A Dynamic Semantic Algorithm for Interpreting Action Verbs Based on a Sequential Model
| Project/Area Number |
10610457
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
英語・英米文学
|
|---|
| Research Institution | Saitama University |
|---|
Principal Investigator |
NISHINA Hiroyuki Saitama University, Faculty of Liberal Arts, Professor, 教養学部, 教授 (20125777)
|
|---|
| Project Period (FY) |
1998 – 1999
|
| Project Status |
Completed (Fiscal Year 1999)
|
| Budget Amount *help |
¥1,300,000 (Direct Cost: ¥1,300,000)
Fiscal Year 1999: ¥600,000 (Direct Cost: ¥600,000)
Fiscal Year 1998: ¥700,000 (Direct Cost: ¥700,000)
|
|---|
| Keywords | interpretation procedure / semantic model / moving parts / rotation formula / formal semantics / action verb / individual as a set / movements / 行為関数 / 運動複合体 / 内容点集合関数 / 回帰移動関数 / 他動行為 |
|---|
| Research Abstract |
We conducted a preliminary research to construct a cognitive semantic interpretation theory for action verbs based on a formal model as follows :
1. A new semantic model was proposed in which each individual is defined as a set forming a volume shifting in a temporal space. Focusing on the subsets of an individual performing an action, its parts involved in each interval were abstracted for the semantic definition of the relevant action verb, i.e. the name for that action. With these, we were enabled to describe an action complex spatiotemporally.
2. An interpretive theory for action verbs based on spatiotemporal model was proposed (Nishina '98). To define an action verb, we proposed a sequence of feature sets referring to the parts involved in the action denoted by the verb. An algorithm using a feature sequence was devised for deciding which action is contained in a complex of movements. Several interpretive rules reflecting this algorithm was also defined.
3. This model is being further refined as a wire-frame model and its effectiveness was discussed (Nishina '00). The skeleton of an individual was described in a wire frame and its movements were shown to be expressed in rotation formulae, which describe each point as rotating round some axis. Complex action structure was achieved by converting these formulae to individual functions on axis. We confirmed that by comparing this structure with some operations added on it with the defining sequence for an action verb, we can decide if that complex contains the action whose name is that verb. Thus, this model proved to be a test bench for a cognitive formal model for semantic interpretation.
|
Report
(3 results)
Research Products
(14 results)
