Hierarchical Neural Network Generative Models for Movie Dialogues

By Iulian Serban, Alessandro Sordoni, Yoshua Bengio, Aaron C. Courville, Joelle Pineau. 2015

195 citations recorded by the source corpus.

Subfields: Natural Language Processing, Generative Models

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

In Generative Models

Citations in the corpus

Exploring the Limits of Language Modeling (2016, methodology): “For example, when trained on movie subtitles (Serban et al., 2015; Vinyals & Le, 2015), these language models are able to generate basic answers to questions about object colors, facts about people, etc. Lastly, recently proposed sequence-to-sequence models employ conditional language models…”
Generating Informative and Diverse Conversational Responses via Adversarial Information Maximization (2018, methodology): “For embedding-based evaluation, we used a word2vec embedding trained on GoogleNews Corpus 3 , recommended by [44].”
Deep Reinforcement Learning for Dialogue Generation (2016, background): “One concrete example is that SEQ2SEQ models tend to generate highly generic responses such as“I don’t know” regardless of the input (Sordoni et al., 2015; Serban et al., 2015b; Serban et al., 2015c; Li et al., 2015).”
Smart Reply: Automated Response Suggestion for Email (2016, methodology): “In [21] the authors rely on having an SMT system to generate n-best lists, while [19] and [24], like this work, develop fully generative models.”
Continuously Learning Neural Dialogue Management (2016, background): “Recently an emerging line of research has focused on training a network-based dialogue model, mostly in text-input schemes (Vinyals and Le, 2015; Serban et al., 2015; Wen et al., 2016; Bordes and Weston, 2016).”
GuessWhat?! Visual Object Discovery through Multi-modal Dialogue (2016, methodology): “This can be encoded into a fixed size vector by using either an LSTM encoder [15] or an HRED encoder [36].”
Adversarial Evaluation of Dialogue Models (2017, methodology): “Much recent work has employed RNN encoder-decoder models to translate from utterance to response ([9], [8], [11]).”
Challenges in Building Intelligent Open-domain Dialog Systems (2019, background): “Early work has focused on better representing dialog contexts [119, 121] using hierarchical models, which can be viewed as implicit modeling of contextual consistency.”
Sample-efficient Actor-Critic Reinforcement Learning with Supervised Data for Dialogue Management (2017, background): “Recently an emerging line of research has focused on training neural network-based dialogue models, mostly in text-based systems (Vinyals and Le, 2015; Shang et al., 2015; Serban et al., 2015; Wen et al., 2017; Bordes et al., 2017).”
Commonsense Knowledge Aware Conversation Generation with Graph Attention (2018, methodology): “…models [Sutskever et al. , 2014; Bahdanau et al. , 2014] have been successfully applied to large-scale conversation generation, including neural responding machine [Shang et al. , 2015], hierarchical recurrent models [Serban et al. , 2015], and many others [Sor-doni et al. , 2015].”
An Attentional Neural Conversation Model with Improved Specificity (2016, methodology): “This suggests that the benefit of incorporating hierarchical structure in both HRED and AWI is more apparent with larger training data.”
Incorporating Unstructured Textual Knowledge Sources into Neural Dialogue Systems (2015, background): “Deep learning approaches that are trainable end-to-end are beginning to surface in the dialogue community [16, 15], which essentially combine parsing, dialogue planning and natural language generation modules.”
A Large-Scale Chinese Short-Text Conversation Dataset (2020, methodology): “As an important public resource of dialogue corpora, movie scripts such as OpenSubtitles have been used to construct dialogue datasets [14,26] .”
Personalizing a Dialogue System With Transfer Reinforcement Learning (2016, background): “Dialogue systems can be classified into two classes: open domain dialogue systems (Ritter, Cherry, and Dolan 2011; Galley et al. 2015; Serban et al. 2015; Li et al. 2016b; Mou et al. 2016) and task-oriented dialogue systems (Levin, Pieraccini, and Eckert 1997; Young et al.”