Language is messy. Ask any person who has ever had to learn a second language and they will tell you that the most difficult aspect isn’t learning all the rules, but understanding the exceptions to the rules — the real-world application of the language.Continue reading
Since 2016, there has been a sharp increase in open source machine translation projects based on neural networks or Neural Machine Translation (NMT) led by companies such as Google, Facebook and SYSTRAN. Why have machine translation and NMT-related innovations become the new Holy Grail for tech companies? And does the future of these companies rely on machine translation?
Never before has a technological field undergone so much disruption in such a short time. Invented in the 1960s, machine translation was first based on grammatical and syntactical rules until 2007. Statistical modelling (known as statistical translation or SMT), which matured particularly due to the abundance of data, then took over. Although statistical translation was introduced by IBM in the 1990s, it took 15 years for the technology to reach mass adoption. Neural Machine Translation on the other hand, only took two years to be widely adopted by the industry after being introduced by academia in 2014, showing the acceleration of innovation in this field. Machine translation is currently experiencing a golden age of technology.
From Big Data to Good Data
Not only have these successive waves of technology differed in their pace of development and adoption, but their key strengths or “core values” have also changed. In rule-based translation, value was brought by code and accumulated linguistic resources. For statistical models, the amount of data was paramount. The more data you had, the better the quality of your translation and your evaluation via the BLEU score (Bilingual Evaluation Understudy, the most widely used algorithm measuring machine translation quality). Now, the move to Machine translation based on neural networks and Deep Learning is well underway and has brought about major changes. The engines are trained to learn language as a child does, progressing step by step. The challenge is not only to process exponential data (Big Data) but more importantly to feed the engines the most qualitative data possible. Hence the interest in “Good data.”
[This article originally appeared on Kirti Vashee’s Blog]
This is the final post for the 2017 year, a guest post by Jean Senellart who has been a serious MT practitioner for around 40 years, with deep expertise in all the technology paradigms that have been used to do machine translation. SYSTRAN has recently been running tests building MT systems with different datasets and parameters to evaluate how data and parameter variation affect MT output quality. As Jean said:
” We are continuously feeding data to a collection of models with different parameters – and at each iteration, we change the parameters. We have systems that are being evaluated in this setup for about 2 months and we see that they continue to learn.”
This is more of a vision statement about the future evolution of this (MT) technology, where they continue to learn and improve, rather than a direct reporting of experimental results, and I think is a fitting way to end the year in this blog.
It is very clear to most of us that deep learning based approaches are the way forward for continued MT technology evolution. However, skill with this technology will come with experimentation and understanding of data quality and control parameters. Babies learn by exploration and experimentation, and maybe we need to approach our continued learning, in the same way, learning from purposeful play. Is this not the way that intelligence evolves? Many experts say that AI is going to be driving learning and evolution in business practices in almost every sphere of business.
SYSTRAN’s solution are used every day by various types of companies across many industries to get the most accurate and secure automatic translations on any type of content – from sensitive documents to websites to mobile apps and much more. We’d like to focus today on how one of our clients – Alvarez & Marsal, a consultancy firm- uses SYSTRAN’s platform to manage eDiscovery projects with the highest efficiency and accuracy.
The processes and tools used in eDiscovery scenarios are, most of the time, quite complex given the large volumes of electronic data produced. Unlike hard-copy evidence, e-documents are a lot more dynamic and contain various metadata that demand the highest translation quality in order to eliminate any claims of spoliation at any time in a litigation case.
Phil Beckett, the firm’s Managing Director, who has recently been named ‘Investigation Digital Forensic Expert of the year’ by Who’s Who Legal is talking to us about how SYSTRAN’s solutions plug into their internal processes to manage their projects end to end.
Phil Beckett – Managing Director at Alvarez & Marsal
This article was originally published in inside BIGDATA by Ken Behan. How Neural Machine Translation Will Help Online Marketplaces Turn their Individual Sellers into Global Players.
In this special guest feature, Ken Behan, Chief Growth Officer at SYSTRAN, discusses new technology that is aiming to totally eradicate the problem of language as one of the primary factors limiting small businesses from operating abroad. Powered by big data, AI and deep learning, Neural Machine Translation (NMT) advances from previous models that translated words one at a time, to the more human-like method of reading sentences for context and meaning. Ken is responsible for defining and implementing the growth strategy for SYSTRAN who have a global presence in the US, Europe and Asia. With 20 years’ experience in the language intelligence industry, he is considered a thought leader in language translation having held several Senior Executive roles within the industry. A native of Ireland he is also a serial entrepreneur as well as business mentor to several Irish Start-ups.
“귀걸이가 맘에들어요. 친구들에게 추천하겠습니다.” As an online seller, is this good news or bad? In a survey commissioned by Education First, 49% of executives admitted that language barriers and communication difficulties had prevented significant international business deals from being done. In the same way that language barriers hamper big business, they hamper small business.
