Decoding, fast and slow: A case study on balancing trade-offs in incremental, character-level pragmatic reasoning

Zarrieß S, Buschmeier H, Han T, Schüz S (2021)
In: Proceedings of the 14th International Conference on Natural Language Generation. 371–376.

Konferenzbeitrag | Veröffentlicht | Englisch
 
Download
Es wurden keine Dateien hochgeladen. Nur Publikationsnachweis!
Abstract / Bemerkung
Recent work has adopted models of pragmatic reasoning for the generation of informative language in, e.g., image captioning. We propose a simple but highly effective relaxation of fully rational decoding, based on an existing incremental and character-level approach to pragmatically informative neural image captioning. We implement a mixed, ‘fast’ and ‘slow’, speaker that applies pragmatic reasoning occasionally (only word-initially), while unrolling the language model. In our evaluation, we find that increased informativeness through pragmatic decoding generally lowers quality and, somewhat counter-intuitively, increases repetitiveness in captions. Our mixed speaker, however, achieves a good balance between quality and informativeness.
Erscheinungsjahr
2021
Titel des Konferenzbandes
Proceedings of the 14th International Conference on Natural Language Generation
Seite(n)
371–376
Konferenz
14th International Conference on Natural Language Generation
Konferenzort
Aberdeen, Scotland, UK
Konferenzdatum
2021-09-20 – 2021-09-24
Page URI
https://pub.uni-bielefeld.de/record/2956772

Zitieren

Zarrieß S, Buschmeier H, Han T, Schüz S. Decoding, fast and slow: A case study on balancing trade-offs in incremental, character-level pragmatic reasoning. In: Proceedings of the 14th International Conference on Natural Language Generation. 2021: 371–376.
Zarrieß, S., Buschmeier, H., Han, T., & Schüz, S. (2021). Decoding, fast and slow: A case study on balancing trade-offs in incremental, character-level pragmatic reasoning. Proceedings of the 14th International Conference on Natural Language Generation, 371–376.
Zarrieß, S., Buschmeier, H., Han, T., and Schüz, S. (2021). “Decoding, fast and slow: A case study on balancing trade-offs in incremental, character-level pragmatic reasoning” in Proceedings of the 14th International Conference on Natural Language Generation 371–376.
Zarrieß, S., et al., 2021. Decoding, fast and slow: A case study on balancing trade-offs in incremental, character-level pragmatic reasoning. In Proceedings of the 14th International Conference on Natural Language Generation. pp. 371–376.
S. Zarrieß, et al., “Decoding, fast and slow: A case study on balancing trade-offs in incremental, character-level pragmatic reasoning”, Proceedings of the 14th International Conference on Natural Language Generation, 2021, pp.371–376.
Zarrieß, S., Buschmeier, H., Han, T., Schüz, S.: Decoding, fast and slow: A case study on balancing trade-offs in incremental, character-level pragmatic reasoning. Proceedings of the 14th International Conference on Natural Language Generation. p. 371–376. (2021).
Zarrieß, Sina, Buschmeier, Hendrik, Han, Ting, and Schüz, Simeon. “Decoding, fast and slow: A case study on balancing trade-offs in incremental, character-level pragmatic reasoning”. Proceedings of the 14th International Conference on Natural Language Generation. 2021. 371–376.
Link(s) zu Volltext(en)
Access Level
OA Open Access

Export

Markieren/ Markierung löschen
Markierte Publikationen

Open Data PUB

Suchen in

Google Scholar