What’s happened in AI: August 27th-September 3rd

By | September 4, 2018

Busy week leading into Labor Day weekend. Highlights are centered around the autonomous vehicle world with one of Apple’s self-driving cars getting rear-ended, and Toyota investing $500mm into Uber.

Very interesting move by Toyota given their own autonomous vehicle program and Uber’s have struggled compared to peers. Maybe this partnership provides synergies that aren’t immediately obvious and propels them forward. The obvious synergy is that Toyota has a new customer locked in for its vehicles. Can’t go wrong in that aspect.

Company developments:

Apple Lexus RX 450h self-driving test car rear-ended during road testing in Silicon Valley – Aug. 31, 2018 (ABC)

  • On August 24, one of Apple’s Lexus RX 450h self-driving test vehicles in “autonomous mode” was merging south on the Lawrence Expressway in Sunnyvale, California at less than 1.6kph when it was rear-ended by a 2016 Nissan Leaf going about 24kph, according to the report posted on the California Department of Motor Vehicles website. Both vehicles sustained damage but there were no injuries, the report said
  • Apple executives have never publicly spoken about the company’s self-driving car program, but filings in a criminal court case last month confirmed the company had at least 5,000 employees working on the project and it was working on circuit boards and a “proprietary chip” related to self-driving cars

Hail an autonomous taxi: Yandex rolls out ‘Europe’s first’ self-driving cabs in Russia – Aug. 30, 2018 (ZD Net)

  • Russian technology giant Yandex has begun to test self-driving taxis in Innopolis, a small town built a few years ago to foster education in technology and entrepreneurship. About 100 residents have already signed up for the pilot scheme, with the first rides in two vehicles taking place this week
  • At the moment, passengers who hail a taxi can travel free of charge between five standard locations in Innopolis, including the university, a local business center, and a residential area. The average ride is just over 2km or 1.25 miles. The two autonomous cars, Toyota Prius models equipped with Yandex technology, detect objects around them with centimeter precision and compare them against preloaded maps. Yandex told ZDNet that its taxis run in all weather and traffic conditions. However, for the time being they will have a safety engineer in the passenger seat

Diffbot launches AI-powered knowledge graph of 1 trillion facts about people, places, and things – Aug. 30, 2018 (VentureBeat)

  • Leveraging a combination of computer vision and natural language processing, Diffbot’s web crawler can parse the layout and structure of virtually any webpage — about 90 percent of the web and 20 or so page types, Tung claims — for facts, figures, and abstract relationships between objects. (Typical examples include a product page on Amazon.com or an executive bio on a company’s webpage)
  • Data extracted by Diffbot’s crawler feeds into an enormous database called the Diffbot Knowledge Graph, or DKG, comprising more than a trillion facts and 10 billion entities. (Tung said it’s adding facts at a rate of 130 million per month.) Core categories include people (skills, employment history, education, social profile), companies, locations (mapping data, addresses, business types, zoning information), articles (every news article, dateline, byline from anywhere on the web, in any language), discussions (chats, social sharing, and conversations), and images (organized using image recognition and metadata collection)

JPMorgan just made another high-profile artificial intelligence hire – Aug. 28, 2018 (CNN Money)

  • Apoorv Saxena, a Google executive in charge of cloud-based AI products, joins the bank later this week as head of artificial intelligence and machine learning services. The role also places him in charge of asset and wealth management artificial intelligence technology, according to an internal memo seen by CNNMoney
  • Saxena, who will work in Palo Alto, is the company’s second high-profile AI hire in recent months. The company named Manuela Veloso, who led the machine learning department at Carnegie Mellon University, its first head of artificial intelligence research in May

‘I hate them’: Locals reportedly are frustrated with Alphabet’s self-driving cars – Aug. 28, 2018 (CNBC)

  • Alphabet’s self-driving cars are said to be annoying their neighbors in Arizona, where Waymo has been testing its vehicles for the last year. More than a dozen locals told The Information they they hated the cars, which often struggle to cross a T-intersection near the company’s office
  • Waymo has been testing its vehicles in the Phoenix suburbs for little more than a year and is widely seen as the furthest along in the self-driving car space, but its safety drivers have to take control of the vehicles regularly, people with direct knowledge of the issues tell The Information

Fundraising / investment:

Phrasee Lands $4M Series A Funding to Take AI-Powered Copywriting to America – Aug. 30, 2018 (OA Online)

  • Phrasee, the London-headquartered marketing technology company that uses artificial intelligence (AI) to generate optimised marketing copy, has closed a $4m Series A funding round. The investment will fuel Phrasee’s international growth plans, including scaling up its new US operation in San Francisco and the recently announced expansion of its UK team
  • Albion Capital Group LLP led the round, with follow-on funds from existing investors including global digital communications business Next Fifteen PLC (Next 15) and angel syndicate Galvanise Capital. Next 15 and Galvanise announced their initial investments in 2016