There are three ways to increase revenue: acquire more customers, increase the average spend per customer, and increase the number of transactions per customer.
Here’s the question for eCommerce platforms with millions of creative, ambitious, well-intended people looking to increase their income: can your seller in France close a deal in Japan? Can your seller in China provide customer support to his buyers in Spain? Can your users see reviews in their native language, no matter what the source language was?
Historically, these capabilities have been reserved for the captains of industry. eBay has been using their proprietary Machine Translation for years. But machine learning is making it available to everyone.
Technology firms have been working on the language problem for decades, but the last several years have seen significant advances that merit the attention of business leaders who are eyeing international markets. The technology is called Neural Machine Translation (NMT), a deep learning system that captures meaning in the context of translated sentences, not the single word. The net result is fluency, where previously, only a “gist” was possible. By combining NMT with existing big data tools that scrape, structure and analyze, new value propositions have suddenly become more attainable.
For example, let’s take a case from a peer-to-peer e-commerce site where a Korean customer inquires of an American Vendor, “미국에서 귀걸이를 한국으로 배송하면 얼마나 걸릴까요?”
When translated with NMT the result is “How long will it take to get earrings from America to Korea?” where as a statistical engine will return “In the United States, how long will it take to deliver the Korean earrings,” making commerce far more difficult to transact and probably several emails to clarify.
Similarly, reviews on web sites are extremely important with 88% of people saying that they now incorporate reviews as part of their buying process. In an “English only” world this tends not to be a challenge but, as HBR reported, 72.4% of people said they would be more likely to buy a product with information in their own language. NMT makes this possible today and with annual E-commerce revenue growth in double digits, e-tailers have a tremendous opportunity to accelerate revenue growth with minimal investment.
Customer self-service has also exploded over the last few years with many companies relying on “super users” to solve their clients’ problems. Again those outside the top 4 global languages find themselves in the dark most of the time. By implementing NMT solutions, companies are not only benefiting from happier customers in languages they struggled with, but also fend off local “copycat” technologies.
Of the three revenue growth options mentioned at the top of this piece, increasing spend per transaction and increasing number of transactions per customer are the easiest of the three. At the opposite end, losing a customer is a heavier cost then all. We live in a world where if two tech companies are created equal – the user will buy from the one with the better experience or the more relatable values. Offering cross-language chat, multilanguage reviews and knowledge based would put you in the upper echelon of competitors.
By the way, this ‘reads’ this. It was good news. This text means this: “I like the earrings. I’ll recommend them to friends.”
This article written by Ken Behan was originally published in inside BIGDATA. How Neural Machine Translation Will Help Online Marketplaces Turn their Individual Sellers into Global Players.
This article was originally published in Entrepreneur Magazine.
The ability to translate between languages quickly and inexpensively opens vast new possibilities.
Behind-the-scenes technology is not usually the sexy stuff that makes big headlines. Unless you are the IT guy in the back room, this kind of difficult-to-explain stuff is not the leading topic of discussion at your dinner party. Neural Machine Translation (NMT) is different. Few things on the horizon currently have as much importance or as much appeal. What it does behind the scenes changes the face of the whole economy.
In short, NMT is a deep learning technology that translates within context, not just one word at a time. Recent advancements have made this approach nearly fluent, making previous iterations of machine translation irrelevant overnight. The usual language translation heavyweights, Google and SYSTRAN, are pioneering this technology and making it available to different segments of the market.
So what does this have to do with business? In short, everything. Here are four ways NMT will impact the market.
1. Small businesses with global reach.
Small businesses are the driving force of the U.S. economy. According to the Small Business Administration, these companies employ 99.7 percent of America’s workforce. Their impact on the economy is far ranging, from innovative products to essential services. But they are also limited in reach, typically restricted by small operational budgets. That means they don’t frequently sell to international markets, and especially not to foreign language economies.
“After the internet arrived, we started hearing the term ‘global economy,’” says Denis Gachot, CEO at SYSTRAN. “It implies an ability to communicate, connect and transact with anyone in the world. But most don’t have that ability because, despite the internet removing geographic barriers, there is still very much a communication barrier in language.”
A large corporation can hire multilingual professionals to run remote offices and provide customer service in numerous languages. That is a luxury most small businesses cannot afford. NMT allows these businesses to immediately translate their web pages and online communications into more than 100 languages. “Neural Machine Translation is going to change the economy by giving more businesses a language capability they can use to communicate and understand in real time,” says Gachot.
That means that the shop owner in Milwaukee, Wisconsin, can market her products to people in Germany, Japan, Brazil and dozens of other countries.
2. Automatic translation of thousands of documents.
But NMT is not just an opportunity for growth for small business. Larger corporations stand to benefit from the quick processing capabilities of NMT as well. Company documents can quickly be translated into multiple languages with reliable accuracy and precision.
