Learning ROILA. Image from roila.org.
Science fiction is full of examples of humans and machines interacting verbally, but until recently any attempts at voice recognition have been met with limited success. Even today the level of success attained by advanced tools like Siri is up for debate. If you’re lucky enough to speak English without an accent and can avoid using an expansive vocabulary, then Siri may work well enough for you, but a quick search on YouTube will give plenty of examples of Siri falling squarely on her face. Maybe Google’s upcoming project, code-named Majel, will do better, but I’m not going to bet any money on it.
One team in The Netherlands is working to meet the machines half way by adopting what could be dubbed Esperanto for robots. The language is named ROILA and it is an attempt to provide a simple spoken language that is both easy to learn, and easy for computers to interpret. Think of something along the lines of Palm’s Graffiti for the Palm Pilots, but updated for the spoken world.
They summed up the process on how ROILA was developed on their website:
We reviewed the most successful artificial and natural languages across the dimensions of morphology and phonology (see overview in the form of a large table) and composed a language that is extremely easy to learn. The simple grammar has no irregularities and the words are composed of phonemes that are shared amongst the majority of natural languages. The set of major phonemes was generated from the overview of natural languages. Moreover, we composed a genetic algorithm that generated ROILA’s words in a way that they are easy to pronounce. The same algorithm makes sure that the words in the dictionary sound as different from each other as [much] possible. This helps the speech recognizer to accurately understand the human speaker.
For more information check out roila.org or snag the book Learning ROILA.
Unfortunately none of my current robots have the processing power to understand ROILA, but it is definitely something that I will try implementing in the near future.
