jurafsky&martin_3rdEd_17 (1).pdf

This idea was first developed by using phrase based

Info icon This preview shows pages 425–427. Sign up to view the full content.

View Full Document Right Arrow Icon
transduce a question to an answer. This idea was first developed by using phrase-based machine translation (Ritter et al., 2011) to translate a user turn to a system response. It quickly became clear, however, that the task of response generation was too different from machine trans- lation. In machine translation words or phrases in the source and target sentences tend to align well with each other; but in conversation, a user utterance may share no words or phrases with a coherent response. Instead, (roughly contemporaneously by Shang et al. 2015 , Vinyals and Le 2015 , and Sordoni et al. 2015 ) transduction models for response generation were modeled instead using sequence to sequence (seq2seq) models (Chapter 25), as shown in Fig. 28.6 . How are you ? I’m fine . EOS Encoding Decoding EOS I’m fine . Figure 28.6 A sequence to sequence model for neural response generation in dialog. A number of modifications are required to the basic seq2seq model to adapt it for the task of response generation. For example basic seq2seq models have a tendency to produce predictable but repetitive and therefore dull responses like “I’m OK” or “I don’t know” that shut down the conversation. This can be addressed by changing the objective function for seq2seq model training to a mutual information objective,
Image of page 425

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
426 C HAPTER 28 D IALOG S YSTEMS AND C HATBOTS or by modifying a beam decoder to keep more diverse responses in the beam (Li et al., 2016a) . Another problem with the simple S EQ 2S EQ response generation architecture is its inability to model the longer prior context of the conversation. This can be done by allowing the model to see prior turns, such as by using a hierarchical model that summarizes information over multiple prior turns (Lowe et al., 2017b) . Finally, S EQ 2S EQ response generators focus on generating single responses, and so don’t tend to do a good job of continuously generating responses that cohere across multiple turns. This can be addressed by using reinforcement learning, as well as techniques like adversarial networks, to learn to choose responses that make the overall conversation more natural ( Li et al. 2016b , Li et al. 2017 ). Fig. 28.7 shows some sample responses generated by a vanilla S EQ 2S EQ model, and from a model trained by an adversarial algorithm to produce responses that are harder to distinguish from human responses (Li et al., 2017) . Input tell me ... how long have you had this falling sickness ? Vanilla-S EQ 2S EQ i’m not a doctor. Adversarial a few months, i guess . Input so i had the doctors test sammy ’s response to conditioning . Vanilla-S EQ 2S EQ sammy wrote the test sammy wrote the test . Adversarial so he took the pills . Input they didn ’t have much success with this problem commander . Vanilla-S EQ 2S EQ they ’re not the only ones who have been in the system .
Image of page 426
Image of page 427
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern