Grounding language models in spatiotemporal context
Natural language is rich and varied, but also highly structured. The rules of grammar are a primary source of linguistic regularity, but there are many other factors that govern patterns of language use. Language models attempt to capture linguistic regularities, typically by modeling the statistics...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
International Speech Communication Association
2014
|
| Online Access: | http://hdl.handle.net/1721.1/91490 https://orcid.org/0000-0002-2564-8909 https://orcid.org/0000-0002-4333-7194 |
| _version_ | 1826208874379608064 |
|---|---|
| author | Roy, Brandon C. Vosoughi, Soroush Roy, Deb K |
| author2 | Program in Media Arts and Sciences (Massachusetts Institute of Technology) |
| author_facet | Program in Media Arts and Sciences (Massachusetts Institute of Technology) Roy, Brandon C. Vosoughi, Soroush Roy, Deb K |
| author_sort | Roy, Brandon C. |
| collection | MIT |
| description | Natural language is rich and varied, but also highly structured. The rules of grammar are a primary source of linguistic regularity, but there are many other factors that govern patterns of language use. Language models attempt to capture linguistic regularities, typically by modeling the statistics of word use, thereby folding in some aspects of grammar and style. Spoken language is an important and interesting subset of natural language that is temporally and spatially grounded. While time and space may directly contribute to a speaker’s choice of words, they may also serve as indicators for communicative intent or other contextual and situational factors. To investigate the value of spatial and temporal information, we build a series of language models using a large, naturalistic corpus of spatially and temporally coded speech collected from a home environment. We incorporate this extralinguistic information by building spatiotemporal word classifiers that are mixed with traditional unigram and bigram models. Our evaluation shows that both perplexity and word error rate can be significantly improved by incorporating this information in a simple framework. The underlying principles of this work could be applied in a wide range of scenarios in which temporal or spatial information is available. |
| first_indexed | 2024-09-23T14:14:00Z |
| format | Article |
| id | mit-1721.1/91490 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T14:14:00Z |
| publishDate | 2014 |
| publisher | International Speech Communication Association |
| record_format | dspace |
| spelling | mit-1721.1/914902022-10-01T19:55:49Z Grounding language models in spatiotemporal context Roy, Brandon C. Vosoughi, Soroush Roy, Deb K Program in Media Arts and Sciences (Massachusetts Institute of Technology) Vosoughi, Soroush Roy, Brandon C. Vosoughi, Soroush Roy, Deb K. Natural language is rich and varied, but also highly structured. The rules of grammar are a primary source of linguistic regularity, but there are many other factors that govern patterns of language use. Language models attempt to capture linguistic regularities, typically by modeling the statistics of word use, thereby folding in some aspects of grammar and style. Spoken language is an important and interesting subset of natural language that is temporally and spatially grounded. While time and space may directly contribute to a speaker’s choice of words, they may also serve as indicators for communicative intent or other contextual and situational factors. To investigate the value of spatial and temporal information, we build a series of language models using a large, naturalistic corpus of spatially and temporally coded speech collected from a home environment. We incorporate this extralinguistic information by building spatiotemporal word classifiers that are mixed with traditional unigram and bigram models. Our evaluation shows that both perplexity and word error rate can be significantly improved by incorporating this information in a simple framework. The underlying principles of this work could be applied in a wide range of scenarios in which temporal or spatial information is available. 2014-11-07T15:00:20Z 2014-11-07T15:00:20Z 2014-09 Article http://purl.org/eprint/type/ConferencePaper http://hdl.handle.net/1721.1/91490 Roy, Brandon C., Soroush Vosoughi, and Deb Roy. "Grounding language in spatiotemporal context." The 15th Annual Conference of the International Speech Communication Association, September 14-18, 2014. https://orcid.org/0000-0002-2564-8909 https://orcid.org/0000-0002-4333-7194 en_US http://www.interspeech2014.org/public.php?page=program_details.html Proceedings of the 15th Annual Conference of the International Speech Communication Association Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf International Speech Communication Association Vosoughi |
| spellingShingle | Roy, Brandon C. Vosoughi, Soroush Roy, Deb K Grounding language models in spatiotemporal context |
| title | Grounding language models in spatiotemporal context |
| title_full | Grounding language models in spatiotemporal context |
| title_fullStr | Grounding language models in spatiotemporal context |
| title_full_unstemmed | Grounding language models in spatiotemporal context |
| title_short | Grounding language models in spatiotemporal context |
| title_sort | grounding language models in spatiotemporal context |
| url | http://hdl.handle.net/1721.1/91490 https://orcid.org/0000-0002-2564-8909 https://orcid.org/0000-0002-4333-7194 |
| work_keys_str_mv | AT roybrandonc groundinglanguagemodelsinspatiotemporalcontext AT vosoughisoroush groundinglanguagemodelsinspatiotemporalcontext AT roydebk groundinglanguagemodelsinspatiotemporalcontext |