Archiv der Kategorie: Artificial Intelligence

Machine Learning macht Analysen besser

Kommentar verfassen

Zuerst analysierten lernende Maschinen das Nutzerverhalten in Suchmaschinen, um passende Werbung anzuzeigen. Heute optimieren sie Verkehrsflüsse, die Stahlherstellung und planen die Flugzeugwartung. Experten von Allianz, Trip Advisor, GfK und Boeing erklären, wie ihnen Machine Learning hilft.

http://www.computerwoche.de/a/machine-learning-soll-analysen-besser-machen,3217540

Bei der Münchener Allianz Versicherung ist Andreas Braun, Head of Global Data and Analytics, zufrieden mit den Ergebnissen seiner Experimente mit den neuen Analytics-Ansätzen aus der künstlichen Intelligenz. „Wir haben bei uns ein Ökosystem aus verschiedenen Bestandteilen im Einsatz. Big-Data-Technologien und Machine Learning bieten uns bessere Möglichkeiten, mit unseren Daten umzugehen, und liefern konsistent gute Ergebnisse“, sagte er auf der Konferenz der Yandex Data Factory zum Thema „Machine Learning and Big Data“ in Berlin. Zum Beispiel im Gebäude-Management: Zusammen mit Studenten der TU München hat die Versicherung eine App entwickelt, die eine Vielzahl von Gegenständen über Sensoren vernetzt.

„Das System kalibriert sich selbst, lernt normales Verhalten im Haus, und kann so einen Einbruch von anderen ungewöhnlichen, aber unkritischen Vorfällen unterscheiden.“ Außerdem wollen die Experten die Bilderkennung weiter verbessern. Eingereichte Fotos sollen bei Versicherungsschäden automatisch durch Maschinen beurteilt werden.

Die Experten, die der russische Suchmaschinen-Anbieter Yandex nach Berlin eingeladen hatte, tauschten sich unter dem Motto „Business Challenges“ auch über die Schwierigkeiten und Risiken rund um Machine Learning aus. Jeff Palmucci, Director of Machine Intelligence beim Reiseportal Trip Advisor, schilderte, wie sein Unternehmen maschinelles Lernen in die Geschäftsprozesse implementiert. So hilft die Technik, Restaurants und Hotels automatisiert mit passenden Tags wie „romantisch“ oder „charmant“ zu versehen, damit Suchende schnell das richtige Angebot finden. Auch um Betrug etwa bei den Bewertungen rasch zu erkennen, setzt das Portal Machine Learning ein.

Menschliches Verhalten vorhersagen

Machine Learning stellt Unternehmen vor vielfältige Herausforderungen. Nicht alle Branchen eignen sich gleich gut, erklärte Jane Zavalishina, CEO der Yandex Data Factory: „Es geht vor allem darum, menschliches Verhalten vorherzusagen.“ Bei Ergebnissen, die auf Machine Learning basieren, könne man aber durch die hohe Komplexität und die großen Datenmengen nie genau nachvollziehen, wie sie zustande gekommen sind. In der Praxis müsse man mit den Empfehlungen experimentieren, um herauszufinden, ob sie der bisherigen Vorgehensweise überlegen sind. Das gehe aus ethischen und praktischen Gründen allerdings nicht immer.

Jane Zavalishina CEO, Yandex Data Factory „Viele Unternehmen befinden sich aber noch an dem Punkt, an dem sie versuchen, Big Data Analytics überhaupt zu verstehen.“
Jane Zavalishina CEO, Yandex Data Factory „Viele Unternehmen befinden sich aber noch an dem Punkt, an dem sie versuchen, Big Data Analytics überhaupt zu verstehen.“
Foto: Yandex

In Echtzeit Web-Inhalte zu personalisieren oder Vorhersagen zu treffen, ist für die russische Suchmaschine Yandex nichts Neues. Das Wissen des Konzerns, das aus der Suchtechnik und dem kontextuellen Einspielen passender Werbung entstanden ist, und die dafür entwickelten Algorithmen stellt sie seit 2014 auch extern zur Verfügung. Zunächst probierte das Tochterunternehmen Yandex Data Factory, das Firmensitze in Moskau und Amsterdam unterhält, die Techniken maschinellen Lernens in der Wissenschaft aus – zum Beispiel, um Big-Data-Probleme des europäischen Kernforschungszentrums CERN zu lösen.

