AXEL.org

artificial intelligence

This New Startup Helps Non-Techies Use Machine Learning

**This is part of our series highlighting startups who share our mission of trying to make people’s lives just a little easier**

Machine learning sounds pretty simple, in theory. For example, if you want to create a tool that identifies faces in a photo, you input tons of pictures of people into the machine learning tool and, after a while, the software will learn what a face looks like in a photo and how to identify it.

But it’s more complicated than it seems. There’s actually a whole load of custom code, different software, and advanced data analytics that go into the recipe for a successful piece of machine learning.

Enter Lobe, a new startup that aims to make machine learning as simple as clicking together a couple of LEGO bricks. It got its name from the platform’s drag-and-drop visual coding interface (known as “lobes”) which manifest as card-based views that can be moved around, put together, and built upon.

Let’s face it, machine learning doesn’t sound like a simple concept you could chat about with your granny over tea. It’s a pretty complex thing that has taken years to develop and even longer for people to master the art of it.

Which leads most people to thinking that it’s not for them. But if you’ve had ideas in the past (and who hasn’t?!), they could probably benefit from machine learning in some way.

To get things started on Lobe, you simply have to add a load of images or sound files into the website and the tool will start processing them immediately to learn what it can from the files you’ve uploaded.

The goal? For any old non-techie person to make their wildest ideas a reality.

Mike Matas, founder of Lobe, says: “There’s been a lot of situations where people have kind of thought about AI and have these cool ideas, but they can’t execute them. So those ideas just get shed, unless you have access to an AI team.”

According to a recent survey of 2,500 developers, 28% of respondents named AI and machine learning as the technologies they were backing the most in 2018. Even non-techies are somewhat enamored with the idea.

But for the tech-minded lot, there are plenty of tools already out there in the developer world; tools that require the inputting of code and the knowledge of different software languages to build modules upon modules. Basically, they are exclusive in that only people with the technical know-how can use them, whereas Lobe aims to offer an inclusive tool for everyone that has an idea.

“You need to know how to piece these things together, there are lots of things you need to download,” says Matas of the tools that are “exclusive” to developers. “I’m one of those people who if I have to do a lot of work, download a bunch of frameworks, I just give up. So as a UI designer I saw the opportunity to take something that’s really complicated and reframe it in a way that’s understandable.”

How Lobe Works

The tech industry has fully embraced AI and machine learning, welcoming it with open arms and metaphorically feeding it up like the guest of honor at a dinner party. In the survey mentioned above, 73% of respondents said they were interested in learning about machine learning platforms, despite only 17% having worked with AI technologies in 2017.

The team behind Lobe have tapped into this idea that people want machine learning technology, but still find it a pretty confusing concept.

By taking the complicated, code-riddled parts of machine learning (we’re talking feature extraction and labelling here) and turning them into a simple, easy-to-use visual interface, Lobe basically offers machine learning for dummies.

Take the example below. Lobe’s platform allows users to create applications that can read hand gestures in real photos and match them up with signs in emojis without having to go anywhere near a single piece of code.

This makes the possibilities endless: a new dimension has opened up where people can create and build their own apps without having the advanced technical knowledge to do so.

Despite the potential it has, Lobe is still pretty basic in its execution and interface.

It relies solely on images and sensors. For example, existing blueprints include the ability to identify plants from images and create a tuner for different string instruments.

“As a UI designer when I first started, I made everything in Photoshop,” says Matas. “Everything I designed was through a static interface. So every solution was a button on the screen. Then I learned UX prototyping tools. With them, we could do iterative prototyping, and suddenly we could solve problems with motions, interaction, and gestures. And you get things like the iPhone X home swipe.”

What This Means for the Future of AI

For Matas, AI is bringing the next influx of user interfaces with it.

He compares the early days of machine learning with those of PCs, when only computer scientists and engineers could operate the new machines: “they were the only people able to use them, so they were the only people able to come up with ideas about how to use them,” he says.

It wasn’t until later in the 80s when computers became more of a creative tool for all, and that was predominantly due to vast improvements in the user interface making them easier for non-techies to get behind.

The aim is to bring machine learning to the masses and break it free from the tech world in some way. “People outside the data science community are going to think about how to apply this to their field,” Matas says. Unlike before, where they needed an AI expert to stand in and help build the learning systems, they will now be able to create a working model themselves.

For a long time, AI was the future of the tech world, but with tools like Lobe popping up, it now seems that machine learning and deep learning capabilities are the future for all industries.

The AI-Based Startup That Knows Your Skin Better Than You Do

**This is part of our series highlighting startups who share our mission of trying to make people’s lives just a little easier**

Anti-aging. Miracle-cure. Dermatologically tested.