Bark raises $9 million for AI that keeps kids safe online – Aug. 29, 2018 (VentureBeat)

  • On the eve of its third birthday, the Atlanta-based startup is announcing a $9 million funding round led by Signal Peak Ventures, with Two Sigma Ventures and participation from existing investors, including Symmetrical Ventures, Fuel Capital, Hallett Capital, and Atlanta Seed Company. Bark will use the capital to “broaden its product offering” and “establish a presence in new markets,” Bason explained
  • Bark, which works on iOS, Android, and Amazon Fire Tablet devices, monitors kids’ communications in real time, automatically alerting parents to signs of depression, suicidal thoughts, cyberbullying, and adult content. The extensive list of supported apps includes Slack, Facebook, Twitter, WhatsApp, YouTube, Reddit, Pinterest, OneDrive, Kik, Houseparty, GroupMe, Google Drive, Flickr, Dropbox, Ask.Fm, and even Snapchat. And that’s all in addition to platform-level features that let parents monitor device location, photos, videos, and browsing history

Toyota to invest $500m in Uber in driverless car deal – Aug. 27, 2018 (BBC)

  • According to a press release issued by the firms, self-driving technology from each company will be integrated into purpose-built Toyota vehicles. The fleet will be based on Toyota’s Sienna Minivan model with pilot trials beginning in 2021
  • Shares in Toyota spiked at reports of the deal. Not surprising. Many analysts think personal car ownership will drop dramatically when the self-driving, ride-sharing future is fully upon us – with major companies instead purchasing enormous fleets of vehicles. Toyota, then, may have just secured its biggest ever customer

Google releases open source reinforcement learning framework for training AI models – Aug. 27, 2018 (VentureBeat)

  • Reinforcement learning — an artificial intelligence (AI) technique that uses rewards (or punishments) to drive agents in the direction of specific goals — trained the systems that defeated Alpha Go world champions and mastered Valve’s Dota 2. And it’s a core part of Google subsidiary DeepMind’s deep Q-network (DQN), which can distribute learning across multiple workers in the pursuit of, for example, achieving “superhuman” performance in Atari 2600 games. The trouble is, reinforcement learning frameworks take time to master a goal, tend to be inflexible, and aren’t always stable
  • “Inspired by one of the main components in reward-motivated behavior in the brain and reflecting the strong historical connection between neuroscience and reinforcement learning research, this platform aims to enable the kind of speculative research that can drive radical discoveries,” Pablo Samuel Castro and Marc G. Bellemare, researchers on the Google Brain Team, wrote in a blog post. “This release also includes a set of colabs that clarify how to use our framework.”

Chinese AI startups raised $5B in VC funding last year, outpacing the US – Aug. 27, 2018 (Tech Republic)

  • Artificial intelligence (AI) startups in China raised nearly $5 billion in venture capital (VC) funding in 2017, overtaking their US counterparts in yet another part of the AI race, according to a Monday report from ABI Research
  • AI startups in the US raised $4.4 billion from 155 investments. Meanwhile, Chinese startups raised $4.9 billion from only 19 investments, indicating a focus on mature AI applications with strong commercial viability and successful use cases, the report noted

Partnerships:

Google and Harvard use AI to predict earthquake aftershocks – Aug. 29, 2018 (VentureBeat)

  • Researchers from Google’s AI division and Harvard University have created an AI model capable of predicting the location of aftershocks up to one year after a major earthquake. The model was trained on 199 major earthquake events, followed by 130,000 aftershocks, and was found to be more accurate than a method used to predict aftershocks today
  • “We found that after feeding these model stress changes into the neural network, the neural network could sort of predict aftershock locations in the testing dataset more accurately than the sort of baseline Coulomb failure stress change criterion that’s used a lot in studies of aftershock locations,” Phoebe DeVries of the Department of Earth and Planetary Sciences at Harvard University told VentureBeat in a phone interview

Cheetah Mobile and OrionStar release GreetBot AI receptionist robots – Aug. 29, 2018 (VentureBeat)

  • The companies have engaged in sales agreements with 13 companies across 10 industries to deploy the robots as greeters at company and hotel front desks
  • Cheetah Mobile chairman and CEO Fu Sheng, along with leaders from China’s largest shared workspace company, Ucommune; office building services company Sunrise Technologies; travel app developer Etoury; and robot-based hotel services provider New Human Technologies attended the order-signing ceremony at the 2018 World Robot Conference (WRC) in Beijing on August 17

Amazon Web Services boosts machine learning to treat depression – Aug. 29, 2018 (Healthcare IT News)

