Highlights this week include two big autonomous vehicle partnerships and some heavy fundraising rounds. Honda and GM, in addition to Toyota and Softbank, launched autonomous vehicle partnerships with major commitments. The space continues to build and scale based on who partners with who.
On the fundraising side, ZipRecruiter and Hopper both had over $100mm fundraising rounds, pushing them to near unicorn status in terms of their valuations.
Company developments:
Google to release DeepMind’s StreetLearn for teaching machine-learning agents to navigate cities – Oct. 5, 2018 (Tech Republic)
- The StreetLearn environment relies on images from Google Street View and has been used by Google DeepMind to train a software agent to navigate various western cities without reference to a map or GPS co-ordinates, using only visual clues such as landmarks as it wanders the streets
- The StreetLearn environment encompasses multiple regions within the centers of the cities of London, Paris and New York. It is made up of cropped 360-degree panoramic images of street scenes from Street View, each measuring 84 x 84 pixels. Each panoramic image is a node in larger network or graph of images, with up to 65,000 nodes per 5km city region, and multiple regions per city. Each region has a distinct urban setting, for instance differing amount of construction and varying numbers of parks and bridges. For example, in New York the four distinct environments used for training were Harlem, Central Park, Midtown, and Greenwich Village
Burger King’s ‘AI-written’ ads show we’re still very confused about artificial intelligence – Oct. 3, 2018 (The Verge)
- The chicken crossed the road to become a sandwich. Burger King encouraged the chicken,” says the voice. “The Whopper lives in a bun mansion, just like you,” it chirps. They’re good ads! And, of course, they’re lies. In a press release, Burger King claims the videos are the work of a “new deep learning algorithm,” but an article from AdAge makes it clear that humans — not machines — are responsible for the funnies. “Artificial intelligence is not a substitute for a great creative idea coming from a real person,” Burger King’s global head of brand marketing, Marcelo Pascoa, told the publication
DeepMind AI can generate convincing photos of burgers, dogs, and butterflies – Oct. 2, 2018 (VentureBeat)
- A system that can produce convincing photos out of whole cloth is the holy grail of image synthesis, and thanks to artificial intelligence (AI), researchers at Google subsidiary DeepMind and Heriot-Watt University believe they’ve come pretty darn close to creating it. In a paper published this week on the preprint server Arxiv.org (“Large Scale GAN training for high fidelity natural image synthesis“), they describe AI that approximates food pics, landscapes portraits, and candid pet photos with impressive consistency. In some cases, the novel samples are almost impossible to distinguish from those taken with a camera
- The trick was employing large, highly optimized generative adversarial networks (GAN), or two-part neural networks consisting of generators that produce samples and discriminators that attempt to distinguish between the generated samples and real-world samples. The teams’ system, which they dubbed “BigGANs,” benefited from architectural tweaks, an increased batch size (2,048 images), and four times as many parameters (158 million) — the algorithmic levers used to control certain properties of the model — compared to prior art
Ford shares proposed standard for autonomous cars to signal intentions – Oct. 2, 2018 (CNET)
- In a blog post on Tuesday, Ford said that other road users shouldn’t be expected to change their behavior around autonomous vehicles. But the company believes that any Level 4 autonomous vehicle — that’s one that operates in certain situations without a human driver needing to take control — should “communicate intent.” As a result, Ford wants to develop a standard for all autonomous cars to use.
