Saturday, February 11, 2017

Babel Fish: The App that Lets You Talk to Animals!

“The Babel Fish is a small, leech-like, yellow fish, and by putting this into one's ear one can instantly understand anything said in any language... The Babel fish has led to significant and profound consequences for the Universe… The practical upshot of all this is that if you stick a Babel fish in your ear you can instantly understand anything said to you in any form of language."
Douglas Adams. Hitchhiker’s Guide to the Galaxy.


“For millions of years mankind lived just like the animals
Then something happened which unleashed the power of our imagination
We learned to talk”
Pink Floyd. Keep Talking. The Division Bell.


“Any sufficiently advanced technology is indistinguishable from magic.”
Arthur C. Clarke. 1917-2008.


Prediction: Humans will soon implement animal voice and gesture recognition algorithms based on existing research enabling us to communicate with many intelligent species such as apes, dolphins, whales, and dogs. "Google Translate for Animals" may have been a clever April Fools joke a few years ago but everything we need to implement it is, in fact, available today. Such a tool would have significant scientific benefits in our understanding of the animals around us as well as opportunities for monetization. Today, we can automatically translate between dozens of human languages each containing millions of words and phrases. We can use pattern matching to distinguish minute facial and hand gestures in humans. It should be a cinch to teach a computer to recognize a few dozen simplistic phrases per species.

Scientists have already shown that many animal species possess rudimentary language skills and are capable of vocalizing dozens of specific words and phrases. Here, for example, is the sound-frequency analysis of prairie dog alarm calls, differentiated according to predator (Con Slobodchikoff). I’m sure you’ve also seen documentaries about chimps that use sign language and dogs that recognize over a thousand words.




We spend billions of dollars a year (in human brain power as well as computing cycles) searching for potentially non-existent extra-terrestrials, hoping that they have advanced civilizations, and trying to communicate with them using radio technology across thousands of light years. Our chances of success are (no pun intended) astronomically low. What if we take just a fraction of that money and spend it on attempts to communicate more effectively with real inhabitants of this planet instead of (or in addition) to searching for our imaginary friends in the night sky?


We already have the technology to digitally recognize nuanced verbal cues in dozens of human languages, we have the technology to recognize tiny and complex facial expressions and hand gestures of humans. Are you telling me we can't recognize a few hundred barks and grunts and gestures that these animals routinely use to communicate with each other? Their primitive language is too hard for us to recognize? Communicating. back should be relatively easy as well, by replaying pre-recorded sound bites.


Who'd pay for it, you ask. The monetization opportunities are almost endless. Imagine if your dog could actually tell you what she wants instead of barking incessantly. How much would you pay that? How much would your annoyed neighbor pay for the same privilege?


We could also offer specialized solutions and packaging for various vertical markets. Veterinarians can finally diagnose pet diseases without guesswork. Farmers can upgrade their barns to the 21st century. SeaWorld can finally offer equal employment opportunities to dolphins and killer whales. Zoos can attach loud speakers to animal cages and crank up the volume. The possibilities are endless. Unfortunately, television personalities like The Dog Whisperer would be out of a job.


An obvious, but perhaps unintentional consequence, is that we can also let the animals talk to each other. Not only can we have “Google translate Dog:English” but also “Google translate Dog:Pig”. Have hours of fun with the whole family the next time you visit a farm. Imagine the implications for world inter-species peace if we could air drop a few thousand iPhones in the African Sahara that simply run “Google translate Lion:Gazelle”.


So much for my feeble attempts at humor. But, really, I shouldn't have to come up with ridiculous business propositions to convince you that this is a worthwhile effort. Imagine the contributions to primate research, to environmental science, to veterinary science. Imagine how many more puppies would get adopted at the shelter if only you could hear them first hand. Scientists have already done most of the heavy lifting, identifying hundreds of very specific barks and gestures in several commonly studied species. All we have to do is write a “Google translate” language adapter.


To be clear, some of this work is already happening in research communities: "A computer science colleague of mine and I are using artificial intelligence techniques to keep a computer record of the call that the prairie dogs were making, analyze it with these AI techniques, and then spit back the answer to us, which potentially could be in English. So the prairie dogs could say something like 'thin brown coyote approaching quickly.' And then we could tell the computer something that we wanted to convey to the prairie dogs. And the computer could then synthesize the sounds and play it back to the prairie dogs."

So, I haven't totally lost my mind. But we do need commercial applications and a more concerted effort to move the work from the research arena to the real world.

My brother-in-law had the last laugh when I told him about my latest (ahem) hare-brained idea. “Why don't you work on a Republican:Democrat translator instead?”, he asked. That, I’m afraid, would be an impossible task.

No comments:

Post a Comment