GM’s Cruise is launching the Origin electric autonomous vehicle and sliding doors today, the product of 3 years of work between General Motors, Cruise, and Honda. The Cruise Origin is designed without a combustible engine or a driver, decisions which Cruise CEO Dan Ammann says give people more space. The Origin can seat up to 6 people, with 3 passengers on each side looking toward each other in the rear-wheel drive vehicle.The vehicle also comes with air bags, “Start Ride” buttons, an SOS button, and a camera on the roof of the interior, presumably to deploy computer vision models that can do things like analyze the sentiment of riders. VentureBeat reached out to Cruise for more details on how each of these interior features works.

IBM formally announced the IBM Policy Lab — an initiative aimed at providing policymakers with recommendations for emerging problems in technology — ahead of a panel discussion to be held tomorrow at the World Economic Forum in Davos. The panel will be hosted by IBM CEO Ginni Rometty, with Siemens CEO Joe Kaeser, White House Deputy Chief of Staff for Policy Coordination Chris Liddell, and OECD Secretary General Angel Gurría. IBM also outlined a set of priorities for AI regulation, including several aimed at compliance and explainability.The Policy Lab — which soft-launched in November 2019 — serves as a forum for establishing a “vision” and actionable suggestions to “harness the benefits of innovation while ensuring trust,” according to press materials published this morning. It is under the leadership of codirectors Ryan Hagemann, a former senior policy fellow at the International Center for Law and Economics at the Niskanen Center, and Jean-Marc Leclerc, who currently vice-chairs the American Chamber of Commerce to the European Union’s Digital Economy Committee and chairs the Software Alliance’s Europe, Middle East, and Africa Policy Committee. To execute its mandate, the think tank convenes stakeholders and leaders in public policy, academia, civil society, and tech to formulate ideas to help tackle global challenges.

Microsoft Research AI today said it plans to open-source an optimized version of Google’s popular BERT natural language model designed to work with the ONNX Runtime inference engine. Microsoft uses to the same model to lower latency for BERT when powering language representation for the Bing search engine. The model, which “delivers its largest improvement in search experience” for Bing users, was detailed in a post last fall.This means developers can deploy BERT at scale using ONNX Runtime and an Nvidia V100 GPU with as little as 1.7 milliseconds in latency, something previously only available in production for large tech companies, a company spokesperson told VentureBeat in an email.Microsoft joined Facebook to create ONNX in 2017 to fuel interoperability across AI hardware like semiconductors and software like machine learning frameworks. The BERT-optimized tool joins a number of ONNX Runtime accelerators like one for Nvidia TensorRT and Intel’s OpenVINO. Using the ONNX standard means the optimized models can run with PyTorch, TensorFlow, and other popular machine learning models.

Equipping an AI self-driving car to successfully navigate a car wash is and edge priority for developers, but would provide value to car owners.

Platforms that let shoppers virtually try on cosmetics, apparel, and accessories have exploded in popularity over the past decade, and it’s easy to see why. According to a survey conducted by banking company Klarna, 29% of shoppers prefer to browse for items online before actually buying them, while 49% are interested in solutions that take their measurements so they can be sure something will fit before buying.With this top of mind, a team of researchers hailing from Adobe, the Indian Institute of Technology, and Stanford University explored what they describe as an “image-based virtual try-on” for fashion. Called SieveNet, it’s able to retain the characteristics of an article of clothing (including wrinkles and folds) as it maps the item to virtual bodies — without introducing blurry or bleeding textures.

Will the proliferation of AI and machine learning reinforce the worldwide digital divide? It’s one of the questions the Global Talent Competitiveness Index (GTCI) and Global Cities Talent Competitiveness Index (GCTCI) seek to answer by benchmarking the ability of countries and cities to compete for talent. An answer has historically proven elusive, but the 7th annual reports published by Insead, Adecco Group, and Google suggest it might instead provide “significant” opportunities despite the fact that AI skills are “scarce” and “unequally distributed” across nations.“AI is changing many facets of business and society and, if properly used and governed, has potential to foster sustainable development,” said Katell Le Goulven, executive director of the Insead Hoffmann Global Institute for Business and Society. “The GTCI report argues that with multi-stakeholder cooperation the technology could help achieve some of the SDGs [the United Nations’ Sustainable Development Goals] such as those related to health (via personalized remote diagnosis and big data analysis to track and reduce endemic disease). But it also points to the imperative of closing the global digital skills gap to harness the potential of AI for good.”

Antarctica’s emperor penguin population has suffered such severe climate-related breeding issues that it’s at risk of disappearing by the year 2100, according to a 2019 study by the British Antarctic Survey. In search of a solution, an Intel-led group of tech companies developed a computer vision solution to help ecologists count the remaining penguins faster and more accurately than before.The coalition consists of Intel’s AI Builders — the company’s curated ecosystem of software vendors — along with Microsoft’s AI for Earth initiative and data science consultancy Gramener. The consultancy sourced a corpus containing photos of Antarctica’s penguin colonies from Oxford University’s Penguin Watch Project, which over the past decade has sourced millions of time-lapse images from camera traps in over 40 locations and recruited online volunteers to annotate them. It then fed the data through a convolutional neural network —  a type of AI model most commonly applied to analyzing visual imagery — that preserved spatial information while localizing penguin counts and estimating overall tallies.