Inzwischen besprechen die Datenexperten mit Firmen, die viele Kunden und große Datenmengen haben, wie sich deren Services, Prozesse und Produkte ver­bessern lassen. „Die Anwendungsmöglichkeiten für maschinelles Lernen in Unternehmen sind fast unbegrenzt“, sagte Zavalishina. „Viele Unternehmen befin­den sich aber noch an dem Punkt, an dem sie versuchen, Big Data Analytics überhaupt zu verstehen.“

Eine der ersten Firmen, die Wissen und Technologie von Yandex nutzte, war die russische Straßenverwaltungsbehörde Rosavtodor, die Vorhersagen zur Verkehrsdichte und zu Unfällen benötigte. Im Stahlwerk Magnitogorsk Iron and Steel Works optimieren heute Algorithmen die Stahlproduktion. Zu wenige Zusätze ergeben eine schlechte Qualität, zu viele treiben die Kosten in die Höhe. Bisher nutzten die Stahlkocher für ihre Mischungsvorhersagen komplizierte Modelle. Yandex Data Factory verwendete zur Optimierung historische Daten aus den zurückliegenden zehn Jahren. Vergleichsweise einfach scheint es dagegen, mit Machine Learning Websites zu optimieren und Online-Werbung auszusenden.

Business ist datengetrieben

„Wir sind ein komplett datengetriebenes Business“, sagt Norbert Wirth, Global Head of Data and Science beim Marktforschungsinstitut GfK, „Machine-Learning-Algorithmen sind für uns ein Werkzeug im Kanon mit anderen, das aber für die Vorhersage und für Klassifizierungsprobleme zunehmend wichtiger wird.“ GfK nutzt es derzeit vor allem für die Analyse von Social-Media-Daten und um Marktanteile und Marktperformance vorherzusagen.

„Wir setzen es ein, wenn nicht die Frage nach dem Warum entscheidend ist, sondern die Qualität der Vorhersage“, so Wirth. Sind Aussagen über eine Marke tendenziell eher positiv oder negativ? Und um welche Themen geht es? Bei kleineren Datenbeständen könne man das noch selbst herausfinden, wird es jedoch umfangreicher, seien die Algorithmen „extrem spannend – und sie werden immer leistungsfähiger“. Das sei kein Hype, sagt der Marktforscher, „Machine Learning wird an Bedeutung zunehmen. Mit wachsender Computerpower kann man damit jetzt wirklich arbeiten.“ Die eine Sache sei ein toller Algorithmus, die andere, ob man die dafür nötigen Maschinen auch am Start habe.

In Zukunft werden Analysten laut Wirth zusätzliche Daten verwenden, um Algorithmen zu trainieren und die Modelle leistungsfähiger zu machen. „Es geht in die Richtung, im Analyseprozess mit mehreren Datenquellen zu arbeiten. Natürlich mit solchen, die auch legal genutzt werden dürfen.“ Data Privacy sei ein sehr wichtiges Thema rund um Machine Learning – aber auch die Stabilität und die Qualität der Daten.

Der Flugzeughersteller Boeing nutzt Machine Learning, um seine Services und die interne Produktion zu verbessern, berichtete Sergey Kravchenko, President Russia and CIS von Boeing. Das Flugzeug 787 verfüge über mehr als zehntausend mit dem Internet verbundene Sensoren, die den Mechanikern am Boden schon während des Fluges melden, wenn zum Beispiel eine Lampe oder eine Pumpe ausgetauscht werden muss. So können Fluggesellschaften ihre Wartungskosten reduzieren und im Betrieb effizienter arbeiten.

Boeing arbeitet mit Big Data und Machine Learning, um den Fluggesellschaften mit den während eines Flugs gesammelten Daten zu helfen, Treibstoffkosten zu senken und die Piloten bei schlechtem Wetter zu unterstützen. Nun werden die Daten auch in der Produktion verwendet, um etwa für bestimmte Prozesse die besten Ingenieure zu finden. Daten der Personalabteilung würden genutzt, um zu verstehen, wie die Lebensdauer und die Qualität der Flugzeuge mit dem Training und der Mischung der Menschen im Produktionsteam korrelieren. Gibt es bei Prozessen, die aufwendige Nacharbeiten erfordern, Zusammenhänge mit den bereitgestellten Werkzeugen oder mit dem Team? Kravchenko will mit Big-Data-Analysen den gesamten Zyklus von Design, Produktion und Wartung verbessern.

Ein neues Big-Data-Projekt ist die Flight Training Academy, die 2016 eröffnet werden soll. Hier werden Daten der drei Flugsimulatoren gesammelt und ausgewertet, um die Gestaltung des Cockpits und das Design der Flugzeug­software zu verbessern. Kravchenko will seinen russischen Kunden auch anbieten, in Zukunft Daten auszutauschen und sie gemeinsam auszuwerten.

