Posts from machine learning

Clarifai Mobile SDK - Machine Learning On The Phone

Our portfolio company Clarifai introduced something exciting yesterday.

It is a mobile SDK that any developer can put into their mobile app and it will allow for machine learning on the device:

Machine learning (the process by which computers can get smarter through data examples instead of explicit programming) requires massive computational power, the kind usually found in clusters of computer servers in massive datacenters (ooooh, the cloud). This means that machine learning technology is usually only available to those who can connect to the cloud.

Not anymore! Clarifai’s Mobile SDK gives users the power to train and use AI in the palms of their hands by installing machine learning capability directly on their devices, bypassing the traditional requirement of internet connectivity and massive computing power. After all, these days we have tiny supercomputers in our pockets – our mobile phones. Starting with an iOS SDK, Clarifai is on a mission to make user experiences uniquely personalized on any device from your cellphone to your toaster, anywhere in the world.

Here’s a slideshow that explains how this works:

#machine learning

Numeraire Is Live

Back in February, I posted about Numeraire.

I wrote:

the Numerai team has now gone a step further and issued a crypto-token called Numeraire to incent these data scientists to work together to build the best models instead of just competing with each other

And roughy four months later, I am happy to write that the Numeraire token is live on the Ethereum blockchain.

You can read more about this here.

Well done Numerai team.

#crypto#machine learning#stocks

Using Your Data To Make Your Product Better

One of my favorite uses of AI is to use the data in your product to make your product better. I am talking about making a better UI using AI on your data.

Our portfolio company Quizlet did just that and wrote about it here.

They used the ~150mm study sets that their users have put into Quizlet over the last ~12 years to predict suggested definitions during the create study set mode.

Here is what it looks like.

 

 

 

 

 

 

 

 

 

 

I think this is super cool and a great way to make your product better.

#machine learning

Video Of The Week: Why Toddlers Are Smarter Than Computers

I saw Gary Marcus give a talk at the NYU AI Event I blogged about this past week. In that

In that talk, he suggested that Deep Learning wasn’t going to get us all the way to where we want to get in AI.

I thought it was an interesting take on AI, particularly right now when the buzz and hype is so high.

This TedX talk makes the same argument and so I am sharing it with you.

#machine learning

The AI NexusLab

NYC is an emerging hub for AI and AI startups. That is because of the large number of mathematicians, scientists, and programmers trained in AI who work on wall street, because of leading institutes like NYU’s Courant School that work on cutting-edge science in the field, and because of a number of programs aimed at AI startups here in NYC.

A few weeks ago I was at the Future Labs AI Summit to hear about AI from Yann LeCun, Gary Marcus, and many others. Below is a short highlight video of the summit.

Here is a short highlight video of the summit.
 

 
The Future Labs at NYU Tandon are now accepting applications for the second cohort or the AI NexusLab to find AI companies to support.

Applications for the next AI NexusLab cohort close Wednesday, May 3rd and conclude with the next Future Labs AI Summit in November.
 
If you are and AI startup or you are familiar with any early stage artificial intelligence startups who you think could benefit from our program, please have them apply at www.nexuslab.ai/
 
Accepted companies receive
• $100K in funding
• An NYU student fellow for the duration of the program
• mentorship from leading AI faculty and industry experts
• Access to papers and academic research
• Access to data sets
• Partner opportunity to pilot partners (the last cohort included Daimler, Tough Mudder, Quontopian, and others)
• More than 400K worth of support and services.
• Present at the next Future Labs AI Summit (last speakers included Yann LeCun, Gary Marcus, and others)

The Future Labs are also hosting office hours this Friday, April 28th from 1:00pm-5:30pm for teams who have questions about the program at the Data Future Lab – 137 Varick Street, 2nd Floor.

#machine learning

Machine Learning For Beginners

So you are hearing a lot about machine learning these days.

You are hearing words like models, training, forks, splits, branches, leafs, recursion, test data, and overfitting, and you don’t know what any of them mean.

Well I have some help, courtesy of my colleague Jacqueline who shared this scrolling lesson in machine learning with her colleagues at USV (me included) this weekend.