Resampling methods are designed to add or remove examples from the training dataset in order to change the class distribution. Once the class distributions are more balanced, the suite of standard machine learning classification algorithms can be fit successfully on the transformed datasets. Oversampling methods duplicate or create new synthetic examples in the minority class, …

Amid the start of an impeachment trial; talk of mounting hostility with Iran; new trade deals with China, Canada, and Mexico; and the final presidential debate before the start of the Democratic presidential primary season, you might’ve missed it, but it was also a momentous week for facial recognition regulation.A bipartisan group in Congress wants action, roughly a dozen state governments are considering legislation, and overseas news broke Thursday that the European Commission is considering a five-year moratorium on facial recognition among potential next steps. This would make the EU the largest government worldwide to halt deployment of the technology.In Washington, DC this week, the House Oversight and Reform Committee pledged to introduce legislation in the “very near future” that could regulate facial recognition use by law enforcement agencies in the US. Just like in hearings held last summer, members of Congress exhibited a fairly unified, bipartisan position that facial recognition use by the government should be regulated and in some cases limited. There was talk of regulation, but until this week, the future of sweeping facial recognition regulation seemed uncertain.

Supplier discovery often isn’t a walk in the park. According to a survey by Thomasnet, nearly 50% of buyers have worked with a supplier that unexpectedly went out of business. That’s why in 2015, four German entrepreneurs — Christian Heinrich, Fabian Heinrich, Gregor Stühler, and Lee Galbraith — founded Scoutbee, which develops and sells access to an AI-driven supplier discovery platform. After raising $12 million in June 2019 to lay the runway for future growth, the company this week closed a $60 million series B round led by Atomico with participation from Lakestar and Siemens-backed venture firm Next47.CEO Stühler said the fresh capital, which brings Scoutbee’s total raised to $76 million, will bolster R&D efforts as the company explores strategic acquisitions. Additionally, it will enable Scoutbee to expand its engineering, AI and machine learning, sales and marketing, and product development departments from 160 total staff members to around 260 across offices in Berlin and Washington, D.C. by 2021.

(Reuters) — The EU’s proposal for a temporary ban on facial recognition technology won backing from Alphabet Chief Executive Sundar Pichai on Monday but got a cool response from Microsoft President Brad Smith. While Pichai cited the possibility that the technology could be used for nefarious purposes as a reason for a moratorium, Smith said a ban was akin to using a meat cleaver instead of a scalpel to solve potential problems.“I think it is important that governments and regulations tackle it sooner rather than later and give a framework for it,” Pichai told a conference in Brussels organized by think-tank Bruegel. “It can be immediate but maybe there’s a waiting period before we really think about how it’s being used,” he said. “It’s up to governments to chart the course” for the use of such technology.

Google CEO Sundar Pichai has called for sensible AI regulation that does not limit the huge potential benefits to society. Writing in a FT editorial, Pichai said: “…there is no question in my mind that artificial intelligence needs to be regulated. It is too important not to.” Few people debate the need for AI regulation […] The post Google CEO: We need sensible AI regulation that does not limit its potential appeared first on AI News.

We’ve all been there: You’re snapping pics with your phone — perhaps of a high-speed bike ride or of a hockey match — and don’t think to check whether the autofocus is in lockstep with the action. It isn’t, as you later discover, and you’re stuck with a gallery of unusably blurry photos.In search of a solution, scientists at the Inception Institute of Artificial Intelligence in the United Arab Emirates, the Beijing Institute of Technology, and Stony Brook University developed an AI system that removes blur from images in post-production. They note in a paper that it’s human-aware, meaning it’s able to deblur human faces, and that it performs “favorably” against state-of-the-art motion deblurring methods.

ServiceNow today announced that it has acquired Loom Systems to strengthen automated insights for customers using the Now platform in order to help them quickly fix IT issues. Loom Systems CEO Gabby Menachem and some members of the Loom Systems team will join ServiceNow as part of the deal. The financial terms of the sale were not disclosed.Loom Systems tech will be integrated with ServiceNow’s IT Service Management (ITSM) and IT Operations Management (ITOM) solutions.Based in Israel, Loom Systems can analyze software log data across applications in the cloud or on premises to help organizations manage operations. The resulting data can be used for anomaly detection, recommend task remediation in plain language, automate the creation of incident reports, or automate tasks with ServiceNow workflows.

Measuring foot traffic in real time is the pursuit of countless businesses seeking greater insight into buyer behavior. Startups like PlaceIQ, Factual, GroundTruth, Skyhook, and others have risen up to meet the need — they occupy a market that’s estimated to be worth $22.8 billion by 2024. A somewhat newer entrant is Placer.ai, which was cofounded in 2016 by Noam Ben-Zvi, Oded Fossfeld, Ofir Lemel, and Zohar Bar-Yehuda. It might not rival location data giants like FourSquare and ThinkNear, but it managed to nab $12 million this week in a funding round led by BV Capital with participation from Aleph, Reciprocal Ventures, OCA Ventures, existing investors, and an undisclosed group of new strategic investors.

