AI: What comes into your mind? Robots, Siri, maybe autopilot cars? Once we give it a bit of thought, we realize we have a very fixed idea of what AI is (I’m not blaming all those TV shows and books we went through as children, not at all). Machine Learning is a term just as obscure, if not more. But there’s more to AI and machine learning than we think.
AI is present in our everyday lives, everywhere. As of 2017, with 25% of companies investing at least 15% of their IT budget on this sector, it better be. Sometimes, it’s what we expect, other times, it’s with a subtle intricacy which is the definition of unexpected. Either way, these are a few surprises up the sleeves of this (mechanical) person.
Predictions and recommendations
This is one normal and expected. Our online shopping sprees are often littered with “you may also like”s and “customers who bought this also bought”s. Sites like Amazon.com use artificial neural networks to generate these product recommendations. Amazon’s technology is called DSSTNE, and they recently made it open source, much to the delight of every tech tinkerer around.
Their system is quite different to the ones Google, Facebook and Microsoft use, which focus primarily on image and speech recognition. In particular, Facebook’s 7 layer deep neural network deserves a mention here. This is because it provides facial recognition with a 95% accuracy rate.
In addition, Google’s new smart reply function (which you may have noticed or used) also employs machine learning to determine the contents of an email and suggest replies.
PickMe is robot driven?
Everyone calm down, no you’re not being driven around by humanoid robots. However, for transport services like PickMe and Uber, machine learning is a fundamental part of the system. This is used to determine arrival times, fare details, and your locations (try getting picked up from the same place a few times).
These systems rely on data derived from scores of previous trips to provide you with accurate times regarding your journey. Uber states that this analysis has improved their estimations by 26%. Of course, this means higher customer satisfaction rates as well.
News Generation with AI
AI being integrated into the media world is again, something expected. Speeding up research and cross referencing data, it helps media personnel on a whole new level. However, that’s not all it can do. Nowadays, it can even take over writing articles! (Don’t worry none of us are going to lose our jobs.)
While AI isn’t at a level to mimic the levels of human creativity, the everyday mundane tasks are handled by it easily. This includes financial reports and sports recaps among others. The technology still requires the use of a human brain to figure out what data you need and what changes in said data you observe. However, the tedious task of writing it out is now out of our hands.
The age when we’re too slow to take notes
Yes, we’ve reached the era where we’re either too lazy or slow to take our own notes. Meetings and conference calls in particular, are challenging to many. Fortunately for us, several applications exist to help us get rid of the challenging job of translating audio conversations in text.
The first of these, Clarke.ai records and transcripts conference calls. Clarke the bot dials into your conference calls and silently takes notes for you while you concentrate on what is being said. Once the call is over, Clarke also mails the notes to your inbox.
On the other hand, Otter, is something we’ve recently tried out. Otter uses AI smarts for automatic speech recognition, speaker separation and identification as well as deep content search. Essentially what it does is record a conversation (even between several people) and transcribe it into text.
Making up stories just got a whole new meaning
Generations upon generations upon generations later, we Sri Lankans have pretty much heard every story our ancestors could possibly dream of (let’s be honest, one could only dream of going to heaven hanging on an elephant tail). Other countries are done documenting all the folklore they ever knew. So what’s new?
What’s new is Sheldon County. A podcast which uses AI to generate a story from random characters, settings and relationships. A user accessing this podcast will do so by typing a random number each time. This number is used to gather a few random characters and plot points from a large database, which is then made into a story. Each time the story is a unique one.
Beethoven just woke up to see the new AI composer
Usually, nowadays we end up keeping separate playlists on our phone or podcast for various moods. But is that necessary? Well, with the introduction of brain.fm, its not. Brain fm is an AI composer, with the specialty being that it can compose music for various moods you might be in. These include focus, relaxation, meditation, sleep and nap. Whichever you decide, you will get music composed just for that purpose.
On the other, not so light end of the spectrum, we have AI made black metal albums. “Coditany of timeness” is an album created by two music technologists using a deep learning software. This ingests a music album, breaks it into segments and analyzes it, and spits out an imitation of the style.
No? Good! This was just brushing the surface of what AI does in our world. In an almost invisible way, it is impacting our lives daily. And it’s not always how you expect.
Know of any more intriguing AI applications? Let us know in the comments below!