- Ford’s proposed standard would have autonomous cars use the following signal lights: two white lights moving side-to-side would signal the car is yielding and is about to stop. A solid white light would signal the vehicle is driving autonomously. And rapidly blinking white lights would serve as a “start-to-go” warning before the car accelerates. These were tested in the real-world study as well as in virtual-reality simulations
These ex-Apple engineers are building sensors for self-driving cars – Oct. 1, 2018 (CNBC)
- Silicon Valley start-up Aeva has been secretive about the technology it’s been building to help self-driving cars understand what’s happening around them. On Monday, the company finally revealed its creation: a system that detects velocity, depth and reflectivity more than 200 meters away, without using a ton of power — in a unit that’s the size of a tissue box
- With around 50 employees and $45 million in funding from investors including Canaan Partners and Lux Capital, Aeva is still a small outfit. But it’s working with a handful of big-name automakers on integrating the product into various types of vehicles, said Soroush Salehian, who co-founded Aeva with Mina Rezk. A customer can employ as many as five of the boxes on a single vehicle, depending on the level of automation that’s desired, with each unit costing $200 or $300, Salehian said
M&A:
Social intelligence platforms Brandwatch and Crimson Hexagon to merge – Oct. 4, 2018 (TechCrunch)
- Brandwatch — the UK-HQ’d social intelligence company that raised $65 million from European VCs Nauta Capital, Highland Europe and Partech — is to merge with another leading player in the space, Crimson Hexagon . The merger, expected to close in Q4, will, according to the companies, create a business with around $100 million in recurring annual revenues. The merged entity will simply be called Brandwatch
- The idea is to create new, artificial intelligence-driven products to help brands understand consumer behavior, especially via social media. In a statement, Giles Palmer, founder and CEO of Brandwatch, said: “In this digitally connected world, our vision is to transform how organizations understand their consumers through products that bring structure and meaning to the public voices of billions of people. This merger allows us to accelerate towards that vision and move beyond social listening to innovate at the cross section of brand, market and consumer intelligence.”
Google acquires customer service automation startup Onward – Oct. 2, 2018 (TechCrunch)
- Google has acquired a small startup that builds tools for businesses looking to automate their customer service or sales workflows. Onward and some of its key employees, including co-founders Rémi Cossart and Pramod Thammaiah, as well as CTO Aaron Podolny, will be joining Google. Terms of the deal weren’t disclosed
- The startup gives businesses an AI-powered chat solution to get customers what they want while making the most efficient use of their resources through automation. Cossart and Thammaiah pivoted the startup from a more consumer-facing product called Agent Q, which was a sort of shopping virtual assistant that people could text and quickly ask for product recommendations. The team billed it as a product marriage of Magic and Consumer Reports
LivePerson acquires Conversable, a conversational AI company – Oct. 1, 2018 (VentureBeat)
- Conversable, maker of automated bots for popular food chains such as McDonald’s, TGI Friday’s, Pizza Hut, Dunkin’ Donuts, and Denny’s, was acquired today by LivePerson. Other companies using Conversable have been as varied as Sam’s Club, Whole Foods, and government contractor Booz Allen. A platform to help customers better train the machine learning they use was introduced in March 2017. Financial terms of the deal were not disclosed
- Whereas LivePerson is a publicly traded company that once focused solely on connecting people for web chat — as indicated by the name LivePerson — in 2016 the company started to offer customers automated bots and extend a series of AI services based on years of chat logs between customers and businesses. Conversable will help LivePerson continue to accelerate its goal of providing conversational commerce products to customers, CEO Robert Locascio told VentureBeat in an email
Fundraising / investment:
Honda to invest $2.8bn in GM’s self-driving car unit – Oct. 4, 2018 (BBC)
- Honda is to invest $2.75bn (£2.1bn) and take a stake in General Motors’ self-driving unit, GM Cruise, as firms continue to team up in the race to develop autonomous vehicles
- Under the tie-up with GM, Honda will contribute approximately $2bn over 12 years to self-driving vehicle initiatives, which together with a $750m equity investment in Cruise, brings its total commitment to the project to $2.75bn
ZipRecruiter picks up $156M, now at a $1B valuation, for its AI-based job-finding marketplace – Oct. 4, 2018 (TechCrunch)