Soft Robotics raised another $23 million to continue developing its solutions to the seemingly intractable problems gripping and sorting machines face.As work published by MIT and others has established, picker robots struggle with complex poses and unfamiliar objects. That’s because they not only have to locate objects and understand how to grasp them, but because they’ve got to set them down such that they don’t sustain damage or disturb their surroundings. Truly versatile picker robots could transform warehouses in industries from ecommerce to manufacturing.CEO Carl Vause said this capital infusion, which comes after Soft Robotics partnered with industrial giant FANUC to integrate its systems with the latter’s robots through a new controller, will accelerate the startup’s next phase of growth. “Variability is the kryptonite of the robotics industry,” he added. “By offering a system that is able to grasp and manipulate items that vary in size, shape, and weight, we are able to solve the problem of high variability in both products and processes.”

Presented by Ople.aiIn recent years, AI has emerged as an undisputed competitive advantage, which means smart business leaders facing the digital transformation have some really familiar questions: How do you measure success? What questions do you need to ask to ensure that you’re making the right decisions right out of the gate? And most importantly, how do you implement AI in your business to deliver a positive impact?The answers to those questions all fall under one umbrella, essentially: Do you choose a centralized approach or a decentralized approach to embedding AI into your business? There are pros and cons to both, says Pedro Alves, founder and CEO at Ople.ai.The first step of any data science project is to define a problem within the business — often referred to as strategic problem formulation — and develop a data set for that problem. Once that is complete, you enter what Alves calls the data science technical chasm. The chasm is the point at which the data scientist finishes all the technical steps, such as feature engineering and algorithm selection, and steps back. Once the AI model is built, the end business user begins to use the solution.

Computer vision systems generally excel at detecting objects but struggle to make sense of the environments in which those objects are used. That’s because they separate observed actions from physical context — even those that do model environments fail to discriminate between elements relevant to actions versus those that aren’t (e.g., a cutting board on the counter versus a random patch of floor).That’s why a team of researchers from the University of Texas and Facebook AI Research investigated in a paper Ego-Topo, a technique that decomposes a space captured in a video into a topological map of activities before organizing the video into a series of visits to different zones. By reorganizing scenes into these “visits” as opposed to a series of footage, they assert, Ego-Topo is able to reason about first-person behavior (e.g., what are the most likely actions a person will do in the future?) and the environment itself (e.g., what are the possible object interactions that are likely in a particular zone, even if not observed there yet?).

LumApps has raised a $70 million round of venture capital as the Paris-based startup seeks to bring more AI and machine learning capabilities to its workplace collaboration platform.The company’s Social Intranet service is deployed internally by companies with the goal of connecting employees so they can share information and communicate across all departments. Social Intranet integrates with tools such as Google’s G Suite and Microsoft’s Office 365 to allow employees to work simultaneously on writing, scheduling, or planning projects either in the office or remotely on any device.“It’s crucial for companies today to connect and engage all of their employees, and that can only happen when they have a way to centralize all communications, workflows, and applications,” said Sébastien Ricard, CEO of LumApps, in a statement.

Industrial robots and warehouse automation are lucrative intermingling markets, as evidenced by Berkshire Grey. The Lexington, Massachusetts-based company, which combines AI and robotics to automate omnichannel fulfillment for retailers, ecommerce, and logistics enterprises, today announced that it has secured a mammoth $263 million in series B funding led by SoftBank. Khosla Ventures, New Enterprise Associates, and Canaan participated in the round. CEO Tom Wagner says the fresh capital will fuel the startup’s global expansion, acquisitions, and team growth.“With our intelligent robotic automation, our clients see faster and more efficient supply chain operations that enable them to address the wants of today’s savvy consumer,” said Wagner, who previously served as iRobot’s chief technology officer.

Ever heard of FlyHash? It’s an algorithm inspired by fruit flies’ olfactory circuits that’s been shown to generate hash codes — numeric representations of objects — with performance superior to classical algorithms. Unfortunately, because FlyHash uses random projections, it can’t learn from data. To overcome this limitation, researchers at Princeton, the University of San Diego, IBM Research, and the MIT-IBM Watson AI Lab developed BioHash, which applies “local” and “biologically plausible” synaptic plasticity rules to produce hash codes. They say that it outperforms previously published benchmarks for various hashing methods and that it could yield binary representations of things useful for similarity searches.

As part of an annual look at global AI investment trends, CB Insights today reported that AI startups raised a record $26.6 billion in 2019, spanning more than 2,200 deals worldwide. That’s compared to roughly 1,900 deals totaling $22.1 billion in 2018 and about 1,700 deals totaling $16.8 billion in 2017.The reported high recorded by CB Insights in the AI in Numbers report is in line with analysis by other organizations keeping an eye on investment in the AI ecosystem. The National Venture Capital Association earlier this month said that although overall venture capital spending took a dip last year, investors spent a record $18.4 billion on AI startups in the United States in 2019.With investment highest in fields like autonomous driving, drug research, finance, and facial recognition, the AI Index 2019 report released last month found more than $70 billion in global private investment in AI.

