|
Symbyonics, LLC |
|
|
Augus 8, 2016 Issue |
||
|
CEOCFO MAGAZINE |
||
|
|
||
|
Machine Translation Quality Estimation Technology Revolutionizing the Machine-Mediated Language Services Market |
||
|
Co-Founder
Symbyonics, LLC
Interview conducted by: Lynn Fosse, Senior Editor CEOCFO Magazine
CEOCFO: Mr. Marcu, what is the idea behind Symbyonics? Mr. Marcu: The idea behind Symbyonics is to improve the efficiency of machine mediated labor markets while ensuring that people are paid fairly in the process.
CEOCFO: On a very fundamental level, what do you understand about accomplishing these goals? Mr. Marcu: I think the best way to articulate this is by means of a few examples. I remember reading a fascinating book several years back which is called “Shadow Work: The Unpaid, Unseen Jobs That Fill Your Day”. The premise of that book, written by Craig Lambert, is that there is a tremendous upsurge in shadow work, i.e., uncompensated work that we came to do on a daily basis: we make our own travel arrangements; we prepare our taxes; we bag our groceries; we fill up the gas tanks of our cars; and we manage personal banking operations. It is a great deal of stuff that, in some instances, we prefer to do because we get a better deal. However, as machines become increasingly smarter, there is also some shadow work that happens because machines are just not good enough and we would definitely not want to do that sort of shadow work. Let us consider, for example, what happens when you call a customer support center: you call and you are taken in baby steps through a long menu of choices and, after ten minutes, you may end up back at the beginning of the conversation. Those ten minutes are wasted because speech recognition and natural language understanding technologies are not yet good enough. When immature technologies reduce our spare time, that is annoying; it gets worse though and it becomes unfair when immature technologies impact negatively one’s compensation in professional environments.
CEOCFO: Where does Symbyonics come in? Mr. Marcu: Symbyonics is focusing on markets where this happens and is optimizing and making these markets more efficient by quantifying the amount of shadow work that may be hidden in a employer-contractor/freelancer engagement. By quantifying the amount of shadow, unpaid work induced by the adoption of advanced, emerging technologies, Symbyonics accomplishes two goals: (i) It helps work doers estimate the amount of shadow work that they may be undertaking and it empowers them to make better choices that, over time, increase their compensation. For example, when evaluating two jobs, they can say, “I am going to take job A that is worth $150 because Symbyonics helped me understand that job B, which is priced at $200, incorporates $80 worth of shadow, uncompensated work. During the time I would work on the $80 uncompensated, shadow work, I could increase my compensation by working on another job C.” (ii) Understanding the costs of shadow work also helps employers & job givers: they can automatically evaluate business quotes with higher accuracy, reduce churn by keeping their employees and contractors happy, and increase the overall quality of their products.
CEOCFO: How does this theory turn into a business or a business model? Mr. Marcu: This theory turns into a business model when it is implemented in a market where the adoption of an emerging technology is already creating significant pain. In the Language Services Market, businesses are seeking to obtain human level translations from a work force of more than half a million professional translators spread around the globe. In the past, a business would go to a professional translator and say: “I have a five-thousand-word document. Would you please translate it for me?” The parties would agree that the professional translator could do the work for, let’s say, 10 or 15 cents/word and both parties would end up happy: the translator for being compensated fairly and the business for receiving a high quality product. As Statistical Machine Translation technology has taken the world by storm, businesses have attempted to capitalize on it: some are obtaining a first draft translation from a machine and they are then asking professionals to fix it for a lower per-word price. When the automatic translation is of very high quality, paying professional translators a lower price makes complete sense, because there is less work to do. However, when translations are of low quality, paying a lower price is unfair to the professional translator because it creates unpaid, shadow work; and it is bad for the business because under unrealistic time constraints, the quality of the work suffers. In most cases, the quality of a translation produced by a machine is unknown: so both parties in a professional engagement may end up not getting their desired outcomes. Symbyonics brings transparency into the language services market by providing a suite of online tools & services that enables language service market participants to identify in each translation job the sentences that are translated with high accuracy by machines, which can be edited by humans at a discounted price, and the sentences that cannot be translated accurately by machines, which need to be translated by professionals at a full price. The suite of tools provides price estimates that are personalized to the skills and knowledge of the professionals who are using them and that learn to adapt over time in order to maximize the value a professional translator may get from using Machine Translation technology as an aid. In other words, Symbyonics eliminates commercial risk and enables professional translators to increase their compensation over time by focusing on jobs with no shadow, uncompensated elements.
CEOCFO: This is your Fair Trade Translation service? Mr. Marcu: Yes, this is FairTradeTranslation.com.
CEOCFO: It is fairly new. What has been the response? How are you gaining attention? Mr. Marcu: It is indeed a very new concept. I am actually familiar with bringing new concepts to market: I am one of the original founders of Language Weaver Inc., which was the first company that brought to market a Statistical Machine Translation product. Fifteen years ago, Statistical Machine Translation was faced with skepticism; it first gained traction only in government markets. But today, Google alone, one of the main providers, helps 500 million users translate more than 100 billion words daily. By comparison, Fair Trade Translation was received from the beginning with significantly more enthusiasm: almost 1,000 alpha users have provided us with more than 300 product update requests, which we have incorporated into the online service that went live a few weeks ago.
CEOCFO: Are you funded for your next steps? Are you looking at partnerships or financing? Mr. Marcu: We are self-funded. We have used proceeds we generated from being engaged in successful entrepreneurial activities in the past to fund Fair Trade Translation. Our initial goal was to put in front of professional translators something that they like and want to use. We are now transitioning to the next phase and work on creating partnerships with organizations through which we can reach larger groups of professional translators faster.
CEOCFO: What is the takeaway for our readers? Mr. Marcu: I think it depends on who the readers are. For professional translators, Fair Trade Translation is an online service that enables them to optimize and maximize their compensation over time. For Language Service Providers, Fair Trade Translation is a service that can be used to validate business quotes and insure that their translation workforces are used efficiently; Fair Trade Translation also enables Language Service Providers deal optimally with their long tails of translation projects and extend the set of services they can provide. For readers who care about technology, the takeaway is that the boundary between humans and machines is an area worth paying attention to: research and business advances will lead to significant disruption in the way we, humans, interact and collaborate with machines. |
“By quantifying the amount of shadow, unpaid work induced by the adoption of advanced, emerging technologies, Symbyonics accomplishes two goals: (i) It helps work doers estimate the amount of shadow work that they may be undertaking and it empowers them to make better choices that, over time, increase their compensation. (ii) Understanding the cost of shadow work also helps employers & job givers: they can automatically evaluate business quotes with higher accuracy, reduce churn by keeping their employees and contractors happy, and increase the overall quality of their products.”- Daniel Marcu
Symbyonics, LLC
Contact: Daniel Marcu +1 310 844 6985
google plus press releases news - events
|
|
|
|
||
|
|
||
|
|
||
|
|
||
Daniel
Marcu