Teaching machines to converse
Abstract/Contents
- Abstract
- The ability of a machine to communicate with humans has long been associated with the general success of AI. This dates back to Alan Turing's epoch-making work in the early 1950s, which proposes that a machine's intelligence can be tested by how well it, the machine, can fool a human into believing that the machine is a human through dialogue conversations. Despite progress in the field of dialogue learning over the past decades, conventional dialog systems still face a variety of major challenges such as robustness, scalability and domain adaptation: many systems learn generation rules from a minimal set of authored rules or labels on top of handcoded rules or templates, and thus are both expensive and difficult to extend to open-domain scenarios. Meanwhile, dialogue systems have become increasingly complicated: they usually involve building many different complex components separately, rendering them unable to accommodate the large amount of data that we have to date. Recently, the emergence of neural network models the potential to solve many of the problems in dialogue learning that earlier systems cannot tackle: the end-to-end neural frameworks offer the promise of scalability and language-independence, together with the ability to track the dialogue state and then mapping between states and dialogue actions in a way not possible with conventional systems. On the other hand, neural systems bring about new challenges: they tend to output dull and generic responses such as ``I don't know what you are talking about"; they lack a consistent or a coherent persona; they are usually optimized through single-turn conversations and are incapable of handling the long-term success of a conversation; and they are not able to take the advantage of the interactions with humans. This dissertation attempts to tackle these challenges: Contributions are twofold: (1) we address new challenges presented by neural network models in open-domain dialogue generation systems, which includes (a) using mutual information to avoid dull and generic responses; (b) addressing user consistency issues to avoid inconsistent responses generated by the same user; (c) developing reinforcement learning methods to foster the long-term success of conversations; and (d) using adversarial learning methods to push machines to generate responses that are indistinguishable from human-generated responses; (2) we develop interactive question-answering dialogue systems by (a) giving the agent the ability to ask questions and (b) training a conversation agent through interactions with humans in an online fashion, where a bot improves through communicating with humans and learning from the mistakes that it makes.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2017 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Li, Jiwei |
|---|---|
| Associated with | Stanford University, Computer Science Department. |
| Primary advisor | Jurafsky, Dan, 1962- |
| Thesis advisor | Jurafsky, Dan, 1962- |
| Thesis advisor | Brunskill, Emma |
| Thesis advisor | Potts, Christopher, 1977- |
| Advisor | Brunskill, Emma |
| Advisor | Potts, Christopher, 1977- |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Jiwei Li. |
|---|---|
| Note | Submitted to the Department of Computer Science. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2017. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2017 by Jiwei Li
Also listed in
Loading usage metrics...