Say Hello, Hola, Bonjour to the Robots
September 16, 2021
By Ofer Tirosh
Will we live and work with robots in the future? It's an odd question to ask really, as robots and artificial intelligence (AI) are a big part of our daily lives already! We are very familiar with big industrial robots dancing in synchronized formation in assembly plants, but the near future offers us much more human experience.
If you’ve ever asked your mobile app for directions or ordered a package via Amazon, or tried to get to speak to a doctor on the phone, well, you’ve been talking to a machine. Not so much an R2D2 whooping and whirring, but a disembodied virtual assistant with voice search software intelligently scanning the internet.
Maybe the better question is how will robots learn to work with us in the future? Humans are a lot more complex and diverse when it comes to language, culture, abilities, and personalities. Even we get confused trying to deal with each other.
Think Alexa will wake up and order your shopping in Mandarin? Hmm not yet.
Think Google Assistant has all the answers, even in Swahili? Think again.
In fact, Alexa supports 8 languages as of 2021. Alexa also supports dialects, in English, French, and Spanish.
- English dialects (5): Australia, Canada, India, UK, US
- French dialects (2): Canada and France
- Spanish dialects (3): Spain, Mexico, US
As of April, 2021 Google Assistant supports the following languages:
Apple’s Siri now supports 21 languages. In 2020, they introduced iOS 14, which offered translations in real-time for 11 languages and even worked offline. However, Hindi wasn’t one of them, and it is the third most spoken language in the world today.
So despite the progress in voice assistants over the last decade, we still have a long way to go before we have that Star Trek-style Universal Translator that can understand us like a native, and respond in kind to our voice, needs, and moods.
Robots to Look After the Family
If you are reading this and excited about robots in your home, then you probably can’t wait for the robot trend to hit your local stores.
You don’t have to wait much longer for the robot revolution. 2022 is set to be the year of the ‘bots.
Thanks to Starship Technologies we will soon see local delivery robots buzzing around, designed to quickly deliver parcels and post within a two-mile radius. This self-driving delivery bot is a ground-based alternative to delivery drones.
Some of us are already familiar with the Tesla self-navigating cars that drive themselves, automatically perform perfect parallel parking, and seek your favorite radio station with only a few simple voice commands and the power of neural networking. Tesla intends to ignore the warnings of popular science fiction and introduce humanoid Tesla Bots next year.
ElliQ is described as an "active aging companion" robot.
The ElliQ is a two-device package that includes an Android screen and an Alexa-like digital assistant bot. It is designed to be a companion that sits on a desk but can swivel around, speak verbal notifications, respond to messages, set reminders for meds, answer video calls, and monitor activities.
ROYBI the robot provides kids with a self-guided tutoring experience that uses AI and the power of up-to-date virtual learning resources. ROYBI robot introduces children to technology, math, science, and multiple languages including English, Mandarin, and Spanish.
Natural Language Understanding vs Human Translation Services
Now we understand that what you say, and how you say it, is at the core of the technological cold war for voice recognition dominance in robotics, the question becomes why aren’t more companies able to develop and localize machine-translation services?
The best answer to that, as it happens, came in 2017 from Google itself.
Even a sophisticated AI that listens, recognizes, and transcribes words with accuracy, will still have a processor-intensive grind of localization to go through, known as natural language processing (NLP) —so, Spanish may be Spanish in Spain, but not quite the same in Mexico, or the United States. NLP works to resolve each dialect and all the cultural differences that require distinct linguistic models.
Google’s answer is natural language understanding (NLU). This is why Google may well soon take the lead in multilingual voice search: it has already put in the work of creating an enormous backlog of more than 115 languages capable of automatic speech recognition (ASR) through Google Assistant.
ASR and localization of Google Home can help reduce the processing load on Google Assistant, as it only needs to build on the intent of the transcribed sentence.
NLU is where the intent behind our words is found, however; this speech data needs to be tailored to the local use of words, grammar, and colloquial expressions. Correct transcription and annotation are essential blocks in building an accurate product with reliable results.
Tomedes is one of only 3 translation providers partnering with Google on localizing apps in dozens of common, and rare, languages, and understands the challenges of NLP, and the importance of the human dimension to robots comprehending and interpreting our speech.
“Voice recognition, virtual assistants, and voice search are rapidly growing and evolving fields,” said Maya Ronen, COO of Tomedes. “Adding additional languages to virtual assistant applications multiplies the project complexity, and the data being generated, by an order of magnitude. It takes real human expertise to understand barriers to language, and to assure the quality of these interactions in a way we can all understand: human or robot.”
So don’t be shy to say “hola, bonjour, or ni hao” to the car, tablet, tv, smartwatch, or refrigerator -they may just greet you back!
Why choose us
24/7 Human Support
1 Year Guarantee
95,000 Business Customers