Abstract
Reinforcement learning gives a way to learn under what circumstances to perform which actions. However, this approach lacks a formal framework for specifying hand-crafted restrictions, for specifying the effects of the system actions, or for specifying the user simulation. The information state approach, in contrast, allows system and user behavior to be specified as update rules, with preconditions and effects. This approach can be used to specify complex dialogue behavior in a systematic way. We propose combining these two approaches, thus allowing a formal specification of the dialogue behavior, and allowing hand-crafted preconditions, with remaining ones determined via reinforcement learning so as to minimize dialogue cost.
| Original language | English (US) |
|---|---|
| Title of host publication | NAACL HLT 2007 - Human Language Technologies 2007: The Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Main Conference |
| Pages | 268-275 |
| Number of pages | 8 |
| State | Published - 2007 |
| Event | Human Language Technologies 2007: The Conference of the North American Chapter of the Association for Computational Linguistics, NAACL HLT 2007 - Rochester, NY, United States Duration: Apr 22 2007 → Apr 27 2007 |
Other
| Other | Human Language Technologies 2007: The Conference of the North American Chapter of the Association for Computational Linguistics, NAACL HLT 2007 |
|---|---|
| Country | United States |
| City | Rochester, NY |
| Period | 4/22/07 → 4/27/07 |
Fingerprint
ASJC Scopus subject areas
- Language and Linguistics
- Linguistics and Language
Cite this
Combining reinforcement learning with information-state update rules. / Heeman, Peter.
NAACL HLT 2007 - Human Language Technologies 2007: The Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Main Conference. 2007. p. 268-275.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Combining reinforcement learning with information-state update rules
AU - Heeman, Peter
PY - 2007
Y1 - 2007
N2 - Reinforcement learning gives a way to learn under what circumstances to perform which actions. However, this approach lacks a formal framework for specifying hand-crafted restrictions, for specifying the effects of the system actions, or for specifying the user simulation. The information state approach, in contrast, allows system and user behavior to be specified as update rules, with preconditions and effects. This approach can be used to specify complex dialogue behavior in a systematic way. We propose combining these two approaches, thus allowing a formal specification of the dialogue behavior, and allowing hand-crafted preconditions, with remaining ones determined via reinforcement learning so as to minimize dialogue cost.
AB - Reinforcement learning gives a way to learn under what circumstances to perform which actions. However, this approach lacks a formal framework for specifying hand-crafted restrictions, for specifying the effects of the system actions, or for specifying the user simulation. The information state approach, in contrast, allows system and user behavior to be specified as update rules, with preconditions and effects. This approach can be used to specify complex dialogue behavior in a systematic way. We propose combining these two approaches, thus allowing a formal specification of the dialogue behavior, and allowing hand-crafted preconditions, with remaining ones determined via reinforcement learning so as to minimize dialogue cost.
UR - http://www.scopus.com/inward/record.url?scp=70349253156&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=70349253156&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:70349253156
SP - 268
EP - 275
BT - NAACL HLT 2007 - Human Language Technologies 2007: The Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Main Conference
ER -