Containers — sets of self-contained executables, binary code, libraries, and configuration files — are ubiquitous in cloud computing. According to a survey by managed service provider Diamanti, cloud-native apps were the No. 1 container use case for 33% of organizations in 2019. That’s surely music to the ears of Santa Barbara, California-based Anchore, which Saïd Ziouani and Daniel Nurmi cofounded in 2016 to help organizations implement secure container-based workflows. The company today announced the close of a $20 million in series A financing led by SignalFire, bringing its total raised to around $30 million.“We created Anchore to empower developers to build secure software quickly and more efficiently,” said Anchore CEO Ziouani, who said the fresh capital will be put toward R&D. “Given our progress to date backed by the strong adoption this past year, we’re now ready to scale our sales, professional services, and engineering resources globally to meet the need and achieve our mission to bring the benefits of container workflow security to customers around the world.”

Machine learning algorithms aren’t just technological novelties relegated to tasks like picking out faces in crowded places. In the enterprise, they can surface patterns and relationships that would otherwise have been missed. To do just that, Outlier.ai’s business analysis platform extracts data from internal and external sources and analyzes it to spot critical changes in behavior.Investors see potential — a year after experiencing 400% growth, Oakland, California-based Outlier today announced that it has raised $22.1 million in a series B funding round led by Emergence, with participation from existing investors Ridge Ventures, 11.2 Capital, First Round Capital, Homebrew, Susa Ventures, and SV Angel, bringing its total raised to over $30 million.

Snyk, a cybersecurity platform that helps developers find vulnerabilities in their open source applications, has raised $150 million in a round of funding led by New York-based private equity firm Stripes, with participation from Salesforce Ventures, Coatue, Tiger Global, BoldStart, Trend Forward, and Amity.This takes Snyk’s total funding to $250 million from backers including Alphabet’s GV and Accel, including a $22 million series B round in 2018 and a $70 million follow-on round just a few months ago. A Snyk spokesperson said that the company is now worth more than $1 billion, which is at least double the $500 million it was valued at back in September.Founded in 2015, London-based Snyk targets developers — rather than cybersecurity personnel — to help them find and fix flaws in their source code, as well as their containers and Kubernetes applications.

Social networks including Facebook, Twitter, and Pinterest tap AI and machine learning systems to detect and remove abusive content, as does LinkedIn. The Microsoft-owned platform — which has over 660 million users, 303 million of whom are active monthly — today detailed its approach to handling profiles containing inappropriate content, which ranges from profanity to advertisements for illegal services.As software engineer Daniel Gorham explained in a blog post, LinkedIn initially relied on a block list — a set of human-curated words and phrases that ran afoul of its Terms of Service and Community Guidelines — to identify and remove potentially fraudulent accounts. However, maintaining it required a significant amount of engineering effort, and the list tended to handle context rather poorly. (For instance, while the word “escort” was sometimes associated with prostitution, it was also used in contexts like a “security escort” or “medical escort.”)

More AI automation, natural language processing and trust seen by IBM AI Research in its Predictions for 2020.

One of the trends that came into sharp focus in 2019 was, ironically, a woeful lack of clarity around AI ethics. The AI field at large was paying attention to ethics, creating and applying frameworks for AI research, development, policy, and law, but there was no unified approach. The committees and groups, from every sort of organization related to AI, that sought to address AI ethics were coming up with their own definitions (or falling apart with nothing to show for their efforts). But working out ethics in AI is not just a feel-good endeavor — it’s critical to helping lawmakers create just policies and laws and guiding the work of scholars and researchers. It also helps businesses stay in compliance and avoid costly pitfalls, know where they should and should not invest their resources, and how to apply AI to their products and services. Which is to say, there’s a profound humanity to it all.

Apple has acquired Seattle-based edge AI experts Xnor.ai for a reported $200 million. If you recognise Xnor.ai, it’s likely because the company’s technology once powered the person-detection feature on Wyze’s popular cameras. Xnor.ai abruptly cancelled their contract with Wyze back in November – and now we know why. In layman’s terms, edge computing means the […] The post Apple buys edge AI experts Xnor.ai for a reported $200 million appeared first on AI News.

Deepfakes — media that takes a person in an existing image, audio recording, or video and replaces them with someone else’s likeness — are becoming increasingly convincing. In late 2019, researchers at Seoul-based Hyperconnect developed a tool (MarioNETte) that could manipulate the facial features of a historical figure, a politician, or a CEO using nothing but a webcam and still images. More recently, a team hailing from Hong Kong-based tech giant SenseTIme, Nanyang Technological University, and the Chinese Academy of Sciences’ Institute of Automation proposed a method of editing target portrait footage by taking sequences of audio to synthesize photo-realistic videos. As apposed to MarioNETte, SenseTime’s technique is dynamic, meaning it’s able to better handle media it hasn’t before encountered. And the results are impressive, albeit worrisome in light of recent developments involving deepfakes.

