Microsoft acquires conversational AI startup Semantic Machines to help bots sound more lifelike

Microsoft announced today that it has acquired Semantic Machines, a Berkeley-based startup that wants to solve one of the biggest challenges in conversational AI: making chatbots sound more human and less like, well, bots.

In a blog post, Microsoft AI & Research chief technology officer David Ku wrote that “with the acquisition of Semantic Machines, we will establish a conversational AI center of excellence in Berkeley to push forward the boundaries of what is possible in language interfaces.”

According to Crunchbase, Semantic Machines was founded in 2014 and raised about $20.9 million in funding from investors including General Catalyst and Bain Capital Ventures.

In a 2016 profile, co-founder and chief scientist Dan Klein told TechCrunch that “today’s dialog technology is mostly orthogonal. You want a conversational system to be contextual so when you interpret a sentence things don’t stand in isolation.” By focusing on memory, Semantic Machines’ AI can produce conversations that not only answer or predict questions more accurately, but also flow naturally.

Instead of building its own consumer products, Semantic Machines focused on enterprise customers. This means it will fit in well with Microsoft’s conversational AI-based products, including Microsoft Cognitive Services and Azure Bot Service, which are used by one million and 300,000 developers, respectively, and virtual assistants Cortana and Xiaolce.

Women alleging sexual assault by Uber drivers asked to be freed from forced arbitration

A group of women alleging sexual violence from Uber drivers have sent an open letter to the company’s board, asking to be released from the mandatory arbitration clause in the Uber app’s terms of service. The letter was posted on the website of Wigdor LLP, a New York law firm that filed a class action lawsuit against Uber last year on behalf of women who said they were assaulted or raped by Uber drivers and blame the company’s background check procedures.

“Uber’s message to the public are: ‘we help improve access to transportation, and make streets safer’ [and] ‘We do the right thing, period,’” read part of the letter, which was signed by fourteen women. “Secret arbitration is the opposite of transparency. Forcing female riders, as a condition of using Uber’s app, to pursue claims of sexual assault and rape in secret arbitration proceedings does not ‘make streets safer.’ Silencing our stories deprives customers and potential investors from the knowledge that our horrific experiences are part of a widespread problem at Uber.”

They added “when we created Uber accounts, we believed Uber’s promise to provide a ‘safe ride.’ We trusted a company operating in the space of transportation for hire to mean what it says, and we never thought that Uber would perpetuate physical violence against women. But this is exactly what Uber is doing and what is has been doing for years.”

Once a relatively obscure legal issue, mandatory arbitration agreements are now under scrutiny by activists who say they force victims of harassment and discrimination into silence. Opponents of mandatory arbitration say that the closed hearings, which include non-disclosure clauses and are often performed by a third-party arbitrator paid by the company itself, prevent victims from taking further action even as social movements like #MeToo continue to gain ground.

Many companies require employees to sign mandatory arbitration agreements as a condition of employment. According to the Economic Policy Institute, the number of non-union, private sector employees covered by mandatory arbitration clauses has increased dramatically since the early 2000s.

Wigdor LLP noted that some companies, like Microsoft, are ending forced arbitration clauses, especially for sexual harassment, while a bipartisan bill has been introduced in the United States Congress that would end forced arbitration of sexual harassment cases in workplaces, and called on Uber to follow suit.

In an exchange last month with former Uber engineer Susan Fowler on Twitter, Dara Khosrowshahi, who succeeded Travis Kalanick as Uber’s chief executive officer last August, signaled that he is willing to consider ending forced arbitration. “I will take it seriously, but we have to take all of our constituents into consideration,” he wrote.

Fowler, whose blog post detailing sexual harassment at the company led to an internal investigation and contributed to the resignation of Kalanick, is now backing a new bill in California that would forbid companies from forcing employees to enter mandatory arbitration agreements in response to sexual harassment and other workplace discrimination complaints.

