<The state of AI today. If anyone wants to see what it's like, check out DALL-E>
Let's not go to this nonsense as an example of the State of AI today. That's like buying a $2000 pickup truck from China, then adding a state-of-the-art $189 AMR500 Roots Compressor Supercharger from Wish so that you can haul bees REALLY FAST.
A.I. is used in a lot of things these days, and when properly constructed and applied, it can do some amazing things. You might be surprised at how many devices and services rely on some form of Artificial Intelligence. In the case of our acoustic app, the appropriate comparison is not to some entertainment toy, but to Dragon Speaking Naturally and its successors, such as Alexa, Siri, Google Assistant, and Cortana, all of which were introduced starting around 2011. These programs, as does our app, use known speech (or bee sounds) to train their A.I.s. Have 100 people say Hello in American English, and the program will begin to recognize Hello. Have 10,000 people say Hello, and it will get better. Have a few million use it, and it really begins to shine.
Training for speech recognition is not a quick process. When I first was at the Batelle Pacific NW labs and in our own UM Computer labs in the 1980s, programmers were trying to look at speech sonograms to figure out how to develop speech recognition programs. In 1975, Dr. James Baker laid out a speech understanding program called DRAGON. It wasn't until 1997 that the first Dragon System's Naturally Speaking was released as their first dictation produce. I tried it, and it was so-so. Dragon kept after this, and it slowly got better. Dragon also had some setbacks when it was bought out in 2000 by Lernout & Hauspie, who went bankrupt with charges of perpetuating financial fraud.
The rights to Dragon products were acquired by ScanSoft and rebranded as Nuance in 2005. In 2012, LG Smart TVs adopted the Dragon voice recognition feature for its Smart Tvs. In 2022, Microsoft acquired Nuance. In the meantime, despite the financial setbacks, Dragon's speech recognition advanced to the point where it allowed users to input their regional speech patterns.
These days, many of us use a Smartphone or a Smart Home device that uses one of the following Voice Assistants: Apple Siri, 2011; Google Now, 2012; Microsoft Cortana, 2014; Amazon Alexa, 2014; Google Assistant, 2016. They are not perfect, but the difference between now and when each first came out, and between now and 1997, when Dragon released the first Dictation Product is remarkable. I have an Alexa system that wirelessly connects every room in my two-story house, displays our multiple outside surveillance cameras (I've got a stalker - lucky me), ties in notifications from the cameras and my CO2/Fire alarms on each floor, and controls a bank of 6 picture window and patio slider window shades, turns on and off and dims lights, as well as distributes any music I want to any or all rooms or groups of rooms. Does it work - yes, so much so that we don't mention the Wake-Up Word Alexa other than when we want to use Alexa. We have to call her A, or else she'll go into an electronic tailspin trying to figure out what we want.
The best thing about Alexa, as we get deeper into our senior years, we can set up a system so that another party can check in on us, or if we fall we can call for help from wherever we are in the house or near any of the outside cameras. We use small, voice-activated pucks in most rooms, but have an Echo Show 10 in key rooms like the Kitchen. It has a slider that manually blocks the camera for absolute video privacy. When the slider is open, the Alex word prompt will cause the show to pivot, looking for the speaker. Do we give up some privacy - probably? We do get to pick and choose what to activate in any part of the house. Does it provide some security and safety for us as senior citizens - absolutely!
All of this ties into my Samsung Smart Phone. I can check on the house, family, etc. from anywhere in the world where I have cell service - all by voice. The biggest downside of Alexa, it offers up some of the lamest jokes I've ever heard. On Upside, I can listen to any genre of music and thousands of different artists and bands in any room, just by asking Alexa to play something - Alexa play the Blues, or Alexa, play Jazz, or Alexa play ..Name of Performer. Piped into my Classic AR speakers, sub-woofer, and Yamaha Amp. I'm set for the day in my home office.
As per my nut-case stalker, all of the camera streams go to the Cloud and are stored for a month. The moment a person appears near any camera, I get a notification on my PC, on my cell phone. Here again, an AI system screens the images and sounds and classifies them as Motion, Animal, and Person. The sensitivity is such that small birds and large insects like moths show up on the video and are classified as animals. We get a lot of hits on neighborhood cats, raccoons, deer, and recently a beaver. Plus the mail, UPS, and Fed Ex delivery persons. The cameras work on hot summer days and sub-zero winter days and nights. They can spot a domestic cat 30 feet away from the camera. The cameras have night vision. Resolution is really good. I can read the pocket names on the uniforms of delivery and service people. Cameras are mounted high and overlap fields of view. This is not a grainy, can barely tell ts a person, system. All of this motion-triggered video, sound, and motion information is sorted, stored in the cloud, and appropriate notifications are sent, much of which uses AI discrimination. If we want to talk about the state of AI today, let's at least be up-to-date.
Jerry
P.S. James has told us how to properly set up a much more secure system, but he hasn't volunteered to come out to MT to set one up for me. In the meantime, I'll stick with Alexa. I can always unplug it. And I still use physical deal bolt locks.
***********************************************
The BEE-L mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software. For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html
|