Can sensors see behind the corners of obstacles in real time? As it turns out, yes. A study by researchers at Stanford, Rice University, Princeton, and Southern Methodist University published in the journal Optica proposes a system that’s capable of producing around-the-bend images at high resolutions and speeds. It’s able to distinguish the submillimeter details of hidden objects from 1 meter away, and according to coauthor Felix Heide, it could be used to make out things like the license plates of hidden moving vehicles and personnel badges worn by walking individuals.“Non-line-of-sight imaging has important applications in medical imaging, navigation, robotics, and defense,” said Heide. “Our work takes a step toward enabling its use in a variety of such applications.”

Facial recognition is being used in Shanghai to help catch individuals suspected of abusing pharmaceuticals. South China Morning Post reports that Shanghai is testing facial recognition terminals in pharmacies that will verify a person’s identity prior to dispensing controlled substances. Some legal medications can be turned into banned drugs. Cold and allergy medications, for example, […] The post Shanghai uses facial recognition to help catch drug offenders appeared first on AI News.

AI was showcased in many areas at the recent CES trade show, including in manufacturing, robotics and in new apps.

An Airbus plane has successfully taken off using autonomous technology powered by computer vision systems, marking another step toward self-piloting flights.The company announced the tests today, but the flights actually occurred on December 18, 2019. The test crew included two pilots, two flight engineers, and a test flight engineer. The crew staged eight takeoffs during a four-hour period.Typically, a plane communicates with an Instrument Landing System (ILS), which relies on radio waves broadcast across the runway to guide the pilot during the takeoff process. Traditional airport infrastructure is required to initiate the flight.In this case, the aeronautics giant installed an image recognition system in the aircraft to allow it to take off without a pilot operating the controls or any need to communicate with the runway system.

The value of building data-driven businesses with AI at their core is well known today, and business executives are rushing to implement the technology into their operations and gain a competitive advantage, but it’s not as simple as creating a data lake and creating AI models.A large number of AI companies attempting to implement more AI models or build AI-first businesses have experienced challenges. A December 2018 PwC survey found that only 4% of businesses have successfully implemented AI.That’s why today the World Economic Forum released the AI toolkit for Boards of Directors. The toolkit shares guidance on how AI alters things like branding, operations, and company culture.The AI toolkit for Boards of Directors is being released ahead of the annual WEF meeting in Davos, Switzerland where the toolkit will be formally debuted next week. Work by the Singapore government to create a framework for companies using AI in Singapore will also be announced next week with Microsoft president Brad Smith.

Imbalanced classification involves developing predictive models on classification datasets that have a severe class imbalance. The challenge of working with imbalanced datasets is that most machine learning techniques will ignore, and in turn have poor performance on, the minority class, although typically it is performance on the minority class that is most important. One approach …

 Most discussions of AI infrastructure start and end with compute hardware — the GPUs, general-purpose CPUs, FPGAs, and tensor processing units responsible for training complex algorithms and making predictions based on those models. But AI also demands a lot from your storage. Keeping a potent compute engine well-utilized requires feeding it with vast amounts of information as fast as possible. Anything less and you clog the works and create bottlenecks.Optimizing an AI solution for capacity and cost, while scaling for growth, means taking a fresh look at its data pipeline. Are you ready to ingest petabytes worth of legacy, IoT, and sensor data? Do your servers have the read/write bandwidth for data preparation? Are they ready for the randomized access patterns involved in training?

Every member of the team needs to be aware of what is happening with AI self-driving cars, from small incidents to ones involving deaths.

Presented by IntelAI and privacy needn’t be mutually exclusive. After a decade in the labs, homomorphic encryption (HE) is emerging as a top way to help protect data privacy in machine learning (ML) and cloud computing. It’s a timely breakthrough: Data from ML is doubling yearly. At the same time, concern about related data privacy and security is growing among industry, professionals and the public.“It doesn’t have to be a zero-sum game,” says Casimir Wierzynski, senior director, office of the CTO, AI Products Group at Intel. HE allows AI computation on encrypted data, enabling data scientists and researchers to gain valuable insights without decrypting the underlying data or models. This is particularly important for sensitive medical, financial, and customer data.

Whether it’s language, music, speech, or video, sequential data isn’t easy for AI and machine learning models to comprehend — particularly when there’s dependence on extensive surrounding context. For instance, if a person or an object disappears from view in a video only to re-appear much later, many algorithms will forget how it looked. Researchers at Google set out to solve this with Transformer, an architecture that extended to thousand of words, dramatically improving performance in tasks like song composition, image synthesis, sentence-by-sentence text translation, and document summarization.But Transformer isn’t perfect by any stretch — extending it to larger contexts makes apparent its limitations. Applications that use large windows have memory requirements from gigabytes to terabytes in size, meaning models can only ingest a few paragraphs of text or generate short pieces of music. That’s why Google today introduced Reformer, an evolution of Transformer that’s designed to handle context windows of up to 1 million words. By leveraging techniques like locality-sensitive-hashing (LSH) and reversible residual layers to use memory efficiently and reduce complexity over long sequences, it’s able to run on a single AI accelerator chip using only 16GB of memory.