Waze signs data-sharing deal with AI-based traffic management startup Waycare

Waze has struck a data-sharing agreement with Waycare, an artificial intelligence-based traffic management startup, the two companies announced today. The deal will allow them to combine anonymized navigation information crowdsourced from the 100 million drivers who use Waze with Waycare’s proprietary traffic analytics.

The collaboration is now active in Nevada, Florida, California and Nevada, with plans to expand over the next year. It is part of Waze’s Connected Citizens Program, which gives cities around the world access to anonymized driver data to help them manage traffic and road infrastructure.

A representative told TechCrunch that data supplied by Waycare to Waze will be incorporated into the app’s usual interface, while data from Waze will be added to Waycare’s platform alongside its other data sets.

[gallery ids="1628899,1628898"]

 

Founded in 2016, Waycare is a cloud-based platform that enables municipalities to gather data from many sources, including on-board devices, navigation apps, sensors and road camera feeds, and analyze them using proprietary deep learning algorithms to figure out how to improve traffic and road conditions. The new partnership means cities that use Waycare will be able to send urgent alerts to drivers through Waze, while giving Waycare a new trove of data.

Like Waze, which was acquired by Google in 2013, Waycare is based in Tel Aviv with operations in the U.S. The startup has raised $2.3 million so far, according to Crunchbase, and currently has projects in Nevada, Florida, Delaware and California.

 

YouTube releases its first report about how it handles flagged videos and policy violations

YouTube has released its first quarterly Community Guidelines Enforcement Report and launched a Reporting Dashboard that lets users see the status of videos they’ve flagged for review. The inaugural report, which covers the last quarter of 2017, follows up on a promise YouTube made in December to give users more transparency into how it handles abuse and decides what videos will be removed.

“This regular update will help show the progress we’re making in removing violative content from our platform,” the company said in a post on its official blog. “By the end of the year, we plan to refine our reporting systems and add additional data, including data on comments, speed or removal and policy removal reasons.”

But the report is unlikely to quell complaints from people who believe YouTube’s rules are haphazardly applied in an effort to appease advertisers upset their commercials had played before videos with violent extremist content. The issue came to the forefront last year after a report by The Times, but many content creators say YouTube’s updated policies have made it very difficult to monetize on the platform, even though their videos don’t violate its rules.

YouTube, however, claims that its anti-abuse machine learning algorithm, which it relies on to monitor and handle potential violations at scale, is “paying off across high-risk, low-volume areas (like violent extremism) and in high-volume areas (like spam).”

Its report says that YouTube removed 8.2 million videos during the last quarter of 2017, most of which were spam or contained adult content. Of that number, 6.7 million were automatically flagged by its anti-abuse algorithms first.

Of the videos reported by a person, 1.1 million were flagged by a member of YouTube’s Trusted Flagger program, which includes individuals, government agencies and NGOs that have received training from the platform’s Trust & Safety and Public Policy teams.

YouTube’s report positions views a video received before being removed as a benchmark for the success of its anti-abuse measures. At the beginning of 2017, 8% of videos removed for violent extremist content were taken down before clocking 10 views. After YouTube started using its machine-learning algorithms in June 2017, however, it says that percentage increased to more than 50% (in a footnote, YouTube clarified that this data does not include videos that were automatically and flagged before they could be published and therefore received no views). From October to December, 75.9% of all automatically flagged videos on the platform were removed before they received any views.

During that same period, 9.3 million videos were flagged by people, with nearly 95% coming from YouTube users and the rest from its Trusted Flagger program and government agencies or NGOs. People can select a reason when they flag a video. Most were flagged for sexual content (30.1%) or spam (26.4%).

Last year, YouTube said it wanted to increase the number of people “working to address violative content” to 10,000 across Google by the end of 2018. Now it says it has almost reached that goal and also hired more full-time anti-abuse experts and expanded their regional teams. It also claims that the addition of machine-learning algorithms enables more people to review videos.

In its report, YouTube gave more information about how those algorithms work.

