How Google’s BERT helped me make more progress in 1 day than I’d previously made in 30.
What challenge was this project fixing?
I had previously spent 30 days trying to improve the performance of Company News Sentiment Prediction, but the result was less than satisfactory.
The challenge of this particular classification problem is that of “complex learning task, small training set, large feature space.” The algorithm needed to learn 30+ topics with just 2000+ labeled data in a 100,000+ dimension feature space.
What was your hack day solution?
During hack day, I played with a deep language model, BERT, leveraging transfer learning to boost performance in our news sentiment learning task. …
At CB Insights, we adopt a traditional CSS approach. Here’s why.
It’s 2019, and there are a ton of great declarative libraries and frameworks that can be used to create interfaces. React, Vue, native web components, LitElement, Svelt — the list goes on and on. Just like cookies can’t be great without milk, a great interface can’t be great without styling.
So, which one do you choose?
It’s 2019, so our assumption is that the development community surely would have already rallied around a single solution for styling interfaces, similar to how the community has rallied around React & Redux — right? …
As the digital world evolves, Google is taking a multi-pronged approach to maintaining its dominance in the search and ad business, which makes up the vast majority of its revenue.
Its next big thing will likely come from one of its new priority areas, like cloud, transportation, and healthcare. Each has a massive, global addressable market and plays well to Google’s strength in AI.
Google currently ranks third among cloud providers, after Amazon and Microsoft. However, the company has built out its presence in the space through a number of investments, acquisitions, and internal efforts. …
Sallie Krawcheck spoke at the CB Insights Future of Fintech conference last week in New York City. Watch her full talk and all of our fireside chats with leading financial services executives, fintech startups, and venture investors here.
Sallie Krawcheck wants more women to invest.
It’s the core mission behind her company Ellevest, which provides a digital investment platform that helps women prioritize their specific investment wants and needs.
It also offers its clients other financial services with the help of its certified financial planners, executive coaches, and financial providers.
When Ellevest was created, it was initially mocked (even by Krawcheck herself). She was told that women would not invest, let alone on an online platform. Its tag line “invest like a woman” even drew ire from a big chunk of women Krawcheck heard from. …
Brad Garlinghouse spoke at the CB Insights Future of Fintech conference last week in New York City. Watch his full talk and all of our fireside chats with leading financial services executives, fintech startups, and venture investors here.
Since its earliest days, Ripple has been focused on how it can unlock “the Internet of Value,” according to Brad Garlinghouse, CEO of enterprise blockchain firm Ripple.
“We acknowledge we’re on Mile 1 of a 26 mile race,” stated Garlinghouse while speaking to Fortune’s Jeff Roberts at CB Insights’ annual Future of Fintech conference.
Ripple, which offers a blockchain-based cross-border payments solution, is built on top of its associated cryptocurrency XRP. …
This is excerpted from a long-form overview of how AI is transforming retail. See the full post on our Research Portal, and see if your favorite brands are surviving the “retail apocalypse” in our briefing.
Retail is under pressure to crack the AI code.
After all, corporations in every industry are scrambling to adapt and integrate artificial intelligence into their products — and retail is no exception.
For traditional retail giants, this means entering the playing field with the likes of e-commerce behemoths Amazon and Alibaba, both of which are leveraging big data and powerful AI algorithms to transform the retail space. …
Hang on — feeling confused by blockchain? Check out our blockchain explainer, which breaks down the technology in clear, jargon-free English, and other bitcoin & blockchain content on our Research Portal.
The role of centralized internet giants as “digital identity brokers” is under fire.
While users worry about privacy, tech giants (like Facebook and Google) primarily focus on generating profits and satisfying shareholders.
These companies make money by collecting, selling, and analyzing user data, primarily on behalf of advertisers. …
In venture capital, returns follow the Pareto principle — 80% of the wins come from 20% of the deals.
Great venture capitalists invest knowing they’re going to take a lot of losses in order to hit those wins.
Chris Dixon of top venture firm Andreessen Horowitz has referred to this as the “Babe Ruth effect,” in reference to the legendary 1920s-era baseball player. Babe Ruth would strike out a lot, but also made slugging records.
Likewise, VCs swing hard, and occasionally hit a home run. Those wins often make up for all the losses and then some — they “return the fund.” …
This feature is excerpted from a long-form overview of WeWork’s strategy and forward-looking moves. See the full post on our Research Portal, or check out our teardowns of Amazon, JPMorgan, Apple, Coinbase, and more.
WeWork is a real estate company valued like a tech company.
At least, that’s the rap on WeWork from critics who think it can’t support its $20B valuation in private markets.
Backed by Japanese tech and telecom giant SoftBank Group, WeWork specializes in rent arbitrage — leasing and developing properties at one price, then turning around and renting them out at much higher prices. …
Imagine a future where parents can create bespoke babies, selecting the height and eye color of their yet unborn children. In fact, all traits can be customized to one’s preferences: the size of domestic pets, the longevity of plants, etc.
It sounds like the backdrop of a dystopian science fiction novel. Yet some of this is already happening.
Since its initial discovery in 2012, scientists have marveled at the applications of CRISPR (also known as Cas9 or CRISPR-Cas9). …