Experten müssen zusammenpassen

Die Fertigungsindustrie stehe bei der Anwendung von Machine Learning – verglichen etwa mit Telcos und dem Handel – noch am Anfang. Sie werde aber schnell von ihnen und auch von Firmen wie Amazon und Google, lernen. Wer Erfolg haben wolle, müsse die besten Flugzeug- und IT-Experten zusammenbringen. Das Problem: „Die kommen von verschiedenen Planeten.“

Die Zusammenarbeit kann dennoch gelingen – wenn sich alle auf eine gemeinsame Terminologie einigen. „Die Datenexperten müssen etwas mehr von Flugzeugen und Airlines verstehen und die Flugzeugspezialisten mehr über Data Analytics lernen. Sie müssen sich die Werkzeuge teilen, sich gegenseitig vertrauen und ein gemeinsames Team aufbauen“, sagt der Flugzeugbauer. Ein weiteres Problem sei die Relevanz der Daten. „Hier muss die Industrie ihre riesigen Datenmengen anschauen und entscheiden, welche Daten wirklich wichtig sind, um bestimmte Probleme zu lösen. Das ist nicht einfach, dafür brauchen wir Zeit, Trial and Error, und wir müssen von anderen Branchen lernen.“ Die richtige Auswahl der Daten und die Interpretation der Ergebnisse seien dabei wichtiger als der Algorithmus selbst.

Google built its own computer chip specifically for running deep neural networks, called the Tensor Processing Unit, or TPU

Kommentar verfassen
Google server racks loaded with TPUsGOOGLE
GOOGLE OPERATES WHAT is surely the largest computer network on Earth, a system that comprises custom-built, warehouse-sized data centers spanning 15 locations in four continents. But about six years ago, as the company embraced a new form of voice recognition on Android phones, its engineers worried that this network wasn’t nearly big enough. If each of the world’s Android phones used the new Google voice search for just three minutes a day, these engineers realized, the company would need twice as many data centers.

At that time, Google was just beginning to drive its voice recognition services with deep neural networks, complex mathematical systems that can learn particular tasks by analyzing vast amounts of data. In recent years, this form of machine learning has rapidly reinvented not just voice recognition, but image recognition, machine translation, internet search, and more. In moving to this method, Google saw error rates drop a good 25 percent. But the shift required a lot of extra horsepower.

Rather than double its data center footprint, Google instead built its own computer chip specifically for running deep neural networks, called the Tensor Processing Unit, or TPU. “It makes sense to have a solution there that is much more energy efficient,” says Norm Jouppi, one of the more than 70 engineers who worked on the chip. In fact, the TPU outperforms standard processors by 30 to 80 times in the TOPS/Watt measure, a metric of efficiency.

GoogleChipInline.jpg

GOOGLE

A Neural Network Niche

Google first revealed this custom processor last May, but gave few details. Now, Jouppi and the rest of his team have a released a paper detailing the project, explaining how the chip operates and the particular problems it solves. Google uses the chip solely for executing neural networks, running them the moment when, say, someone barks a command into their Android phone. It’s not used to train the neural network beforehand. But as Jouppi explains, even that still saves the company quite a bit. It didn’t have to build, say, an extra 15 data centers.

The chip also represents a much larger shift in the world of computer processors. As Google, Facebook, Microsoft, and other internet giants build more and more services using deep neural networks, they’ve all needed specialized chips both for training and executing these AI models. Most companies train their models using GPUs, chips that were originally designed for rendering graphics for games and other highly visual applications but are also suited to the kind of math at the heart of neural networks. And some, including Microsoft and Baidu, the Chinese internet giant, use alternative chips when executing these models as well, much as Google does with the TPU.

The difference is that Google built its own chip from scratch. As a way of reducing the cost and improving the efficiency of its vast online empire, the company builds much of its own data center hardware, including servers and networking gear. Now, it has pushed this work all the way down to individual processors.

In the process, it has also shifted the larger market for chips. Since Google designs its own, for instance, it’s not buying other processors to accommodate the extra load from neural networks. Google going in-house even for specialized tasks has wide implications; like Facebook, Amazon, and Microsoft, it’s among the biggest chip buyers on Earth. Meanwhile the big chip makers—including, most notably, Intel—are building a new breed of processor in an effort to move the market back in their direction.

Focused But Versatile

Jouppi joined Google in late 2013 to work on what became the TPU, after serving as a hardware researcher at places like HP and DEC, a kind of breeding ground for many of Google’s top hardware designers. He says the company considered moving its neural networks onto FPGAs, the kind of programmable chip that Microsoft uses. That route wouldn’t have taken as long, and the adaptability of FPGAs means the company could reprogram the chips for other tasks as needed. But tests indicated that these chips wouldn’t provide the necessary speed boost. “There’s a lot overhead with programmable chips,” he explains. “Our analysis showed that an FPGA wouldn’t be any faster than a GPU.”