This scrolling machine learning lesson was made by Stephanie and Tony. It is great work. Thanks!

#machine learning

Brain Computer Interface

The WSJ reported yesterday that Elon Musk is developing yet another company, this one based on neural lace technology, to create a brain computer interface.

Neural lace technology, as I understand it, involves implanting electrodes into the brain so that the brain can control machines directly without the need for an IO device like a mouse, keyboard, or voice interface.

I have no idea how advanced this technology is and whether it is ready for commercialization or if this is basically a research project masquerading as a startup.

But in some ways that doesn’t matter if you believe that at some point someone or some group of scientists and medical professionals will figure out how to directly connect our brain to machines without the need of an IO device.

There are so many times that I have thoughts that I don’t do anything with. They sit idle and maybe go nowhere. But if my brain passively passed those thoughts onto a machine for storage or some other action that could lead to a more productive train of thought that could be incredibly valuable. Or it could drive me insane.

I generally subscribe to the theory that all progress is good as long as we understand the negatives of the technology and we (society) engineer controls and the proper repoanes to it (nuclear weapons being​ an example).

But every time something as mind bending as the idea of connecting our brain to external processing, storage, and communication infrastructure comes before me I do have to pause and ask where this is all going.

At times like this it helps to have a belief system (progress is good). I am all for pushing the envelope of progress as long as we spend an equal amount of time and energy thinking through what might go wrong with things like this.

Hat tip to Niv Dror who read yesterday that I wasn’t sure how I was going to post today and encouraged me to write about this topic.

#machine learning

AI For Legal Cases

Our portfolio Casetext was in the news yesterday for raising $12mm, but the more interesting thing about Casetext is their product, called CARA.

CARA is a research assistant for lawyers that offers a super simple proposition:

Securely upload a brief and discover useful case law

CARA uses Casetext’s wikipedia-like database of >10mm court cases and annotations and sophisticated natural language analysis and artificial intelligence to understand the brief and recommend related cases for a lawyer to analyze and possibly cite in their brief.

Lawyers seem to love CARA. According to Silicon Valley Business Journal:

Casetext’s customers include Quinn Emanuel, Fenwick & West, Ogletree Deakins, Greenberg Traurig and DLA Piper.

“CARA is an invaluable, innovative research tool,” Quinn Emanuel partner David Eiseman said in a statement. “With CARA, we can upload a brief and within seconds receive additional case law suggestions and relevant information on how cases have been used in the past, all in a user-friendly interface.”

We think the legal business is ripe for AI-driven innovation. Much of legal research can and will be automated with tools like CARA.

If you are a lawyer and do a lot of legal research, check out CARA. Securely upload a brief here and check it out.

#machine learning

Machine Learning For Investing In Consumer Goods Startups

Our portfolio company CircleUp has been building a marketplace for startup investing, by accredited and institutional investors, in consumer goods companies (natural foods, personal care, beverage, home goods and apparel). In four years of operation, over $300mm has been raised on CircleUp by entrepreneurs to scale their consumer goods startups.

But underneath all of this has been a sophisticated data science effort designed to track the entire consumer goods sector (all companies, not just the ones on CircleUp) and determine which companies succeed and why. Yesterday CircleUp took the covers off this data science effort, called Helio, and explained what they are up to with it.

Here are some bits from that blog post:

there’s endless data on consumer product and retail companies. And, much of it is public. A quick Google search of the product in your pantry tells you how many SKUs the brand has, price points for each SKU, where they are sold, product reviews, and a great deal more. In an A16Z podcast in 2016, Marc Andreessen commented that machine learning wouldn’t be helpful for tech VC because there isn’t enough data (40:04 mark). We agree. But in the consumer industry, the opposite is true. Data is broadly available. Business models are uniform. That’s the perfect recipe for machine learning. That makes Helio possible.