Facebook released PyTorch 1.4 today with upgrades to audio, vision, and text libraries; customization to the mobile version of its deep learning framework; and limited support of the Java programming language. PyTorch is one of the most popular machine learning frameworks used by researchers and developers today.In the latest version, developers can use fine-grain customization of build scripts with PyTorch Mobile. This is a way to optimize library sizes and greatly reduce the on-device footprint, and in early tests was able to generate a version of MobileNetV2 edge model up to 50% smaller.PyTorch 1.4 also includes a framework for distributed model parallel training and Java support for PyTorch inference based on the PyTorch Mobile for Android interface, but at launch the experimental feature is only available for Linux and for inference. PyTorch currently supports Python and C++.

All across the world, numerous research groups are attempting to build a quantum computer. Such a computer would be able to solve certain problems that cannot be solved with current classical computers, even if we combined all these computers in the world into one.At Aarhus University they share the ambition of building a quantum computer. For this reason, a research group under the direction of Professor Jacob Sherson has just used the computer algorithm AlphaZero to learn to control a quantum system.What makes AlphaZero interesting is that it can learn on its own without any form of human expertise. In this manner, AlphaZero has beaten both humans and specialized computer programs in games such as Go, Shogi, and Chess, and it learned to do so only by playing against itself. After just four hours of playing against itself, AlphaZero managed to beat the leading chess program Stockfish. AlphaZero was so superior that Danish grand master Peter Heine Nielsen compared the program to a superior alien species that had visited the earth just to beat us in chess.

A guide to finding a dataset for your data science project, getting it off the ground and tuning your AI education.

In October 2018, months after a brief reveal, Amazon brought Whisper Mode to select third- and first-party Alexa devices. It expanded the feature to all locales in November 2019, such that all smart home appliances powered by Alexa — the company’s virtual assistant — now respond to whispered speech by whispering back.Amazon was a bit light on the technical details initially, save that Whisper Mode uses a neural network — layers of mathematical functions loosely modeled after the human brain’s neurons — to distinguish among normal and whispered words. But in an academic paper appearing in the January 2020 issue of the journal IEEE Signal Processing Letters and an accompanying blog post, it detailed the research that led to the expansion.

No matter the claimed robustness of AI and machine learning systems in production, none are immune to adversarial attacks, or techniques that attempt to fool algorithms through malicious input. It’s been shown that generating even small perturbations on images can fool the best of classifiers with high probability. And that’s problematic considering the wide proliferation of the “AI as a service” business model, where companies like Amazon, Google, Microsoft, Clarifai, and others have made systems that might be vulnerable to attack available to end users.Researchers at tech giant Baidu propose a partial solution in a recent paper published on Arxiv.org: Advbox. They describe it as an open source toolbox for generating adversarial examples, and they say it’s able to fool models in frameworks like Facebook’s PyTorch and Caffe2, MxNet, Keras, Google’s TensorFlow, and Baidu’s own PaddlePaddle.

Some of the top investment firms in Silicon Valley and Seattle are joining up to put $10 million into the Allen Institute AI Incubator for early-stage AI startups. The news comes a day after Apple acquired Xnor. The edge computing AI startup was one of the first companies to receive financial support and mentorship from the incubator and was reportedly acquired for roughly $200 million.The $10 million fund is being supported by Sequoia Capital, Kleiner Perkins, Madrona Venture Group, and Two Sigma Ventures, alongside a cadre of entrepreneurs and executives, like Tableau CEO Adam Selipsky.This is the first external round of investment for the incubator, which opened in January 2017 with the goal of launching three to five AI startups a year in Seattle and giving them access to the Allen Institute’s 100 or so AI researchers. The funding round will raise the amount of cash the incubator can give AI startups from $250,000 to up to $800,000.

(Reuters) — The European Union is considering banning facial recognition technology in public areas for up to five years to give it time to work out how to prevent abuses, according to proposals seen by Reuters.The plan by the EU’s executive — set out in an 18-page white paper — comes amid a global debate about the systems driven by artificial intelligence and widely used by law enforcement agencies.The EU Commission said new tough rules may have to be introduced to bolster existing regulations protecting Europeans’ privacy and data rights.“Building on these existing provisions, the future regulatory framework could go further and include a time-limited ban on the use of facial recognition technology in public spaces,” the EU document said.

Vimeo has debuted a new AI-powered platform designed to help creatives produce shareable, short-form videos in minutes.This is part of a broader pivot by the New York-based video-streaming company, which is a subsidiary of IAC. Vimeo is working hard to ditch its “YouTube alternative” status and position itself as a place for creatives and businesses to access tools and make videos.Against this backdrop, Vimeo Create soft-launched in beta today to attract companies that want to create “high-impact” videos for their social channels.Underpinning the new platform is a stock video marketplace Vimeo launched back in 2018, with users able to search for relevant videos (or photos) by keywords. They can also upload original creations, of course.