In the end, the team settled on an ASIC, a chip built from the ground up for a particular task. According to Jouppi, because Google designed the chip specifically for neural nets, it can run them 15 to 30 times faster than general purpose chips built with similar manufacturing techniques. That said, the chip is suited to any breed of neural network—at least as they exist today—including everything from the convolutional neural networks used in image recognition to the long-short-term-memory network used to recognize voice commands. “It’s not wired to one model,” he says.

Google has used the TPU for a good two years, applying it to everything from image recognition to machine translationto AlphaGo, the machine that cracked the ancient game of Go last spring. Not bad—especially considering all the data construction it helped avoid in the process.

https://www.wired.com/2017/04/building-ai-chip-saved-google-building-dozen-new-data-centers

Cyber Attacks get worse in 2017

Kommentar verfassen
cyber-attack

Image Credit: SWEviL/Shutterstock

It was speculated 2016 would see even more cybersecurity activity than 2015, and it did not disappoint. Consider the $81 million stolen from Bangladesh Bank, the 500 million accounts swiped from Yahoo, or the 19,000 emails leaked from Democratic Party officials in the run-up to the election. Not to mention the IoT-powered botnets launching record-breaking DDoS attacks that have brought down major parts of the Internet.

But, in reality, this year’s cyber-attack headlines offer just a glimpse of a cyber war between hackers and security personnel that is being waged on a grand scale every day. More than anything, they are harbingers of worse to come.

Here are some of the escalated challenges we will face in 2017.

1. Attackers won’t just steal data — they will change it

Today’s savvy attackers are moving away from pure data theft and website hacking to attacks that have a subtler target: data integrity. They will use their ability to hack information systems not just to make a quick buck but also to cause long-term, reputational damage to individuals or groups through the erosion of trust in the data itself.

In the past six months alone, we’ve seen attacks like the DNC and Yahoo breaches, which focused on influencing political and economic public opinion, rather than simply gaining a profit. And the hackers aren’t done yet – the Russian group thought to be behind the election-related breaches is moving on to Germany’s elections next, according to a recent statement.

The scenario is particularly worrying for industries that rely heavily on public confidence. In fact, data from the analysis of SEC disclosures found 83 percent of publicly traded companies worry most about risk of brand damage via hacks. But it’s not just them. A laboratory that cannot vouch for the fidelity of medical test results, or a bank that has had account balances tampered with, are examples of organizations at particular risk. Governments, as pointed out above, may also suffer significant damage from such attacks, as critical data repositories are altered and public distrust in national institutions rises.

We’re also seeing this kind of manipulation at smaller scale. For example, we were deployed in a manufacturing firm that used biometric scanners to restrict access to their machinery and industrial plants. We noticed an unusual Telnet for a biometric scanner that was hooked to the corporate network. After further investigation, we found that legitimate data was being altered – quite possibly to add new fingerprints. This type of manipulation, had it not been detected early, would have have let attackers right in through the front door.

While some of the recent breaches and the result of this year’s U.S. presidential election may seem straight out of a movie, tomorrow’s cyber-attacks will make it harder than ever to parse fact from fiction.

2. Consumer devices will be held for (cyber) ransom

Ransomware, like Cryptolocker, has plagued companies around the world — experts reckon these attacks have increased fivefold in 2016 alone. They encrypt critical files at a speed that is virtually impossible to keep up with and leave companies facing hefty fees for their release.

Hospitals have suffered particularly at the hands of ransomware attacks. They are prime targets, as they have become digital jungles full of everything from life-saving medical equipment and critical patient records to patient devices and staff computers — all with cyber defenses that have failed to keep pace. The result is organizations that pay up. Hollywood Presbyterian Medical Center in Los Angeles paid the equivalent of $17,000 in Bitcoin to extortionists after its computers were taken offline for over a week.

In 2017 and beyond, we will start to see the beginning of a new type of extortion on a micro level, as consumers are targeted across a range of connected objects. Imagine getting home and turning on your smart TV only to find that cybercriminals are running a ransomware attack on your device. Would you pay $50 to unlock it? Or what if the new GPS system in your car got hacked when you were late for a meeting — how much would you pay to unlock it?

3. Artificial intelligence will be a weapon

Artificial intelligence is exciting for many reasons — self-driving cars, virtual assistants, better weather forecasting, the list goes on. But attackers will use AI to wield highly sophisticated and persistent attacks, attacks that blend into the noise of busy networks.

We have already seen the first glimpses of attacks going this direction in automated polymorphic and metamorphic malware. Polymorphic malware, which changes its attributes mid-attack to evade detection, has reinforced the obsoleteness of signature-based detection methods. It self-learns and understands its environment and network before choosing its next action. Automation has also been a major factor in the resurgence of ransomware. We can anticipate that artificial intelligence threats will be similar. Imagine a piece of artificially intelligent malware sitting silently on a network, observing its surroundings and learning how to disguise itself. If it understands how to completely blend in with the background noise of a network, could it ever be detected?