- ZipRecruiter, a startup based out of LA that has built a marketplace for employers to post open positions, and — by way of AI — help job-seekers to connect to openings that best match their skills and interests, has raised $156 million in a Series B round of funding, at a valuation I understand from sources to be in the region of $1 billion
- This latest round is being led by Wellington Management Company and returning investor IVP, along with participation from unnamed existing investors. ZipRecruiter has in total now raised $219 million, with its previous round, $63 million led by IVP in 2014, being its first significant outside funding
Hopper raises $100M more for its AI-based travel app, now valued at $780M – Oct. 3, 2018 (TechCrunch)
- Hopper — a mobile-only travel booking app cofounded by a former Expedia executive in Montreal, Canada that uses artificial intelligence to help you search for and book hotels and flights — has gained a little elevation of its own today. The startup has raised another $100 million in funding, money that it plans to use to build out its AI algorithms and expand deeper into international markets. Hopper has now passed 30 million installs and 75 million trips planned, and says it’s on track to make nearly $1 billion in sales this year
- Sources very close to the company say Hopper’s valuation with this round is also flying: it’s now close to 1 billion Canadian dollars ($780 million in US dollars). As a point of comparison, Hopper was valued at US$300 million in its last round, in late 2016, and it has raised C$184 million (US$235 million) to date. Throughout that time, it’s been a consistent presence in the top-10 travel apps in the US, according to stats from App Annie
SoftBank leads $35M investment in sports engagement startup Heed – Oct. 2, 2018 (TechCrunch)
- Heed, a startup looking to create new ways for sports leagues and clubs to engage with fans, is announcing that it has raised $35 million led by SoftBank Group International. CEO Danna Rabin emphasized that Heed’s real focus isn’t on building fancy hardware, but rather on the artificial intelligence it uses to take that data (which can also be drawn from video and audio footage of the match) and transform it into a general narrative that can be viewed on the Heed smartphone app
- “One of our core promises is that it’s not editorial driven,” Rabin added. “The AI is selecting what’s interesting in a match. Of course, we have a creative team that designs the formats, the visuals, how the packaging should look like, but that’s incorporated into the technology, which is automatically selecting the moments and creating the experiences with no human interpretation”
AI accelerator Zeroth bags investment from digital media firm Animoca – Oct. 2, 2018 (TechCrunch)
- Asia-based accelerator program Zeroth is getting a major infusion of capital after digital media company Animoca Brands agreed to invest over $3 million into its businesses. The relationship is not new. Animoca previously invested US$1 million (A$1.39 million) in Hong Kong-based Zeroth last December, and now it is following up to take a majority stake in Zeroth’s operational business and also joining its fund as an LP
Boundary Holding invests in facial recognition startup XRVision – Oct. 1, 2018 (Biometric Update)
- Facial recognition and video analytics firm XRVision has secured an undisclosed sum in funding from Boundary Holding, a Luxembourg-based early-stage technology fund. The Singapore-based startup will use the capital to expand into new markets
- XRVision was founded in 2015 by Guy Ron and Yaacov Apelbaum and has previously secured funding from NUS and TNF Ventures. The company reports clients in Singapore, Taiwan, Israel, the U.S., the UK, and Thailand
Partnerships:
Fujitsu Laboratories and MIT’s Center for Brains, Minds and Machines Broaden Partnership – Oct. 4, 2018 (MIT News)
- Fujitsu Laboratories Ltd. and MIT’s Center for Brains, Minds and Machines (CBMM) has announced a multi-year philanthropic partnership focused on advancing the science and engineering of intelligence while supporting the next generation of researchers in this emerging field. The new commitment follows on several years of collaborative research among scientists at the two organizations
- “CBMM faculty, students, and fellows are excited for the opportunity to work alongside scientists from Fujitsu to make advances in complex problems of intelligence, both real and artificial,” said CBMM’s director Tomaso Poggio, who is also an investigator at the McGovern Institute and the Eugene McDermott Professor in MIT’s Department of Brain and Cognitive Sciences. “Both Fujitsu Laboratories and MIT are committed to creating revolutionary tools and systems that will transform many industries, and to do that we are first looking to the extraordinary computations made by the human mind in everyday life.”