Previously, that kind of work would have required a team of highly skilled linguists and would have taken weeks to translate the original and check the resulting copy. But NMT changes that. The open network set-up of NMT technology allows for “soft alignment,” which means the system can search for the context of phrases and sentences instead of translating word by word. The reliability of this kind of machine translation, and the speed in which it is accomplished, can dramatically change the way companies are able to operate and ultimately serve their clientele all over the world.
3. Radically changes specific industries.
A change in translation technology means a huge change in specific industries. For example, legal eDiscovery can be extremely complicated for legal teams trying to access emails, chats and online communications in other languages. Each communication has to be carefully assessed for meaning and intent within the context of colloquial uses of the language and varying forms of slang. This is a nightmarish recipe for anyone working on such a case.
NMT changes this by rapidly learning terminology nuances and then producing high-quality translations at a fraction of the time it takes a human team to do the same work.
By using our own brain as a model, this technology is able to apply human intuition at machine speed. “Techniques for understanding slang include custom dictionaries and custom translation engines,” Gachot explains. “These engines are trained from hundreds of thousands of pieces of human translated content and are able to mimic the fluidity of expressions found in those human translations, including when there is colloquialism involved. We also have custom dictionaries for Information Technology (IT), economics, tourism, dialog and so on.”
4. Opens up isolated areas to the global market.
With the expansion of small businesses in global trade and more accessibility to unique products from a greater range of places, previously unreachable geographical markets will open up.
Right now, many countries are left out of the global marketplace because small businesses have no way of marketing to them or handling transactions across language barriers. Consumers in emerging markets may be the first to see the impact, but more remote countries will feel the effects soon after. Online marketplaces like Etsy and Zazzle can make language translation automatic, allowing users to conduct business in their own language. This reduces the friction in global commerce and increases the opportunities that remote consumers have to products and services around the world.
Business owners in remote corners of the world will now have a grander stage for their products, larger companies will be able to better care for their clients, and service industries will evolve to meet the changing tides. That is a complex technology worth talking about at your dinner party.
This article was originally published in Entrepreneur Magazine. This New Translation Tech Will Smash the Language Barrier to Doing Business Globally.
Don’t let language be a hurdle in your business
Lean Manufacturing involves constant efforts to eliminate or reduce ‘muda’ (Japanese term defining waste or any activity that consumes resources without adding value) in design, manufacturing, distribution and customer service processes.
As an operational system, Lean Manufacturing maximizes added value, reduces essential support and eliminates waste in all processes throughout the value chain. Waste in this regard may include over-production, inventory tasks, waiting time, correction, transportation and over-processing.
In summary, the equation for Lean Management is: Increased profitability equals increasing prices or reducing costs. A big part of the cost is the turnaround time between an order being placed and when it is shipped.
What is your Manufacturing Lead Time from the placement of an order to shipping? What would you expect as the standard for your business? What do you foresee as a likely evolution for both you and your customers?
The world has become a global village: electronic commerce is truly international and must be standardized
The 2nd Logistics Information Standardization Forum held in Seoul, South Korea on September 2016, brought together a host of actors in an effort to standardize international logistics information as a key factor in improving logistic processes in business. The forum put forth the creation of an international consensus for a cooperation system on international logistics information.
At the AFNET association, we promote and develop such standards so as to improve relationships between organizations and enterprises. Standardizing electronic commerce for logistics means defining a common document structure for order processing transactions and product deliveries.
Round-trip translation (RTT), also known as back-and-forth translation, recursive translation and bi-directional translation, is the process of translating a word, phrase or text into another language (forward translation), then translating the result back into the original language (back translation), using machine translation (MT) software.
It is often used by laypeople to evaluate a machine translation system, or to test whether a text is suitable for MT when they are unfamiliar with the target language. Because the resulting text can often differ substantially from the original, RTT can also be a source of entertainment*.
When we translate the paragraph below…
Following the publication of the Basel (II and III) and Solvency Regulations, implementing Governance, Risk and Compliance (GRC) practices within financial institutions has gained predominance. To meet the challenges of data governance stated in the pillars of these regulations, companies have multiplied their efforts to recruit the best in GRC concerning Risk Management, Internal control, Internal Audit and Compliance.
According to a study conducted by the consulting company Optimind Winter and mandated by the Observatoire des métiers de la Banque (Banking Career Observatory), “Banking companies are also victims of new risks and must create new job profiles to deal with these new challenges. GRC departments must face new challenges, such as coverage of systemic risk, development of Cloud Computing or Mobile Bank (technological nomadism).” Here, we refer to digitalizing the banking sector and using the Internet to manage personal bank accounts. Banks and insurance companies have thus implemented solutions to protect their customers’ data against cyber-attacks. While the efforts made to secure customer data have proven effective, the threat now lies within financial companies in their daily workflow.
Along with new working methods, a new phenomenon has emerged, known as “Shadow IT,” which is any application or method of transmitting information used in a business process without the endorsement of the internal IS department. Often unaware of its existence, IT departments don’t provide any support. Such processes generate "informal" and non-controlled data that can contravene existing standards and regulations such as Basel and Solvency. Continue reading