An engineer and oceanographer, Derya Akkaynak, created an algorithm that removes the water from underwater images - meaning it takes away the haze & tint that come with most underwater photos. It doesn't require a color chart either (though it does need distance information it gathers through numerous photographs from different angles).
The Sea-thru method estimates backscatter using the dark pixels and their known range information. Then, it uses an estimate of the spatially varying illuminant to obtain the range-dependent attenuation coefficient. Using more than 1,100 images from two optically different water bodies, which we make available, we show that our method with the revised model outperforms those using the atmospheric model. Consistent removal of water will open up large underwater datasets to powerful computer vision and machine learning algorithms, creating exciting opportunities for the future of underwater exploration and conservation.
Essentially, the algorithm goes through every pixel and color balances the image based on collected data on distance/color degradation.
As with most algorithms of this type - the more data we feed it, the better it gets. Sea-thru already has great applications for furthering ocean-based research, but the follow-up question is can this algorithm be extrapolated outward to deal with other atmospheric conditions outside of water - smog, etc.
What other uses can you imagine? It is not hard to imagine how this could be applied to market data either ... Interesting stuff!
AI has plenty of weaknesses - I've talked about some before, and I'll continue to talk about them in the future, but two specific weaknesses were brought to my attention this week.
AI Portraits - Won't Steal Your Data, But Might Steal Your Soul Dorian Gray-Style
I assume most of you have seen the FaceApp trend - people age-ifying their photos and unwittingly giving the rights to their photos to a shadowy Russian tech company. You've also likely seen AI paintings selling for ridiculous money.
But have you seen their lovechild AI Portraits - a more wholesome experiment run by the MIT-IBM Watson AI Lab. AI Portraits uses approximately 45,000 different Renaissance-esque 15th-century portraits and General Adversarial Networks to translate your selfie into an artistic masterpiece. It's novel because instead of simply drawing over your face it's generating new features and creating an entirely new version of your face.
Mauro Martino via YouTube
It's impressive because it determines the best style for your portrait based on your features, your background, and more.
However, it's not without "flaw". The choice of 15th-century portraiture creates a couple of clear biases. At the time, portraits of smiling or laughing individuals were rare, so your smile will likely not transfer. As well, there's a clear bias towards anglo-saxonification.
My son got excited while playing with the app and sent several of his coworkers, friends, and family through the app. If you look at the bottom right, you'll see my lovely wife Jen's portrait.
Most of you have seen my wife and know that she is Indonesian, something that is very much removed from the translation.
All photos are immediately deleted from their servers after creating your image, so your privacy is safe (this time!)
All biases can be considered quirks of this current iteration of the program - which I do earnestly believe is interesting.
Later, you can imagine an AI choosing between various different styles of art based on a cornicopia of factors - or off human selection - but you have to walk before you can run, and this is a fun way to get people excited about AI.
Computer Answering Systems - No, The Answer Isn't 42
“Yes…Life, the Universe, and Everything. There is an answer. But I’ll have to think about it...the program will take me seven-and-a-half million years to run.” - Deep Thought, Hitchhiker's Guide To The Galaxy
Think of the global excitement when IBM's Watson first beat Ken Jennings in Jeopardy ... it's widely considered one of the holy grails of AI research to create a machine that truly understands the nuances of language and human thought. Yet, if you've talked to Alexa recently, you know there's a long way to go.
Today's question answering systems are basically glorified document retrieval systems. They scan text for related words and send you the most relevant options. Researchers at the University of Maryland recently figured out how to easily create questions that stump AI (without being paradoxical, impossible to answer, requiring empathy etc.) in order to enhance those systems.
A system that understands those questions will be a massive step toward a real understanding and processing of language.
So what's the secret to these "impossible" questions?
The questions revealed six different language phenomena that consistently stump computers. These six phenomena fall into two categories. In the first category are linguistic phenomena: paraphrasing (such as saying “leap from a precipice” instead of “jump from a cliff”), distracting language or unexpected contexts (such as a reference to a political figure appearing in a clue about something unrelated to politics). The second category includes reasoning skills: clues that require logic and calculation, mental triangulation of elements in a question, or putting together multiple steps to form a conclusion [...]
For example, if the author writes “What composer's Variations on a Theme by Haydn was inspired by Karl Ferdinand Pohl?” and the system correctly answers “Johannes Brahms,” the interface highlights the words “Ferdinand Pohl” to show that this phrase led it to the answer. Using that information, the author can edit the question to make it more difficult for the computer without altering the question’s meaning. In this example, the author replaced the name of the man who inspired Brahms, “Karl Ferdinand Pohl,” with a description of his job, “the archivist of the Vienna Musikverein,” and the computer was unable to answer correctly. However, expert human quiz game players could still easily answer the edited question correctly.
The main change is increasing the complexity of the questions by nestling another question. In the above example, the second question forces the AI not only to decide the composer inspired by Karl Ferdinand Pohl, but also to decipher who is inspiring (hint: It's Karl Ferdinand Pohl).
AI isn't great yet at mental triangulation; at putting together multiple steps to form a conclusion. While AI is great at brute force applications - we're still coding the elegance.
June 6th, 1944 ... the day we stormed the beaches of Normandy. It was the largest seaborne invasion in history. The invasion created a foothold that allowed Allied forces to expand through France.
The memorial in France this year was a particularly touching event with the many surviving World War 2 soldiers being well into their 90s.
"A generation whose unconquerable spirit shaped the post-war world. They didn't boast. They didn't fuss. They served" - Theresa May
A 97-year old veteran paratrooper, Tom Rice, commemorated the day by once again dropping into the field he landed in on that fateful night.