Toyota and SoftBank enter into joint venture for driverless vehicle services – Oct. 4, 2018 (The Telegraph)
- The business, which will be called Monet Technologies Corp, will start with 2bn yen (£13.5m) in capital. It will initially be focusing on the Japanese market, offering ride-hailing services for public agencies and private companies, before looking to go global
- SoftBank said the company could eventually perform a whole host of “just in time” services, such as preparing food in transit as it is being delivered, combining its platform for the internet of things with Toyota’s connected vehicle technology. Monet will not develop its own autonomous vehicles, instead looking at the services that such vehicles can provide
Research / studies:
Stanford AI detects even the smallest earthquakes from seismic data – Oct. 5, 2018 (VentureBeat)
- A possible solution to detecting micro earthquakes is described in a new paper from the Department of Geophysics at Stanford University, where scientists have developed an AI system — dubbed Cnn-Rnn Earthquake Detector, or CRED — that can isolate and identify a range of seismic signals from historical and continuous data
- It builds on the work of Harvard and Google, which in August created an AI model capable of predicting the location of aftershocks up to one year after a major earthquake. The researchers’ system consists of neural network layers — interconnected processing nodes that loosely mimic the function of neurons in the brain — of two types: convolutional neural networks and recurrent neural networks. The former extracts features from seismographs, while the latter — which can combine memory and inputs to improve the accuracy of its predictions — learns the sequential characteristics of said seismographs
Government / policy:
Google’s Sundar Pichai secretly met Pentagon leaders over artificial intelligence project: Report – Oct. 7, 2018 (Economic Times)
- Google’s India-born chief executive Sundar Pichai quietly paid a visit to the Pentagon to ease tensions that erupted after employee outrage prompted the tech giant to sever a controversial defence contract to analyse drone video, according to a media report
- Pichai met with a group of civilian and military leaders mostly from the office of the Under Secretary of Defence for Intelligence, the Defense Department directorate that oversees the artificial-intelligence drone system known as Project Maven, according to the people, who spoke on the condition of anonymity to speak freely about the meeting, The Washington Post reported
India plans facial recognition technology at airports – Oct. 4, 2018 (Phys.org)
- The scheme will be rolled out early next year in a bid to help congested airports deal with rapidly rising passenger numbers, the ministry of civil aviation said
- Passengers on domestic flights will be able to choose whether they want to use the biometric authentication system and go paperless, ministry secretary Rajiv Nayan Choubey said in a statement
U.S. plans to rewrite rules that impede self-driving cars – Oct. 4, 2018 (Reuters)
- The Trump administration is moving ahead with plans to revise safety rules that bar fully self-driving cars from the roads without equipment such as steering wheels, pedals and mirrors, according to a document made public on Thursday
- The National Highway Traffic Safety Administration (NHTSA) “intends to reconsider the necessity and appropriateness of its current safety standards” as applied to automated vehicles, the U.S. Department of Transportation said in an 80-page update of its principles dubbed “Automated Vehicles 3.0.”
Dulles facial recognition tech nabs 3 impostors in 40 days – Oct. 3, 2018 (WBDJ7)
- Citing a U.S. Customs and Border Protection release, The Washington Post reports a woman arriving on a Monday flight from Accra, Ghana, presented a U.S. passport, but the facial recognition technology reported a mismatch. A secondary inspection and biometric examination identified her as a 26-year-old citizen of Cameroon, not the United States
China’s latest AI champion says technical standards needed for healthy development of facial recognition industry – Oct. 2, 2018 (South China Morning Post)
- Xu Li said his encounter with bureaucracy highlights a critical issue facing the AI industry today – the lack of technology standards even as the country fully embraces AI-driven identity checks. With standards, technology adopters can better understand the risk involved, just like credit worthiness for individuals and companies,” Xu said. “Providers of facial recognition can be assigned different trust levels, ranging from financial security at the top to entertainment uses.”
- SenseTime is the latest company to be hand-picked by Beijing to spearhead the country’s innovation efforts in intelligent vision. The other national AI champions are Baidu for autonomous driving, Alibaba Group for smart city initiatives, Tencent Holdings for computer vision in medical diagnosis, and iFlyTek for speech recognition. (Alibaba is the parent company of the South China Morning Post)
Events / other:
Q&A: Andrew Ng, the Authority on A.I. – Oct. 6, 2018 (Fortune)
- NG, A 42-YEAR-OLD COMPUTER SCIENCE professor at Stanford University, made his name leading artificial intelligence efforts at two of the world’s biggest tech companies, Google and Baidu. Last year he suddenly left Baidu, and over the next many months he launched three high-profile A.I. initiatives of his own: a series of online A.I. courses called deeplearning.ai, a business called Landing AI that develops artificial intelligence for manufacturing companies, and an incubator for startups known as the AI Fund
Alibaba Cloud, Zhejiang Lab and BML Munjal University Host their First Natural Language Processing (NLP) Hackathon in India – Oct. 5, 2018 (India Education Diary)
- Alibaba Cloud, the cloud computing arm of Alibaba Group, together with Zhejiang Lab and BML Munjal University (BMU), concluded its first Natural Language Processing (NLP) Hackathon in India on 29 September. More than 100 students from BMU participated in the day-long event
- During the hackathon, students solved NLP problems and interacted with experts from Zhejiang Lab and Alibaba Cloud to exchange ideas on artificial intelligence and network information. The organisers hoped that through the hackathon, students would be able to adapt their career aspirations with what jobs might be able to offer to them in future