1 / 4

How to Comine Paradigms for Machine Translation

This work explores the integration of various machine translation paradigms to enhance translation quality. Covering Transfer-based MT (RBMT, TMT), Statistical MT (SMT, PSMT, SBMT), and other approaches like Example-Based MT (EBMT) and Interlingua-Based MT (IL-MT), the paper outlines an evolutionary tree of MT methods from 1950 to 2010. It advocates for Hybrid Approaches to MT (HAMT) that assimilate the strengths of diverse techniques into a dominant system, focusing on context, knowledge, and statistical methods to improve overall translation efficiency.

iram
Télécharger la présentation

How to Comine Paradigms for Machine Translation

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. How to Comine Paradigms for Machine Translation Carnegie Mellon University Jaime Carbonell

  2. Alphabet Soup of MT Paradigms • Transfer-rule-based MT (RBMT, TMT, …) • Statistical MT (SMT) • Phrasal-based SMT (PSMT or SMT) • Syntax-based SMT (SBMT) • Example-Based MT (EBMT) • Interlingua-Based MT (IL-MT) • Knowledge-Based MT (KBMT) • Context-Based MT (CBMT) • Hybrid Approaches to MT (HAMT)

  3. An Evolutionary Tree of MT Paradigms Larger-Scale TMT Large-scale TMT Transfer MT Stat Transfer MT Interlingua MT Context-Based MT Analogy MT Example-based MT Stat Syntax MT Statistical MT DecodingMT Phrasal SMT 1950 1980 2010

  4. What Works • Best Overall Recombinant Group (BORG) • Assimilate the best of other paradigms into a dominant one (e.g. SMT)

More Related