Seems like a friendlier welcome than a sea of German soldiers. As an extra fun fact, Teddy Roosevelt stormed the beaches of Normandy at 56 with a heart condition, arthritis, and a cane. He was the oldest man to take part in the invasion.
Our society is built on the backs of many strong men! I'm thankful for their sacrifices.
I came across a post from 2009 about social media. 10 years later, with the knowledge of how much data we use today, it's quite a read.
Here it is in its full glory.
Social Media Is Changing Everything: October 18th, 2009
My son won't use e-mail the way I did. So how will people communicate and collaborate in the next wave of communications?
Here is a peek into the difference that is taking hold. I was looking at recent phone use. The numbers you are about to see are from the first 20 days of our current billing cycle.
My wife, Jennifer, has used 21 text messages and 38 MB of data.
I have used 120 text messages and 29 MB of data.
My son, at college, used 420 text messages, and is on a WiFi campus so doesn't use 3G data.
My son, in high school, used 5,798 text messages and 472 MB of data.
How can that be? That level of emotional sluttiness makes porn seem downright wholesome.
But, of course, that isn't how he sees it. He is holding many conversations at once. Some are social; some are about the logistics of who, what, when, where and why ... some are even about homework. Yet, most don't use full sentences, let alone paragraphs. There is near instant gratification. And, the next generation of business people will consider this normal.
Is social media a fad? Or is it the biggest shift since the Industrial Revolution?
Welcome to the World of Socialnomics. This video has a bunch of interesting statistics ... and is fun to watch.
Social Media Is Changing Everything: April 20th, 2019
Looking at the stats from 2009 is pretty funny
My son was using 472 MB of data a month
Hulu had grown from to 373 million total streams in April 2009
Only 25% of Americans in the past month said they watched a short video on their phone
For some context, I looked up the comparative numbers for 2018.
I picked a random month in 2018 ... in August my son used 10.85 GB of data. He doesn't text as often - but has sent/received 282,000 snapchats since downloading it 5 years ago.
Hulu has over 20 million subscribers who streamed more than 26 million hours a day in 2018
People spend over five hours a day on their smartphones on average. 70% of web traffic happens on a mobile device, and more than 50% of videos are watched on mobile (93% of twitter videos).
Here's what happens every minute of every day on the internet
One thing that Deep Learning excels in is analyzing pictures & videos, and creating facsimiles or combining styles. If you want to create art with deep learning look no further than the Deep Dream Generator or deepart.io which use Convolutional Neural Networks to combine your photo with an art style (if you want to do it on your phone another cool tool to check out is Prisma).
Deepfake is it's exactly what it sounds like ... the use "Deep Learning" to "Fake" a recording. For example, a machine learning technique called a Generative Adversarial Network can be used to superimpose images onto a source video. That is how they made this fun (and disturbing) Deepfake of Jennifer Lawrence and Steve Buscemi.
While this is a fun example, Deepfakes create very real concerns. They're often used for more "nefarious" purposes (e.g., to create fake celebrity or revenge porn and to otherwise make important figures say things they never said). It's likely you've seen videos of Trump or Obama created with this technology. But it is easy to imagine someone faking evidence used at trial, trying to influence business transactions, or using this to support or slander causes in the media.
Christmas Eve saw Dow's historic low ... but it was immediately followed with a 1086 point rally (its largest single-day gain). The S&P and Dow both rallied approximately 5%.
Here are some of the posts that caught my eye recently. Hope you find something interesting.
This week I went to Chicago to speak at my old alma mater - Kellogg.
It was great to talk with professors and spend time with MBA students interested in trading and how the business of trading is changing.
I plan on spending more time there to benefit research and recruiting.
Unsurprisingly, deep dish pizza is still delicious. We were going to eat at a steakhouse. However, Jennifer and I remembered that the only thing better than gorging on meat ... is gorging on meat and cheese.
While some things haven't changed, others have.
The Universe must be trying to tell me something. The day I got back, I found my old ID card ... and, apparently, some things have changed in the past thirty years.
How tempting is it to photoshop (or embellish) a little? In a recent survey of 2000 British people, more than 75% admitted to lying about themselves on social profiles.
Here is a chart ranking the most common topics people are most dishonest about.
Social Media Is Changing Everything: A Reprisal
I came across a post from 2009 about social media. 10 years later, with the knowledge of how much data we use today, it's quite a read.
Here it is in its full glory.
Social Media Is Changing Everything: October 18th, 2009
My son won't use e-mail the way I did. So how will people communicate and collaborate in the next wave of communications?
Here is a peek into the difference that is taking hold. I was looking at recent phone use. The numbers you are about to see are from the first 20 days of our current billing cycle.
How can that be? That level of emotional sluttiness makes porn seem downright wholesome.
But, of course, that isn't how he sees it. He is holding many conversations at once. Some are social; some are about the logistics of who, what, when, where and why ... some are even about homework. Yet, most don't use full sentences, let alone paragraphs. There is near instant gratification. And, the next generation of business people will consider this normal.
Is social media a fad? Or is it the biggest shift since the Industrial Revolution?
Welcome to the World of Socialnomics. This video has a bunch of interesting statistics ... and is fun to watch.
Other Resources:
Social Media Is Changing Everything: April 20th, 2019
Looking at the stats from 2009 is pretty funny
For some context, I looked up the comparative numbers for 2018.
Here's what happens every minute of every day on the internet
via Lori Lewis and Chadd Callahan
A little different than 2009 ...
Posted at 01:37 PM in Business, Current Affairs, Gadgets, Games, Healthy Lifestyle, Ideas, Just for Fun, Market Commentary, Movies, Music, Personal Development, Pictures, Science, Sports, Web/Tech | Permalink | Comments (0)
Reblog (0)