These are all buzzwords that regularly crop up in the beauty industry, driving women (and men) all around the world to spend thousands of dollars on skincare products that they may or may not need.

In fact, the global skincare industry is currently estimated to be worth around $130 billion by 2019 – that’s a lot of miracle cures. And we can only assume that the industry will snowball even further in this selfie-obsessed age.

People want to look good – fact. But with so many products to choose from, it can often be a quagmire of irrelevant creams and dissatisfying serums out there.

One new startup sees itself as the beauty industry’s knight in shining armor. Fed up with the BS spouted around skincare, Proven uses artificial intelligence to create individual skincare routines based on skin types and needs.

“The average person spends 45 minutes to 1.5 hours researching products before they buy any beauty products,” says Proven’s co-founder Ming Zhao. “And even after they buy based on the research that they’re able to do, 55% of people are still unsatisfied post-purchase.”

And, with all the lofty promises in ads and on billboards, is it any wonder that people are left disappointed when miracles don’t happen?

“No single person is capable of reading the amount of information there is in order to make a sound decision,” continues Zhao, going on to add that this was the reason behind building the largest database in the beauty industry. Using co-founder Amy Yuan’s computational physics background, the duo has put together an AI engine that sifts through reviews of skincare products.

How Proven Gets to Know Your Skin

To date, the engine has analyzed around eight million skincare product reviews, 20,000 ingredients, and 100,000 products using a sophisticated algorithm to eliminate any fake reviews.

When it sifts through this data, it can pick out general patterns and trends to determine what products are suitable for particular skin types.

“After trying numerous products and investing, nothing really worked for me,” admits Zhao. “Eventually what worked for me were customized products that were made for me by a few different facialists. So that’s how the optimize idea of tailoring products to exactly someone’s situation, someone’s skin, first came to my mind numerous years ago.”

And so Proven was born.

Now that the algorithm has large quantities of data in its clutches, the engine can be let loose on the public.

It works by encouraging people to take a skincare routine quiz that culminates in a categorization of what skin type you have before offering a selection of Proven’s custom skincare products (which are mixed on-site by a chemist).

Quiz-takers can then opt-in to Proven’s bundle of personalized goodies that costs $120 every two months.

Though the AI machine still needs input from each individual, like what products they can’t live without, whether they have an oily T-Zone, and skin allergies, it then uses the mammoth amounts of data is has sifted through to provide a customized solution.

Zhao and Yuan’s main idea was to tap into the power of deep learning algorithms to pick out useful, relevant information from millions of testimonials online and turn the overwhelming amounts of data into a formula to predict skincare routines that actually work – no miracles involved here, just cold, hard data and facts.

Why Use AI in Skincare?

“Why is the skincare industry not in great shape right now? Because everything is the same,” says Sue Y Nabi, founder of new skincare brand Orveda. “I’m fed up with ‘miracles’ and I don’t believe in focus groups – they’re good for telling you you’re not making mistakes, but they don’t give you the recipe for success.”

This is where AI swoops in and metaphorically saves the day (or saves the face of the whole industry – get it?).

While focus groups can’t provide a recipe or a formula for success, AI can. Or, at least, it has the capabilities to.

The traditional skincare model of business rests on focus groups, consumer research, and celebrity endorsements which, although are powerful in “speaking” to people who want to look good, don’t provide personalized information for each individual.

And, let’s face it, everyone is unique. No two people have the same skin, so it’s impossible to promote a product that is a one-size-fits-all. New startups like Proven are on a mission to disrupt these attitudes and outdated ways of researching and advertising in the skincare world.

What Proven Means for the Future of the Skincare Industry

Instead of helping women buy into the “dream” of miracle cures and skin that looks ten years younger, the co-founders of Proven are on a mission to bring a rational, logic-based approach to the skincare industry; an approach that systematically and sensibly figures out what ingredients are most appropriate for each individual.

But the main question that keeps cropping up around AI-generated beauty routines is whether the data scraped together can actually lead to useful machine-based decisions.

Again, the skincare industry is incredibly subjective and, while reviews might be a good way to determine if a product is for you on an individual level, large amounts of it might generate conflicting viewpoints.

So far, Proven has definitely had a good go at disrupting the current (and potentially archaic) narratives of the skincare world, where emphasis is placed on miracles and fancy buzzwords are used to exacerbate peoples’ need to look good.

Using the latest technology seems like a good way to bring those archaic notions into the present day, but is something like AI a useful tool for something as personal and as individual as skin?

As of yet, that remains unproven.