Studying how people move to music is a powerful tool for researchers looking to understand how and why music affects us the way it does. Over the last few years, researchers at the Centre for Interdisciplinary Music Research at the University of Jyväskylä in Finland have used motion capture technology -- the same kind used in Hollywood -- to learn that your dance moves say a lot about you, such as how extroverted or neurotic you are, what mood you happen to be in, and even how much you empathize with other people.Recently, however, they discovered something that surprised them. "We actually weren't looking for this result, as we set out to study something completely different," explains Dr. Emily Carlson, the first author of the study. "Our original idea was to see if we could use machine learning to identify which genre of music our participants were dancing to, based on their movements."

AI researchers in healthcare are more cognizant of whether the data they need was collected in accordance with the best permission practices.

Resampling methods are designed to change the composition of a training dataset for an imbalanced classification task. Most of the attention of resampling methods for imbalanced classification is put on oversampling the minority class. Nevertheless, a suite of techniques has been developed for undersampling the majority class that can be used in conjunction with effective …

Managing multiple apps across cloud environments isn’t a walk in the park by any stretch. That’s why Ran Ribenzaft and Demisto veteran Nitzan Shapira cofounded Epsagon, which uses distributed tracing and agent-less code to give engineers greater visibility into apps. Two years after its founding, the company found success, more than quadrupling its customer roster in the past 12 months alone and expanding into verticals including internet of things,  technology, retail, media, insurance, smart mobility, and telecommunications.The momentum impressed investors like U.S. Venture Partners, evidently, which led the $16 million series A funding round that Epsagon announced this morning. Previous backers Lightspeed Venture Partners and StageOne Ventures also contributed to the tranche, which CEO Shapira says will fuel growth on the R&D, sales, and marketing sides of Epsagon’s business. “Cloud microservices have enabled companies to build applications faster than ever before; yet visibility into what’s working and what’s running in production remains a challenge,” he said. “[T]oday we’re continuously evolving our … platform to scale with the growing cloud microservices stack.”

The average person is distracted every 40 seconds when working in front of a computer. Indeed, even the most productive workers get only 11 minutes on average between interruptions. Increasingly, with experiments like Digital Wellbeing and Screen Time, tech giants are advancing usage limits as a potential remedy. But if you ask Mathias Mikkelsen, CEO and founder of Memory, our attention deficit calls for a uniquely digital solution.Available today, Dewo (pronounced “dee-woe”) is a “personal assistant for deep work” and the brainchild of the 45-person team behind Timely, Memory’s first product. Timely launched three years after Mikkelsen bootstrapped the company by selling his apartment. A productivity-tracking tool that taps AI to create up-to-the-minute timesheets on behalf of employees, Timely quickly found success, racking up thousands of paying customers across over 160 countries and attracting the attention of investors like Concentric, Investinor, 500 Startups, and SNÖ Ventures.

Presented by DataikuWhite-box AI is now getting heaps of attention. But what does it mean in practice? And how can businesses start moving away from black-box systems to more explainable AI? Learn why white-box AI brings business value, and how it’s a necessary evolution when you join this VB Live event.Register here for free.Black box AI has been getting some attention in the media for how it has produced undesirable, even unethical results. But the conversation is so much more complex, says Rumman Chowdhury, managing director at Accenture AI. When technologists or data scientists talk about black box algorithms, what they’re specifically referring to is a class of algorithms for which we don’t always understand how the output is achieved, or non-understandable systems.

Imbalanced Classification Crash Course. Get on top of imbalanced classification in 7 days. Classification predictive modeling is the task of assigning a label to an example. Imbalanced classification are those classification tasks where the distribution of examples across the classes is not equal. Practical imbalanced classification requires the use of a suite of specialized techniques, …

Face swapping is a category of deepfakes that extracts the faces of people in existing media and replaces them with other peoples’ features, typically with AI and machine learning. It’s been popularized by apps like MixBooth and SnapChat, and while the underlying techniques have enabled sophisticated image editing for legitimate purposes, they’ve also given rise to concerns about potential misuse or abuse.Various groups have compiled manipulated media to support the development of face swapping detection methods, but the samples that have released so far are relatively few in number or overly artificial. That’s why researchers from SenseTime Research, the R&D division of Hong Kong-based tech startup SenseTime, partnered with Nanyang Technological University in Singapore to design a new large-scale benchmark for face forgery detection. They call it DeeperForensics-1.0, and they say it’s the largest corpora of its kind with over 60,000 videos containing roughly 17.6 million frames.

Business-to-business buyers are typically 57% of the way to a buying decision before they engage with sales and marketing departments. Motivated by the idea that AI and automation might have a role to play in helping seal the deal, five entrepreneurs — Amanda Kahlow, Dustin Chang, Premal Shah, Shane Moriah, and Viral Bajaria — cofounded 6Sense in 2013. The account-based orchestration platform developer today announced it has raised $40 million in series C funding.This latest round — which was led by Insight Partners and brings 6Sense’s total raised to $105 million, following a $27 million funding round in April 2019 — comes after a year in which the San Francisco-based company doubled its customer base, with notable new customers including Box, Cisco, Dell, Zendesk, Sumo Logic, NetApp, Domo, Motorola, Cumulus, Symantec, and Tableau. CEO Zintak says the fresh capital will set the stage for growth in 2020.

