30.8% Better Than Google NMT For Swedish-to-English Contracts

This Swedish-to-English Contracts translator uses a Slate Desktop personalized SMT that’s 30.8% more productive than Google NMT.

A Slate Rocks customer (a translator) created a translation engine with his or her translation memories (TMs) using a personal computer. This page describes the engine and compares the translator’s Slate Desktop experience to an experience using Google’s new-and-improved neural machine translation (NMT) technology. You can read the entire report with thirth (30) more customer experiences by downloading the full report Study of Machine Translated Segment Pairs.

Slate Desktop Engine Details

The customer started with translation memories in the language pair and industry of his or her work, totaling the estimated corpus size number of segments. Slate Desktop cleaned the TMs, prepared a training corpus and built the engine. Note that these processes typically runs overnight. During that processing, Slate Desktop also extracted a representative set consisting of randomly selected segments from the training corpus.

sourcesv
targeten
subject domaincontract
estimated corpus size80,000
segments per representative set2,336
words per source segment14
words per target segment18

Benchmark Score Comparison

The segment pairs in the representative set are representative of the translator’s daily work. By focusing on one translator’s experience, these scores indicate a level of work reduction this customer will likely experience in his or her daily work using the respective MT system (Google or Slate Desktop) with 95% confidence.

 Google
NMT
Slate Desktop
SMT
words per MT segment1718
BLEU score35.3874.86
exact MT match (count)98787
exact MT match (percent)0.0420.337
words per exact MT match (count)413
filtered BLEU score (no exact MT matches)35.0267.18
segments requiring edit (count)2,2381,549
character edits per segment3627
total character edits80,56841,823

These scores indicate this customer using Slate Desktop will likely spend significantly less time editing MT suggestions than if he or she were using Google for this work. This is because Slate Desktop creates engines with the customers translation memories and optimizes them to predict how the customer translates. While on the other hand, Google optimizes its NMT service for millions of customers with countless demands.

Google’s three (3) longest exact MT matches

The exact MT match (count) in the Benchmark Scores table (above) is the number of segments that Google NMT successfully matched to the translator’s actual work, i.e. Google successfully predicted the translator’s actions. The three segments in this table are the exact MT match segments with the longest length. This translator can expect to experience these kinds Google NMT results while translating these kinds of project.

sven (Google and translator)
Scenariospecifika antaganden för utvalda scenarier belyses i Tabell A-3.Scenario-specific assumptions for selected scenarios are illustrated in Table A-3.
Noteringar från tidigare möten i kursiv text.Notes from previous meetings in italics.
Föraren har full kontroll över fordonet.The driver has full control of the vehicle.

Slate’s three (3) longest exact MT matches

The exact MT match (count) in the Benchmark Scores table (above) is the number of segments that this translator’s Slate Desktop engine successfully matched to the translator’s actual work, i.e. Slate Desktop successfully predicted the translator’s actions. The three segments in this table are the exact MT match segments with the longest length. This translator can expect to experience these kinds Slate Desktop results while translating these kinds of project.

sven (Slate and translator)
Denna TSD (teknisk specifikation för driftskompatibilitet) beskriver hur ett visst delsystem ska vara beskaffat för att uppfylla de väsentliga kraven och säkerställa driftskompatibilitet inom unionens järnvägssystem, så som beskrivs i artikel 1 i direktiv 2008/57/EG.This technical specification for interoperability (TSI) is a specification by which a particular subsystem is addressed in order to meet the essential requirements and ensure the interoperability of the Union’s rail system as described in Article 1 of Directive 2008/57/EC.
Ett sådant utförande ställer dock höga krav på utförande och detaljutformning (inkl. brandgasventilationen) och därmed kan en robustare lösning vara att delar av skyddsnivån fördelas till exempelvis brandskyddande sprutbetong, och andra delar fördelas till inklädnad/undertak.However such a specification puts high demands on the specification and components (incl. fire gas ventilation) and as such, a more robust solution can be to allocate parts of the protection level specification to e.g. fire resistant injected concrete and allocate other parts to the cladding/ceiling.
Revisionerna skall utföras minst en gång per år och minst en revision skall utföras under pågående arbete (tillverkning, montering eller installation) inom ramen för det delsystem som är föremål för EG-kontrollförfarandet i punkt 6.The frequency of the audits shall be at least once a year, with at least one audit during the time period of performing relevant activities (manufacture, assembling or installation) for the subsystem being the subject of the EC verification procedure mentioned under point 6.
Advertisements