Today’s translation world struggles with technology-driven business trends from translation quality & work productivity, to competitive edge & customer confidentiality. But these are not challenges for everyone. So…
- If you don’t want to deliver quality translations that make you proud,
- If you don’t want a way to complete 5 days of work in 3 or 4 days,
- If you don’t want to strengthen your competitive edge,
- If you don’t need to protect customers’ confidentiality…
Well if these describe you, I’ll save you some time. Just stop here. There’s no reason to go any further because everything that follows relates to one of these goals.
Great! You’re here. I take that to mean you’re looking for something related to these goals. I have a product, Slate Desktop, that many translators, teams, agencies and small LSPs have found helpful in realizing these goals.
Introducing Slate Desktop
Slate Desktop is software that adds a machine translation service to your personal computer. The service runs in the background as a machine translation provider for computer-assisted translation tools, applications like memoQ, Trados Studio, CafeTran, etc. that also run on your PC.
There’s nothing new about machine translation providers for your computer-assisted translation tool. Online services like Google, Microsoft, KantanMT, DeepL and others have been providers for years.
You sign-up, enter your credit card number and add a provider to your translation project. The provider sends your client’s information to their Internet service. The service returns machine-translated segments. You send post-edited work to customers and pay the service fee on next month’s credit card bill.
Online services are easy, instant and automatic. So maybe these questions that follow crossed your mind: How is Slate Desktop like the online services? How is Slate Desktop different from the online services? Why use Slate Desktop instead of the online services? Get ready because the answers might surprise you.
How is Slate Desktop like the online services?
Slate Desktop is easy. You configure your computer-assisted translation tool to use Slate Desktop just like you configure it to use those online services.
Slate Desktop is automatic. It generates machine translated segments for your computer-assisted translation tool much like online services work automatically.
Slate Desktop has mature machine learning and artificial intelligence technologies. They are exactly the same technologies used by online services for over a decade.
Slate Desktop delivers a “brand” of quality. Some translators prefer the “Google brand.” Others prefer the “DeepL brand.” Some prefer the “Slate Desktop brand.” You get the idea.
How is Slate Desktop different from the online services?
Slate Desktop runs on your PC, not the Internet. It works in the background and delivers machine-translated segments without sending your customer’s information to the Internet.
Slate Desktop’s machine learning and artificial intelligence technologies are configured for one or a few translators. Online services configure their technologies for millions of users.
Out of the box, Slate Desktop does not instantly translate. Unlike online services, you must first build a machine translation model to “teach” Slate Desktop how you want it to translate.
Why use Slate Desktop instead of the online services?
Before answering this question, we need to review some breakthroughs in machine learning and artificial intelligence.
Machine Learning and Artificial Intelligence
Machine learning technology is the heart of modern machine translation services. It “studies” translation memories for hours, days or weeks depending on the size of the translation memories. And it learns the vocabulary, grammar, syntax, glossaries, and how to translate. Finally, it saves that knowledge in translation models.
Most online services “teach” their models with “generic” knowledge from many translation memories covering a broad range of topics. As a result, these services are configured to sacrifice translation quality in favor of serving millions of users across a broad range of topics.
Some online services “teach” their models with specialized knowledge from translation memories about that specialty. As a result, these “customized” services are configured to deliver better quality translations than generic systems, for translation work about the specialty, but quality degrades significantly outside the specialty.
If machine learning is the heart of modern machine translation services, artificial intelligence technology is the soul. It uses the knowledge in a model to generate translation for source language segments. When knowledge is missing from the model, the artificial intelligence generates a senseless and unusable translation.
Complexity and Cost Breakthroughs
Before Slate Desktop, it was complex and expensive to build a customized model for a specialty. It required experts with specialized skills. It required expensive computers. Therefore, customization was only cost-effective for very large volume projects.
Slate Desktop is a breakthrough in complexity. Its simple. It’s easy to use. Translators without specialized technical skills can build customized models using the same Windows PCs that runs their computer-assisted translation tools. They can build customized models using the translation memories created with their computer-assisted translation tools.
Slate Desktop is a breakthrough in cost. You buy a one-time fixed fee license. There are no subscriptions, no usage fees. Translators, teams and small agencies build and use customized models for a fraction of the cost to hire experts with specialized skills or use online services.
Slate Desktop’s breakthroughs open opportunities for all translation projects to benefit from simple and inexpensive customized models. Projects of any size and in any price range have access to better quality from customized models.
These breakthroughs create opportunities, but do you remember the last “difference” between Slate Desktop and online services? That is, online services are instant, but Slate Desktop “out of the box” does not instantly translate.