  • Pharmaceutical company Takeda and research and development data science institute ConvergeHEALTH by Deloitte have partnered to study patient datasets to better understand the etiology, progression and most effective therapies for difficult diseases
  • Using insurance claims information including diagnoses, medical procedures and prescriptions, they ran linear and non-linear models on disease datasets like treatment-resistant depression. The goal was to identify data factors with the highest impact on predicting patient outcomes. By combining the right data and the right questions, the organizations improved the predictability of deep learning models, allowing for the analysis of wider and more complex data sets and a better understanding of patient trajectories

AutoX is using its self-driving vehicles to deliver groceries – Aug. 27, 2018 (TechCrunch)

  • Autonomous vehicle startup AutoX has launched a grocery delivery and mobile store pilot in a partnership with GrubMarket.com and local high-end grocery store DeMartini Orchard
  • The pilot will initially be limited to an area of about 400 homes in north San Jose. The company, which employees nearly 90 people, has just two autonomous vehicles that will be used for the initial launch. Eventually, AutoX aims to expand the pilot west to Mountain View and Palo Alto, with more delivery partners joining soon

Research / studies:

Auckland researchers make world-first discovery into artificial intelligence – Sept. 3, 2018 (NZ Herald)

  • In a world first, AUT researchers have developed an artificial intelligence model, which can predict a person’s choices before they have even made up their mind. The work is based on a new type of artificial intelligence research called spiking neural networks, which was used to develop NeuCube, a machine learning system modelled on how the human brain learns and recognises patterns
  • In experiments, the sisters got 20 participants to watch a video of different beverage logos and recorded their brain data using an EEG headset. That data was synced to the NeuCube algorithm, which learned and classified patterns from the participants’ brains. It was able to predict their beverage choice 0.2 seconds before they consciously perceived the beverage. It also showed a clear difference between logos which were familiar to participants and those which weren’t

Artificial intelligence used to predict cancer growth – Sept. 1, 2018 (BBC News)

  • A team led by the Institute of Cancer Research London (ICR) and the University of Edinburgh developed a new technique known as Revolver (Repeated Evolution of Cancer). This picks out patterns in DNA mutation within cancers and uses the information to forecast future genetic changes. The researchers said the ever-changing nature of tumours was one of the biggest challenges in treating cancer – with cancers often evolving to a drug-resistant form
  • Researchers found that breast tumours which had a sequence of errors in the genetic material that codes for the tumour-suppressing protein p53, followed by mutations in chromosome 8, survived less time than those with other similar trajectories of genetic changes

Baidu researchers develop a new auto-tuning framework for autonomous vehicles – Aug. 29, 2018 (Tech Xplore)

  • Researchers at Chinese multinational tech company Baidu have recently developed a data-driven auto-tuning framework for self-driving vehicles based on the Apollo autonomous driving platform. The framework, presented in a paper pre-published on arXiv, consists of a new reinforcement learning algorithm and an offline training strategy, as well as an automatic method of collecting and labelling data
  • “Motion planning for autonomous driving cars has a lot of challenging issues,” Fan Haoyang, one of the researchers who carried out the study, told Tech Xplore. “One main challenge is that it has to deal with thousands of difference scenarios. Typically, we define a reward/cost functional tuning that can adapt those differences in scenarios. However, we find it is a difficult task.”

Government / policy:

TSA and border officers are testing new facial recognition technology at LAX – Aug. 31, 2018 (LA Times)

  • The 30-day pilot program, a partnership between the Transportation Security Administration and U.S. Customs and Border Protection, is one of several new efforts by the two federal security agencies to use state-of-the-art technology to speed up and bolster the airport screening process
  • Under the program, cameras snap photos of travelers as they prepare to be screened for international flights out of LAX. The photos are optional now but may be mandatory for foreign nationals if the technology is retained after the test. The cameras use facial recognition technology to match up the faces of departing travelers with data collected by government agencies of each foreign national who enters the country

New NEC algorithm dramatically improves accuracy of South Wales Police facial recognition – Aug. 28, 2018 (Biometric Update)

  • South Wales Police have deployed a new algorithm to make their automated facial recognition (AFR) system much more accurate, making only ten false matches out of 44.468 face scanned during the Biggest Weekend event in Swansea, WalesOnline reports after receiving data through a Freedom of Information request
  • Further, two alerts were issued for matches confirmed by the operator, with one person arrested for an outstanding warrant. The match was against one or more of three watch lists created by police specifically for the event. One list was based on suspected organized crime groups associated with music festivals, one was based on individuals wanted by South Wales Police, and one for wanted suspects from the specific area

Events:

KanKan AI Platform Receives Top Ranking at Innovation and Entrepreneurship Competition in Sichuan, China – Aug. 27, 2018 (Business Insider)

  • Remark Holdings, Inc. (NASDAQ: MARK), a diversified global technology company with leading artificial intelligence (AI) solutions and digital media properties, including Vegas.com, today announced that KanKan AI placed third in the Electronic Information category at the 7th Annual China Innovation and Entrepreneurship Competition in the Sichuan region