The next generation of AI-powered attacks to emerge will use customized code to emulate the behaviors of specific users to fool even skilled security personnel. This includes the ability to craft sophisticated and bespoke phishing campaigns that will successfully dupe even the most threat-conscious employee.

Earlier this year, we were deployed in a charity in California for a proof-of-value. One day, a receptionist received an email containing a fake invoice, supposedly coming from a stationary supplier known to the company. The receptionist opened the attachment, as she recognized the company, and typically handled many invoices per day. As soon as she clicked the attachment, her computer immediately connected to a server in Ukraine and downloaded a malware that rapidly began encrypting files. This will only get worse with “smart” malware driving attacks specifically tailored for their victims.

Next year’s attackers can see more than your social media profile. They’ll know that your 10 a.m. meeting with your supplier is being held at its new headquarters. At 9:15 a.m., an email with the subject line “Directions to our office” arrives in your inbox, apparently from the person you are meeting, as you get off the train. Do you click the map link in the email?

 

Here’s how cyber attacks get worse in 2017

Will IBM be your AI and machine learning platform?

Kommentar verfassen

Here’s how IBM got its start in artificial intelligence, and what it brings to the table for your business or organization.

watsonceo.jpg
IBM CEO Ginni Rometty and IBM Watson Group Senior VP Mike Rhodin.

Image: IBM

Of all the tech giants throwing their weight behind artificial intelligence (AI) and machine learning, few receive the kind of attention garnered by IBM. After its seminal Jeopardy win in 2011, IBM Watson became synonymous with technologies such as cognitive computing and AI.

Upon losing to Watson, former Jeopardy champion Ken Jennings famously wrote „I, for one, welcome our new computer overlords“ under one of his responses. All of a sudden, Watson was a household name, igniting conversations about what could be accomplished with AI.

While Watson is a major part of IBM’s approach to AI solutions, it’s only a piece of the puzzle. Here’s a deeper look at the bigger picture, so businesses can determine if IBM is the right AI vendor for their needs.

The history

IBM Research, the company’s research division, dates back to 1945, when it opened the Watson Scientific Computing Laboratory at Columbia University. IBM’s work in AI began in the 1950s, according to its website. Around that time, an IBM employee named Arthur Samuel wrote a self-learning program for playing checkers, and would later be recognized as a pioneer in AI and machine learning.

In the 1970s, IBM built its first robot, and advanced its work in the field in the 1980s with the IBM RS 1. In the 1990s, IBM Researcher Gerry Tesauro used reinforcement learning (RL) to create a self-learning game of backgammon. Then, in 1997, IBM’s Deep Blue computer famously beat World Chess Champion Garry Kasparov at chess.

The company’s development of actual AI products began more recently. Mike Gualtieri, of Forrester Research, said that IBM’s journey toward AI solutions began in 2009, when it acquired two companies: ILOG and SPSS. ILOG is a business rules engine, Gualtieri said, which used to be called an expert system, while SPSS provides advanced analytics. Both of these purchases helped jumpstart IBM’s work on AI solutions for businesses.

Today, IBM’s AI initiatives are centered around the Watson platform. IBM has Watson solutions for analytics and machine learning, data search and discovery, and conversation tools like chat bots.

The vision

IBM views AI as „augmented intelligence,“ said Guru Banavar, vice president and chief science officer for cognitive computing at IBM Research. To take that concept a little further, 451 Research analyst Nick Patience explained augmented intelligence as „AI — and machine learning in particular — acting as a force multiplier for humans.“

Currently, Banavar said, there are „thousands“ of engineers working on the Watson platform. On a high level, the team is split into two very distinctive camps. One on end of the spectrum is a group working on „very concrete, commercial development and deployment,“ which happens usually on a weekly, monthly, or quarterly basis.

„Then, at the other end of the spectrum, we have teams of people that are working on advanced new technologies — some of which are being invented — all the way up to mathematicians who are developing the underlying techniques for them,“ Banavar said.

After Watson won Jeopardy, Banavar said the team at IBM was focused on building custom systems for specific clients or industry niches. However, they recently had to make a conscious strategic decision to move away from that model to focus on APIs.

Banavar said that IBM realized it wouldn’t be possible to build out all of the applications they wanted to with their existing strategy. So, they turned some of their capabilities into a platform with open APIs, „in order to attract and nurture a larger ecosystem of developers that can build many applications that IBM cannot build by ourselves,“ Banavar said.