Phenom People, a human resources (HR) platform that leverages artificial intelligence (AI) to help companies attract new talent, has raised $30 million in a series C round of funding led by WestBridge Capital, with participation from eBay founder Pierre Omidyar’s VC firm Omidyar Ventures, AXA Venture Partners, Sierra Ventures, Sigma Prime Ventures, Karlani Capital, and a fund belonging to AllianceBernstein.Founded in 2010, Philadelphia-based Phenom People touts its “talent experience management” (TXM) platform as an all-in-one solution for companies looking to build career websites with personalized job and content recommendations, chatbots, and a content management system (CMS) for pushing fresh content to the site.

Demis Hassabis founded DeepMind with the goal of unlocking answers to some of the world’s toughest questions by recreating intelligence itself. His ambition remains just that — an ambition — but Hassabis and colleagues inched closer to realizing it this week with the publication of papers in Nature addressing two formidable challenges in biomedicine.The first paper originated from DeepMind’s neuroscience team, and it advances the notion that an AI research development might serve as a framework for understanding how the brain learns. The other paper focuses on DeepMind’s work with respect to protein folding — work which it detailed in December 2018. Both follow on the heels of DeepMind’s work in applying AI to the prediction of acute kidney injury, or AKI, and to challenging game environments such as Go, shogi, chess, dozens of Atari games, and Activision Blizzard’s StarCraft II.

Apple has acquired AI startup Xnor.ai, a source familiar with the matter told VentureBeat today. Xnor focuses on the efficient deployment of AI in edge devices like smartphones, cameras, and drones. Apple reportedly paid roughly $200 million for the company, according to Geekwire. Xnor’s technology could be used to improve Apple’s deployment of AI stored locally on devices or power edge computing in Apple’s Core ML 3 toolkit for app developers.Interest in more efficient AI models has grown recently as a way to drive innovation, reduce costs, and reduce AI’s carbon footprint.Last year, Xnor demonstrated its edge computing capabilities by slapping a solar panel on the back of an FPGA chip to power a computer vision model that needs no battery, and in May 2019, Xnor launched AI2Go, a platform for state-of-the-art edge computing.

Workplace automation tools are expected to see an uptick in adoption in the next few years. One company leading the charge is San Francisco-based Zinier, which was founded in October 2015 by Andrew Wolf and former TripAdvisor market development manager Arka Dhar. A developer of intelligent field service automation, the startup provides a platform — intelligent service automation and control, or ISAC — aimed at fixing machinery before it breaks and maintaining mission-critical client services.After raising $30 million across three funding rounds, the first of which closed in January 2016, Zinier is gearing up for a major expansion with fresh capital. Today the startup announced that it has raised $90 million in series C funding — nearly quadruple its series B total — led by new investor Iconiq Capital, with participation from Tiger Global Management and return investors Accel, Founders Fund, Nokia-backed NGP Capital, France-based Newfund Capital, and Qualcomm Ventures. The round brings Zinier’s total raised to over $120 million, and CEO Dhar says it will support the company’s customer acquisition strategy and accelerate expansion of its services across telecom, energy, utilities, and beyond.

The voice and speech recognition tech market is anticipated to be worth $31.82 billion by 2025, driven by new applications in the banking, health care, and automotive industries. And Tom Livne, who cofounded Verbit.ai with Eric Shellef and Kobi Ben Tzvi in 2017, asserts the Tel Aviv and New York-based startup will contribute substantially to the segment’s rise. Verbit’s adaptive speech recognition tech, which can generate detailed transcriptions with a claimed over 99% accuracy, recently attracted the attention of VCs that include Vertex Ventures and Oryzn Capital, both of which participated in the startup’s $23 million series A round. Now, roughly a year later, Verbit hopes to lay the groundwork for growth with an infusion of fresh capital.

A bipartisan group in Congress is working on legislation that could regulate the use of facial recognition by the private sector, federal government, and law enforcement.“We have a responsibility to not only encourage innovation, but to protect the privacy and safety of Americans consumers,” Rep. Carolyn Maloney (D — NY) said today, while acknowledging a need to educate others in Congress and explore consumer privacy and data security protections currently in place.“In that vein, I would like to announce today that our committee is committed to introducing, and marking up common sense facial recognition legislation in the very near future,” said Maloney, chairwoman of the House Oversight and Reform Committee.

Medical imaging startup Nanox hopes to reinvent the X-ray with hardware inspired by Star Trek’s biobed. In anticipation of future growth, Nanox recently raised $26 million in a funding round led by strategic investor Foxconn, with participation from previous investors Fujifilm, SK Telecom, and others.The round brings Nanox’s total raised to $55 million, and founder and CEO Ran Poliakine says it will enable the company to pursue partnerships with governments, hospitals, and clinic chains. “We are honored to have Foxconn join other world leaders, Fujifilm and SK Telecom, in investing in our vision of eradicating cancer,” he added.Nanox was founded in 2016 by Japanese venture capital tycoon Hitoshi Masuya as part of a joint investment with Sony. After Sony dropped out, Masuya joined forces with Poliakine, and the two decided to split the company’s operations between Japan and Israel.