Well, now that we know we need translation memories to build a custom model for better quality. So, let’s look at some alternate design strategies before we return to the question “why use Slate Desktop.”
Slate Desktop could package generic models. It would be priced higher to cover the extra work and it would become the same as generic online services that sacrifice quality for convenience.
Slate Desktop could require you to send your translation memories to an expert who would build your models. That’s expensive and your customer’s confidentiality agreements would be at risk.
There’s a reason for this distinction between Slate Desktop and online services. I know that the requirement to build a customized translation model seems like a big inconvenience. In fact, it is one of Slate Desktop’s biggest advantages and it’s our segue to answer the third question.
Why use Slate Desktop instead of the online services? (continued)
I’ve done everything I can to make Slate Desktop easy to use. That’s why machine learning, artificial intelligence and Slate Desktop run easily on MS Windows. That’s what makes Slate Desktop “custom machine translation made simple.” So…
The first answer to “why” is because it’s easy to build custom translation models with Slate Desktop.
When Slate Desktop studies translation memories with translation units that you translated, it learns your bilingual knowledge, your vocabulary and your translating style. This step “teaches” Slate Desktop to translate like you translate, to deliver quality that meets your quality demands, to generated the best possible custom translations for you.
The 2nd answer to “why” is because your custom translation model delivers excellent translations with quality your way.
But after a closer look, these answers are incomplete. So, let’s step back. The results below describe about 90% of all customer experiences. Unfortunately, I can’t guarantee you’ll experience these results because about 10% of customers don’t.
Even though I can’t guarantee your results, I can and do guarantee my commitment to serve you and to help you get your best results. My 60-day money-back guarantee embodies that commitment.
The 3rd answer to “why” might be “because you shouldn’t.” You have a 10% chance that Slate Desktop will not work to your quality demands, BUT you will never now until you to try Slate Desktop. You’re welcome to exercise our 60-day refund policy.
There’s a huge elephant in the room sitting next to the requirement to build your models. If you don’t have translation memories, you simply can not use Slate Desktop. That is, you must have a supply of translation memories. In general, more translation units are better – if they’re in the specialty.
The 4th answer to “why” might be “because you can’t.” I can’t guarantee how many translation units will perform well for you, but I can say with confidence, starting with fewer than 80,000 translation units is highly unlikely to deliver usable results.
Finally, I know that Slate Desktop is not the only game in town. Some generic online services are pretty darn good for some language pairs and some generic subjects. So…
The 5th reason “why” is because you don’t have to use it “instead of.” You’re free to use Slate Desktop “in cooperation with” other services. Slate Desktop is as much an expression of your freedom to choose as it is a pursuit of quality and productivity. With no per-word fees, no subscriptions and extra costs, it’s a pretty good deal.
The Four Original Goals
I opened this conversation with a simple question. Are you looking to achieve any one or more of four goals in your efforts adapt to today’s technology-driven business trends? In light of breakthroughs in technology, complexity and cost, let’s continue our conversation by exploring how Slate Desktop addresses those four goals.
Translation Quality: Thirty-one customers worked with me to analyze 73,150 human translations, their Slate Desktop customized models and a popular online service. The translations covered 16 language pairs with 5 source languages and 13 target languages across 13 different specialties. On average, Slate Desktop’s quality exactly and independently matched the translator’s quality 7 times more often than the online service. This is a link to detailed reports for all 31 customers. https://www.slate.rocks/portfolio/translators-experiences/
Productivity: The analysis also calculated each translator experienced a 50% reduction in the post-edit work to fix the erred segments. Overall, work reduction from exact matches and editing resulted in a 25% to 30% increase in productivity. In practical terms, the customers now finish what were 5-day jobs in 3 or 4 days. This is a link to the analysis of the 31 comparisons. https://www.slate.rocks/downloads/study-of-machine-translated-segment-pairs/
Competitive Edge: Countless translators around the world post-edit every day. They start with the same erred generic segments from online services. They do the same work to fix the errors. Nothing distinguishes one from the other. Slate Desktop and your custom translation model are unique. Your productivity and translation quality distinguish you and give you a competitive edge.
Customer Confidentiality: Slate Desktop runs on your PC. Whether you’re building a customized translation model or working in your computer-assisted translation tool, Slate Desktop never sends your customers’ information to the Internet.
A Right Choice for Everyone
You know your goals. You see how breakthroughs in technology, complexity and cost open new opportunities to achieve them. You have a good idea of how Slate Desktop sets itself apart from online services. It’s a good foundation.
Whether you’re an individual translator, an agency or small language service provider… whether you work with one language pair or dozens… one of these links takes you to a product that’s right for you.