Those APIs are being put to use in areas like retail, finance, law, and even fantasy football. But healthcare is one of the primary focuses for Watson solutions.

„I could see a vision where every hospital, every clinical group, had this Watson service. It becomes as essential as an X-ray, as essential as an MRI,“ Gualtieri said. „So, I think that’s their vision. They’re putting a lot against that.“

Strengths

To understand whether or not IBM would be a good fit for your organization, you must weigh your company’s needs against IBM’s strengths. On the technical side of things, Banavar said, these strengths start with Watson’s language capabilities.

„Watson has image processing capabilities, speech processing capabilities, regular numerical data analytics capabilities across the board — we have the entire spectrum,“ Banavar said. „But, if you ask me what is a really unique, and probably the most advanced, capability in Watson, it is language processing.“

When it comes to business, IBM’s AI strength comes from three key elements: IBM Research, its acquisition prowess, and its consultants.

IBM Research may not always produce a breakthrough, but it does give the company a distinctive edge, said Gualtieri. „The advantage of that is that the largest companies in the world — who IBM wants to sell to — want that edge,“ he added.

The ability to acquire the right companies to broaden its portfolio of offerings is a key differentiator for IBM. Banavar noted that the company has also been leveraging open-source libraries and toolkits to make use of new techniques in neural networking, word embedding, and more.

In addition to its research prowess and acquisition budget, IBM has a large network of consultants. According to Patience, that is key, „because a lot of the early machine learning opportunities involve taking enabling technologies such as machine learning algorithms and turning them into enhanced business processes and applications; something IBM understands well.“

Challenges

One of the biggest challenges facing IBM is managing the expectations that come from terms like cognitive computing and AI. This is further compounded by the public-facing nature of Watson, especially in the wake of its Jeopardy win, and confusion around the capabilities of AI as well.

„Everyone thinks that we’re on the verge of Star Trek, like next week,“ Gualtieri said. So, IBM must have a grand and transformative vision about the future of AI, but they also have to keep the expectations in check so customers don’t regret moving forward, Gualtieri noted.

On the question of safety and ethics, many would share Gualtieri’s view that the technology „is not even close to getting to the point where ethical issues are really a serious concern.“ However, Banavar said that ethical challenges are still something the IBM team must consider.

The first crucial issue that must be addressed, Banavar said, is the idea of explainability. If a doctor or financial advisor uses Watson to make a decision, for example, they must be able to understand why Watson chose a particular solution or set of options.

The other ethical consideration is bias. With machine learning systems, the models are built with training data — but the data has to faithfully represent what you’re trying to model, or it could be biased, Banavar said. Because of that, selecting the proper training data set is an ethical decision of the utmost importance. This is made even more important by how broad a potential impact Banavar sees for AI technologies.

„At the end of the day, I do think that cognitive computing is necessary for us to solve the world’s big problems,“ Banavar said.

http://www.zdnet.com/article/should-ibm-be-your-ai-and-machine-learning-platform

Samsung Plans To Give Galaxy S8 An AI Digital Assistant

Kommentar verfassen

All the cool companies have them: digital assistants. Apple has Siri, Microsoft has Cortana, and Google  has the cleverly named Google Assistant. Now, Samsung plans to bring its own iteration of a virtual assistant in the Galaxy S8 next spring, according to a new report from Reuters.

The assistant will be based on work by Viv Labs, a San Jose-based AI company that Samsung acquired this October (the move immediately fueled speculation that Samsung was moving into the AI space). The founders of Viv Labs already have a strong track record in the field as the creators of Siri, which Apple bought in 2010.

Samsung appears to be tapping into Viv’s existing strengths rather than aiming to revamp the platform. One of Viv’s hallmarks is that it is designed to be a one-stop-shop that works seamlessly with third-party services. “Developers can attach and upload services to our agent,” Samsung Executive Vice President Rhee In-jong said during a briefing, according to Reuters. “Even if Samsung doesn’t do anything on its own, the more services that get attached the smarter this agent will get, learn more new services and provide them to end-users with ease.”

If the digital assistant is a hit, it could help Samsung make up for its financial losses over the Galaxy Note 7 recall, which is projected to cost the company at least $5.4 billion. It could also rebuild consumer confidence after the Note 7 debacle and, more recently, a recall of a Samsung top-loading washing machine due to “impact injuries.”

But the company is entering a crowded market. Apple paved the way with Siri, though its early lead is shrinking after the launch of Google’s Assistant, which can tap into Google’s well-established knowledge graph and search capabilities. And there’s always Amazon Alexa, which already has a home in the smart-home devices the Echo, Dot and Tap.