“With respect to the automated systems that detect extremist content, our teams have manually reviewed over two million videos to provide large volumes of training examples, which improve the machine learning flagging technology,” it said, adding that it has started applying that technology to other content violations as well.

YouTube’s report may not ameliorate the concerns of content creators who saw their revenue drop during what they refer to as the “Adpocalpyse” or help them figure out how to monetize successfully again. On the other hand, it is a victory for people, including free speech activists, who have called for social media platforms to be more transparent about how they handle flagged content and policy violations, and may put more pressure on Facebook and Twitter.

Indian lending platform Capital Float raises $22M Series C extension from Amazon

Capital Float, the fintech startup that says it is India’s largest online lender, announced today that it has raised $22 million in new funding from Amazon. At the end of last year, reports surfaced that Amazon was considering an investment in Capital Float as an extension of its $45 million Series C, which was announced last August. The Bangalore-based startup confirmed to TechCrunch that Amazon’s investment is indeed an extension of that round and brings the total equity it has raised over the past 12 months to $67 million.

Over the same period, Capital Float also raised $80 million of debt from banks and other financial companies, which it combines with its own balance sheet to finance loans to small businesses and other borrowers. Amazon India is among several e-commerce platforms that the company has partnered with to provide loans to sellers, including Snapdeal and Shopclues.

Since its inception in 2013 by co-founders Sashank Rishyasringa and Gaurav Hinduja, Capital Float has raised a total of about $110 million in equity funding from investors, including Ribbit Capital, SAIF Partners, Sequoia India, Creation Investments and Aspada, as well as total debt lines of $130 million.

During the last six months, Capital Float added 50,000 new customers, bringing its total customer base to more than 80,000 people in more than 300 cities. The startup says it currently disburses more than 10,000 loans each month and now has an outstanding loan portfolio of more than $170 million, with a default rate of about 2 percent. About 70 percent of its loans are microloans ranging from 25,000 rupees to 500,000 rupees (about $376 to $7,530).

With the investment from Amazon, the startup has set an ambitious goal of adding 300,000 new customers and originating more than $800 million in loans this year.

In a press statement, Amazon India’s country manager Amit Agarwal said, “We’re excited to work with Capital Float and invest alongside other investors. We are highly impressed with what Gaurav and Sashank have built and we back missionary entrepreneurs and management teams. Credit in India is highly under-penetrated and Capital Float is bringing the right kind of credit solutions to the underserved and informally served segments of SMEs to help realize their full potential.”

Over the last year, Capital Float expanded into more verticals, including products for small- to mid-sized manufacturers, point-of-sale financing for retailers and loans for school construction and self-employed professionals like doctors. It also added new online payment gateways to make it easier for borrowers to repay loans and began piloting deep learning-based underwriting models that use data points like image processing, geotags and new policies such as the Goods and Service Tax (GST), an indirect tax launched last year that is levied at every step of the production chain and the banknote demonetization started by Prime Minister Narendra Modi’s government in 2016.

Amazon’s new blockchain service competes with similar products from Oracle and IBM

Amazon Web Services announced Blockchain Templates late last week, a “blockchain -as-a-service” offering that competes with similar products from Oracle and IBM. The launch shows how eager the biggest enterprise players are to get ahead in the blockchain game even if their customers are still trying to pinpoint exactly what blockchain can do for them (and some investors are starting to temper their initial excitement).

In a blog post about how to use Blockchain Templates, AWS vice president and chief evangelist Jeff Barr acknowledged the lack of clarity by referencing a 1970s “Saturday Night Live” sketch about Shimmer Floor Wax, a floor polish that is also a non-dairy dessert topping.

“Some of the people that I talk to see blockchains as the foundation of a new monetary system and a way to faciliate international payments. Others see blockchains as a distributed ledger and immutable data source that can be applied to logistics, supply chain, land registration, crowdfunding and other use cases,” he wrote. “Either way, it’s clear that there are a lot of intriguing possibilities and we are working to help our customers use this technology more effectively.

