arXiv preprint arXiv:1506.06714

    2015 arXiv

  89. [97]

    Sriram, A., Jun, H., Satheesh, S., and Coates, A. (2017). Cold fusion: Training seq2seq models together with language models. arXiv preprint arXiv:1708.06426

    2017 arXiv

  90. [98]

    E., Krizhevsky, A., Sutskever, I., and Salakhutdinov, R

    Srivastava, N., Hinton, G. E., Krizhevsky, A., Sutskever, I., and Salakhutdinov, R. (2014). Dropout: a simple way to prevent neural networks from overfitting. Journal of machine learning research , 15(1):1929--1958

    2014

  91. [99]

    Sukhbaatar, S., Weston, J., Fergus, R., et al. (2015). End-to-end memory networks. In Advances in neural information processing systems , pages 2440--2448

    2015

  92. [100]

    Sutskever, I., Vinyals, O., and Le, Q. V. (2014). Sequence to sequence learning with neural networks. In Advances in neural information processing systems , pages 3104--3112

    2014

  93. [101]

    Sutton, R. S. and Barto, A. G. (1998). Reinforcement learning: An introduction , volume 1. MIT press Cambridge

    1998

  94. [102]

    Tao, C., Mou, L., Zhao, D., and Yan, R. (2017). Ruber: An unsupervised method for automatic evaluation of open-domain dialog systems. arXiv preprint arXiv:1701.03079

    2017 arXiv

  95. [103]

    Sequence-to-sequence models

    Tensorflow (2017). Sequence-to-sequence models. https://www.tensorflow.org/tutorials/seq2seq. Accessed: 2017-10-08

    2017

  96. [104]

    Tiedemann, J. (2009). News from opus-a collection of multilingual parallel corpora with tools and interfaces. In Recent advances in natural language processing , volume 5, pages 237--248

    2009

  97. [105]

    N., Kaiser, L., and Polosukhin, I

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., and Polosukhin, I. (2017). Attention is all you need. arXiv preprint arXiv:1706.03762

    2017 arXiv

  98. [106]

    and Le, Q

    Vinyals, O. and Le, Q. (2015). A neural conversational model. arXiv preprint arXiv:1506.05869

    2015 arXiv

  99. [107]

    Wallace, R. S. (2009). The anatomy of alice. Parsing the Turing Test , pages 181--210

    2009

  100. [108]

    Weizenbaum, J. (1966). Eliza—a computer program for the study of natural language communication between man and machine. Communications of the ACM , 9(1):36--45

    1966

  101. [109]

    M., Su, P.-H., Ultes, S., and Young, S

    Wen, T.-H., Vandyke, D., Mrksic, N., Gasic, M., Rojas-Barahona, L. M., Su, P.-H., Ultes, S., and Young, S. (2016). A network-based end-to-end trainable task-oriented dialogue system. arXiv preprint arXiv:1604.04562

    2016 arXiv

  102. [110]

    Werbos, P. J. (1990). Backpropagation through time: what it does and how to do it. Proceedings of the IEEE , 78(10):1550--1560

    1990

  103. [111]

    M., van Merri \"e nboer, B., Joulin, A., and Mikolov, T

    Weston, J., Bordes, A., Chopra, S., Rush, A. M., van Merri \"e nboer, B., Joulin, A., and Mikolov, T. (2015). Towards ai-complete question answering: A set of prerequisite toy tasks. arXiv preprint arXiv:1502.05698

    2015 arXiv

  104. [112]

    D., Asadi, K., and Zweig, G

    Williams, J. D., Asadi, K., and Zweig, G. (2017). Hybrid code networks: practical and efficient end-to-end dialog control with supervised and reinforcement learning. arXiv preprint arXiv:1702.03274

    2017 arXiv

  105. [113]

    Williams, R. J. (1992). Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine learning , 8(3-4):229--256

    1992

  106. [114]

    and Rush, A

    Wiseman, S. and Rush, A. M. (2016). Sequence-to-sequence learning as beam-search optimization. arXiv preprint arXiv:1606.02960

    2016 arXiv

  107. [115]

    Worswick, S. (2017). Mitsuku. http://www.mitsuku.com/. Accessed: 2017-10-04

    2017

  108. [116]

    V., Norouzi, M., Macherey, W., Krikun, M., Cao, Y., Gao, Q., Macherey, K., et al

    Wu, Y., Schuster, M., Chen, Z., Le, Q. V., Norouzi, M., Macherey, W., Krikun, M., Cao, Y., Gao, Q., Macherey, K., et al. (2016). Google's neural machine translation system: Bridging the gap between human and machine translation. arXiv preprint arXiv:1609.08144

    2016 arXiv

  109. [117]

    Xing, C., Wu, W., Wu, Y., Liu, J., Huang, Y., Zhou, M., and Ma, W.-Y. (2017a). Topic aware neural response generation. In AAAI , pages 3351--3357

    2017

  110. [118]

    Xing, C., Wu, W., Wu, Y., Zhou, M., Huang, Y., and Ma, W.-Y. (2017b). Hierarchical recurrent attention network for response generation. arXiv preprint arXiv:1701.07149

    2017 arXiv

  111. [119]

    Yao, K., Peng, B., Zweig, G., and Wong, K.-F. (2016). An attentional neural conversation model with improved specificity. arXiv preprint arXiv:1606.01292

    2016 arXiv

  112. [120]

    Yao, K., Zweig, G., and Peng, B. (2015). Attention with intention for a neural network conversation model. arXiv preprint arXiv:1510.08565

    2015 arXiv

  113. [121]

    Yin, Z., Chang, K.-h., and Zhang, R. (2017). Deepprobe: Information directed sequence understanding and chatbot design via recurrent neural networks. In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , pages 2131--2139. ACM

    2017

  114. [122]

    W., and Rudnicky, A

    Yu, Z., Black, A. W., and Rudnicky, A. I. (2017). Learning conversational systems that interleave task and non-task content. arXiv preprint arXiv:1703.00099

    2017 arXiv

  115. [123]

    Zhao, T., Lu, A., Lee, K., and Eskenazi, M. (2017a). Generative encoder-decoder models for task-oriented spoken dialog systems with chatting capability. arXiv preprint arXiv:1706.08476

    2017 arXiv

  116. [124]

    Zhao, T., Zhao, R., and Eskenazi, M. (2017b). Learning discourse-level diversity for neural dialog models using conditional variational autoencoders. arXiv preprint arXiv:1703.10960

    2017 arXiv

  117. [125]

    Zhou, H., Huang, M., Zhang, T., Zhu, X., and Liu, B. (2017). Emotional chatting machine: Emotional conversation generation with internal and external memory. arXiv preprint arXiv:1704.01074

    2017 arXiv

  118. [126]

    Zhu, Q., Li, Y., and Li, X. (2017). Character sequence-to-sequence model with global attention for universal morphological reinflection. Proceedings of the CoNLL SIGMORPHON 2017 Shared Task: Universal Morphological Reinflection , pages 85--89

    2017

  119. [127]

    and Le, Q

    Zoph, B. and Le, Q. V. (2016). Neural architecture search with reinforcement learning. arXiv preprint arXiv:1611.01578

    2016 arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.