„Every door can be unlocked.“  Ellen Fondiler

http://www.forbes.com/sites/shelbycarpenter/2016/11/06/samsung-plan-galaxy-s8-ai-digital-assistant

MIT’s Moral Machine

Kommentar verfassen

It might save your life: MIT’s Moral Machine asks you to answer moral dilemmas

 

mit moral machine google self driving car lexus

We humans err and err often. If it is not a small mistake like leaving the keys in the fridge, then it is a deadly one like leaving the oven on all day. We tend to be reckless, forgetful, overconfident, easily distracted — dangerous traits when steering a two-ton, metal machine across lanes at 70 mph. Four out of the top five causes for car crashes are the result of human error.

Computers, on the other hand, have purely pragmatic minds. They sense data and react in programmed, calculated ways. Self-driving cars already seem to be safer than humans behind the wheel. The rate of progress in artificial intelligence over the past few years has some experts claiming that driving a car will be made illegal by 2030.

“Machine intelligence may have to deal with situations where someone has to die so someone else can live.”

But, as most drivers know, driving can require split-second decisions with no obvious right answer. A squirrel darts into the road — do you swerve and risk hitting other cars or drive straight and hope the squirrel survives. How would you react if a dog ran into the road? Or a criminal? Or a child? Which lives are worth risking? These questions are being asked by teams of researchers around the world. Now they are looking to you for answers.

“Self-driving cars are now practically inevitable,” Massachusetts Institute of Technology graduate student and research assistant Sohan Dsouza told Digital Trends. “That is a good thing, generally, because they would help save countless lives now being lost daily due to human driver error and can offer independent mobility to countless others who cannot drive.”

To that end, Dsouza, Edmond Awad, and their team at the MIT Media Lab developed Moral Machine, a platform that engages users in moral dilemmas and asks them how the self-driving car should respond. A handful of factors play into each scenario, including the age and gender of victims, their social status, and whether they are breaking the law. Participants are asked to make decisions in 13 dilemmas. The results then pooled as crowdsourced data and may one day be used to guide the development of ethical machines. After judging the dilemmas, users can compare their outcomes to others’ and even design their own for others to answer.

“One of our primary goals is provoking debate among the public,” Dsouza said, “and especially dialogue among users, manufacturers, insurers, and transport authorities.

“Not all crashes can be avoided, and the possibility remains that machine intelligence piloting vehicles may have to deal with situations where someone has to die so someone else can live — rather like the classic philosophical thought experiment known as the trolley problem.”

The trolley problem has been pondered for nearly 50 years. In it, a train car is en route to hit five people down the track. You have a switch that can steer the trolley down another set of tracks, where it will hit only one person. Would you intervene or do nothing?

“There are very few experiment-based studies regarding this possibility,” Dsouza said. “Hence, we needed to create a platform that would be able to generate large numbers of multi-factor scenarios and present them to users in an easy-to-understand, easy-to-use, and engaging way, so as to build a model of how people perceive the morality of machine-made decisions.”

“One of our primary goals is provoking debate among the public.”

Moral Machine has gathered answers on more than 11 million scenarios so far. Although the team has yet to perform a deep analysis, they are noticing regional trends that hint at the rocky road ahead. “On average, respondents from western countries place a relatively higher value on minimizing the number of overall casualties — that is, they approve more utilitarian choices — compared to respondents from eastern countries,” Dsouza said.

Revealing these cultural discrepancies fosters debate and dialogue, which is essential to making progress. “We believe we have already made an impact,” Dsouza said. “This dialogue will eventually help the stakeholders in this scene reach an equilibrium of legislation, liability assessment, moral comfort, and public safety.”

http://www.digitaltrends.com/cars/mit-moral-machine/

Google Hits a Samsung Roadblock With New AI Assistant – Viv & Adam Cheyer

Kommentar verfassen

Google just debuted a digital assistant, which it hopes to place inside smartphones, watches, cars and every other imaginable internet-connected device. It’s already hit a snag.

The Alphabet division launched new smartphones last week with the artificially intelligent assistant deeply embedded. It also rolled out a speaker with the feature at its core and announced plans to let other companies tie their apps and services to the assistant.

A day later, Samsung, which just announced it was ending production of its problematic Galaxy Note 7 smartphones, said it was acquiring Viv Labs, a startup building its own AI voice-based assistant.

At first, the deal looked like a counter-punch to Samsung rival Apple — Viv is run by the creators of Apple’s Siri assistant. But buying Viv may be more of a problem for Google, because Samsung is the biggest maker of phones running Google’s Android mobile operating system.

Google strategy is now centered on the assistant, rather than its search engine, because it’s a more natural way for people to interact with smartphones and other connected devices. Getting all Android phone makers to put the Google assistant on their devices would get the technology into millions of hands quickly. But Samsung’s Viv deal suggests assistants are too important for phone makers to let other companies supply this feature.