AWS Blockchain Templates give AWS users working on blockchain apps a faster way to set up Ethereum or Hyperledger Fabric networks. Its launch comes six months after Oracle unveiled its cloud service built on the open-source Hyperledger Fabric project during Oracle OpenWorld and about a year after IBM announced its own Hyperledger-based blockchain-as-a-service offering.

Another new competitor in the BaaS market is Huawei, which announced its Blockchain Service, also built on Hyperledger, last week during its analyst conference in Shenzhen. It joins other Chinese tech companies, including Baidu and Tencent, that already had blockchain platforms.

TaskRabbit CEO posts statement as its app returns following a cybersecurity breach

After taking them down to investigate what it called a “cybersecurity incident,” TaskRabbit’s website and app are back online. The Ikea-owned platform for on-demand labor also posted an update from its chief executive officer Stacy Brown-Philpot about the incident.

“While our investigation is ongoing, preliminary evidence shows that an unauthorized user gained access to our systems. As a result, certain personally identifiable information may have been compromised,” she wrote.

While Brown-Philpot said that an outside forensics team is currently working to identify what information was compromised and will notify all affected users, she urged the platform’s customers and providers, called “taskers,” to monitor online accounts for suspicious activity and change passwords if they used the same login information on other services.

TaskRabbit will add several new security measures because of the incident. Brown-Philpot said they are working on ways to make their login process more secure, reduce the amount of data retained about customers and taskers and “enhance overall network cyber threat detection technology.”

The company will continue posting updates to a dedicated page on its website, which also includes a FAQ for taskers who were unable to complete jobs while the app was offline. TaskRabbit says people who were forced to reschedule or cancel tasks will be compensated.

ZTE postpones earnings report after being slapped with U.S. exports ban

ZTE will postpone the release of its quarterly earnings report after the United States government banned American companies from selling goods to the Chinese telecom and smartphone maker. In a filing with the Hong Kong stock exchange, ZTE said that its earnings report, originally set to be released on Thursday, is now delayed to an undetermined date.

About a year ago, ZTE pleaded guilty to violating U.S. sanctions against Iran and North Korea. Its deal with the U.S. government included penalties and fines totaling more than a billion dollars, but allowed it to continue doing business with U.S. suppliers.

On Monday, however, the U.S. Department of Commerce announced that ZTE had failed to follow the agreement’s terms. It accused the company of making false statements and failing to punish employees and senior management. As a result, the Department of Commerce slapped ZTE with a seven-year export restriction.

This is a huge blow to ZTE, which sources a significant portion of its most important components, including processors, from U.S. companies like Qualcomm . It also means ZTE may lose access to software licensed from Google, including Android.

This is the latest in ZTE’s string of entanglements with the U.S. government. Despite holding fourth place in the U.S. smartphone market share, after Samsung, Apple and LG, ZTE is under scrutiny by U.S. intelligence agencies, which believe that it and fellow Chinese smartphone maker Huawei may pose security concerns.

Funding Societies, a Southeast Asian lending platform, gets $25M Series B led by Softbank Ventures Korea

Funding Societies co-founders Reynold Wijaya and Kelvin Teo.

Funding Societies, a peer-to-peer lending platform in Southeast Asia, said today that it has raised a $25 million Series B led by Softbank Ventures Korea, the Japanese tech conglomerate’s early-stage venture capital unit. The round included returning investors Sequoia India, which led the Singapore-based startup’s Series A two years ago, Golden Gate Ventures and Alpha JWC Ventures, as well as new backers Qualgro and LINE Ventures.

Funding Societies also said it has raised credit lines from banks and financial institutions to lend to small- to medium-sized businesses. Founded in 2015 by Kelvin Teo and Reynold Wijaya, the startup’s name represents its “vision of financial inclusion in Southeast Asia.”

Its Series B was oversubscribed, says Funding Societies, which operates in Singapore, Indonesia, where it is called Modalku, and Malaysia.