Let’s take a look at a few examples:

  • Supergoop! is a sunscreen brand available nationally throughout Sephora, that Helio surfaced due to its quickly growing brand, great distribution and estimated revenue growth. We presented Supergoop! to institutional investors, and shortly after, they raised $3.25 million.
  • REBBL is a line of coconut-milk based beverages made with super herbs known to reduce stress. Aside from being one of the fastest growing products in its category, REBBL donates 2.5% of net sales to initiatives helping eradicate human trafficking. Helio spotted REBBL early and qualified it for investors, showing its compelling brand, team and distribution metrics. Today, REBBL’s lead investors include Powerplant Ventures, led by the ZICO coconut water founder, and Boulder Investment Group Reprise.
  • nutpods plays in the crowded plant-based, dairy alternative category. Helio spotted nutpods for its remarkable product reviews, strong early growth and overall brand, despite it having less than $50,000 in annual sales at the time. After, nutpods got investments from Stray Dog Capital and Melissa Hartwig, founder and CEO of Whole 30, and today is rated #1 on Amazon in its category.
  • Tio Gazpacho is a quickly growing brand in the relatively new category of bottled soups, or more broadly, drinkable meals. Tio Gazpacho was founded in Florida, a place without a robust VC community, but Helio still spotted it, and surfaced it to General Mills, which now is its lead investor.

Helio is currently monitoring over a million brands across natural foods, personal care, beverage, home goods and apparel, and can help find who might be the next Krave Jerky, Seventh Generation or Too Faced. We are talking to likely candidates right now, and not just in the categories above, in all categories we see as promising growth areas in the consumer market.

CircleUp has always taken the view that the entrepreneurs with the best ideas, products and team should win…not the one with the best personal connections. Helio brings us a big step closer towards that ambition.

We are excited to see what happens when entrepreneurs with big ideas meet a capital market that has data science at the core. If you want to participate in that market, visit CircleUp.

#hacking finance#machine learning

The Robot Tax And Basic Income

In my work to prepare for the Future of Labor conversation we had at NewCo Shift a few weeks ago, I talked to a number of experts who are studying job losses due to automation and thinking about what might be done about it. Two ideas that came up a number of times were the “robot tax” and the “basic income.”

The ideas are complementary and one might fund the other.

At its simplest, a “robot tax” is a tax on companies that choose to use automation to replace human jobs. There are obviously many variants of this idea and to my knowledge, no country or other taxing authority has implemented a robot tax yet.

A “basic income” is the idea that everyone receives enough money from the government to pay for their basic needs; housing, food, clothing so that as automation puts people out of work we don’t see millions of people being put out on the street.

What is interesting about these two ideas is that some of the biggest proponents of them are technology entrepreneurs and investors, the very people who are building and funding the automation technologies that have the potential to displace many jobs.

It is certainly true that we don’t know that automation will lead to a jobs crisis. Other technological revolutions like farming and factories produced as many new jobs as they wiped out and incomes increased from these changes. Automation could well do the same.

But smart people are wondering, both privately and publicly, if this time may be different. And so ideas like the robot tax and the basic income are getting traction and are being studied and promoted.

The latest proponent of a robot tax is Bill Gates who said this about it:

You ought to be willing to raise the tax level and even slow down the speed. That’s because the technology and business cases for replacing humans in a wide range of jobs are arriving simultaneously, and it’s important to be able to manage that displacement. You cross the threshold of job replacement of certain activities all sort of at once.

There is a lot of economic surplus that could come from automation. Let’s look at ride sharing. Today I pay something like $15 to go from my home to my office in the morning. Something like $10 of that ride is going to the driver. If the ride is automated, either the price goes to $5, saving me $10 a ride which then is surplus to me, or the profit that Uber is making goes up significantly, which is surplus to them. Some of both is likely to happen. This surplus could be taxed, either at the company level or the individual level, so that the cost of the ride doesn’t go down nearly as much and the driver can continue to compete with the robot or the driver can collect some basic income, funded by the robot tax, while they find a new line of work.

At least that is the idea.

I would not characterize myself as a proponent of a robot tax or a basic income. But I find these ideas interesting and worth studying, debating, discussing, and testing at a small scale to understand their impacts. We should absolutely be doing that.

#economics#employment#machine learning#policy