When it announced its $7.5 million Series A in August 2016, Funding Societies had disbursed $8.7 million Singaporean dollars, a number that has since grown to $145 million SGD, chief executive officer Teo tells TechCrunch. Since its launch, the startup has increased its lender base to more than 60,000 and now claims a default rate of less than 1.5%, down from about 2% to 3% two years ago, thanks to improvements in its underwriting model.

In a press statement, Softbank Ventures Korea partner and managing director Sean Lee said the firm “has been actively investing across Southeast Asia. SME digital lending across Southeast Asia is where we saw huge growth potential. Among many players, we were most impressed with Funding Societies for what it has achieved in a short period of time and its potential to continue to become the number one player.”

Though Teo says Funding Societies is “always exploring other markets, there is still tons of work we need to do in our current three markets.” Despite its considerable growth over the past three years, the startup’s mantra is “slow and steady,” a phrase Teo repeated often during our interview.

“One of the key things we highlight is that it’s more important for us to grow slowly and steadily instead of fast and recklessly, because it’s a trust-based industry,” says Teo.

“We need to give out loans and be able to collect them back, so we focus on learning the market, understanding the market and solving key pain points instead of giving out a bunch of loans to chalk up high numbers and attract VCs.”

For example, though the platform may offer personal loans in the future, Teo said it currently only lends to SMEs because “we believe that we are strategically better suited to serving small businesses and, in terms of our company’s values, we think that serving SMEs is an expansionary effort. Consumer financing, in our personal view, is more consumptive finance. It doesn’t help grow economies.”

Many of the SMEs the company serves are very small. Some of its Indonesian borrowers, for example, make annual revenue of about $5,000 USD per year.

“Many of these borrowers are seeking their first business loan and do not have other sources of financing. A lot of financial institutions take a collateral underwriting approach and a lot of budding businesses would not be able to secure financing that way,” says Teo.

“But we also see some of them come to us as a form of top-up. They already have a bank loan, but it is insufficient for them, so they come to us because they are limited by the size of their collateral. Also, we are able to process financing faster than traditional institutions.”

Funding Societies was created to give SMEs, many of which had previously relied mostly on friends and family loans, access to more means of financing. The company points to a recent study by Ernst & Young, UOB and Dun & Bradstreet that says 65.2% of SMEs in Southeast Asia do not have easy access to traditional business financing, even though most are open to other options, including peer-to-peer lending platforms.

The company says it was the first online peer-to-peer lending platform in Malaysia and that based on third-party data, it is now the leading SME lending platform there, as well as one of Singapore’s three largest peer-to-peer lending platforms. It also holds sizable market share in Indonesia.

Though its platform uses algorithms for initial application screening, a significant portion of work, depending on loan size, is still done by Funding Societies’ employees, who have grown in number from 70 in 2016 to 165 now (Teo says the company is currently hiring in earnest and willing to pay relocation costs for promising talent). Almost all applicants talk directly to someone from the company. Micro-loans, which range in size from $500 USD to $40,000 USD, usually take about two business hours to approve and disburse, while applicants for larger loans may have to wait a few days to about a week.

“We’ve debated and discussed internally a lot if we leave too much money on the table, because our default rate is lower than certain banks in the markets we are serving, but given that we are still at a relatively nascent stage in the lending market and have no control over financial crises, it is more important to stay prudent than to grow recklessly,” says Teo.

This methodical approach is also important when entering new markets. Though many outside observers take the umbrella term “Southeast Asia” a little too literally, ignoring cultural differences between each country, Teo says it is still a fragmented market, so financial service companies need to localize carefully. When Funding Societies enters a new market, it can probably port about 50% of its tech and business model from its previous market, but the other half has to be built from ground up to account for economic and cultural differences, he adds.

“SME financing is a very localized business. With sufficient capital you can win the market and it’s really driven by subsidies and strong marketing,” Teo says. “But for SMEs, you really, really need to understand the local market.”