Last week, despite the Note 7 crisis, Samsung executive Injong Rhee said the company plans to put Viv’s technology in its smartphones next year and then embed it into other electronics and home appliances. A Samsung representative and a Google spokeswoman declined to comment.

That’s a necessity for Samsung, according to some analysts and industry insiders.

„As AI is becoming more sophisticated and valuable to the consumer, there’s no question it will be important for hardware companies,“ said Kirt McMaster, executive chairman of Cyanogen, a startup that makes Android software. Mr. McMaster, a frequent Google critic, said other Android handset makers will likely follow Samsung’s move.

„If you don’t have an AI asset, you’re not going to have a brain,“ he added.

Google may already have known that some Android phone makers — known as original equipment manufacturers, or OEMs — were reluctant to embrace its assistant.

„Other OEMs may want to differentiate“ Google’s Android chief Hiroshi Lockheimer told Bloomberg before it released its own smartphones. „They may want to do their own thing — their own assistant, for example.“

Samsung and Google have sparred in the past over distribution. Google requires Android handset makers to pre-install 11 apps, yet Samsung often puts its own services on its phones. And the South Korean company has released devices that run on its own operating system, called Tizen, not Android.

Viv was frequently on the short-list of startups that could help larger tech companies build assistant technology. Founded four-years ago by Dag Kittlaus, Adam Cheyer and Chris Brigham, the startup was working on voice technology to handle more complex queries than existing offerings.

While it drummed up considerable attention and investment, Viv has not yet released its product to the public. And some analysts are skeptical of Samsung’s ability to convert the technology into a credible service, given its mixed record with software applications.

„It will be very hard to compete with Google’s strength in data and their AI acquisitions,“ said Jitendra Waral, senior analyst with Bloomberg Intelligence. „Samsung would need to prove that its AI solutions are superior to that of Google’s. They are handicapped in this race.“

Samsung is also focused on handling the fallout from its exploding Galaxy Note 7 phones, potentially taking management time away from its Viv integration.

But it’s a race Samsung has to join. In recent years, Samsung acquired mobile-payments and connected-device startups to keep up with Apple, Google and Amazon. Digital voice-based assistants may be more important, if they become the main way people interact with devices.

Silicon Valley titans are rushing into the space because of this potential. Amazon is trying to sign up developers for its Alexa voice technology. Apple has recently touted more Siri capabilities and opened the technology to other developers. And now Google, considered the leader in artificial intelligence, is making its own push.

„I don’t ever remember a time when every single major consumer tech company — and even enterprise companies — have been singularly focused on an identical strategy,“ said Tim Tuttle, chief executive officer of MindMeld Inc., a startup working on voice interaction software. „They’re all following the exact same playbook.“

 

http://adage.com/article/digital/google-hits-a-roadblock-ai-assistant/306244/

Google hired writers from Pixar and The Onion to make Assistant more personable

Kommentar verfassen

Google wants its Assistant to be more than just an order-taking robot — so it hired some clever writers from outside the company to help make it happen.

A new story from the Wall Street Journal’s Christopher Mims details the advancements of different artificial intelligence devices like Amazon Echo and Google’s rival product Home, and how they’re comforting for those who live alone thanks to how personable the AI’s have become.

For Google, that friendly personality is thanks to a team of writers from Pixar and The Onion who helped make the Assistant — which powers Google’s Home device — sound more like a human and less like a robot, according to the Journal. Google’s eventual goal is to help users build an emotional connection with the Assistant, the Journal reports.

Google unveiled its Assistant-enabled Home device last week, a direct competitor to other AI-powered hardware devices like Amazon’s Echo. The Assistant itself is similar to Alexa, which powers the Echo: It has voice-recognition software, natural language recognition, and it gets smarter over time.

You can ask the Assistant to tell you a joke, give you the weather or set a timer, but you can also ask it to do things like remember your favorite sports team or the city you live in. Much like other AI — like Alexa or Apple’s Siri — the Assistant can be equal parts sweet and sassy, which is what helps it seem more relatable and more human. The Assistant lives inside Google Home, but it’s also enabled in Google’s new messaging app, Allo and its new Pixel smartphone.

 

http://www.businessinsider.de/google-assistant-pixar-the-onion-2016-10

Dieter Zetsche beschwört den Wandel der Autobranche

Kommentar verfassen

zetsche-860x860 
Betont lässig: Daimler-Chef Dieter Zetsche brach in Paris mit den Traditionen der Autobranche. Er führte Freizeitmode vor und verkündete das neue Leitbild einer agilen Organisation.

(Foto: Daimler)


Dabei bremste Daimler bislang bei den alternativen Antrieben. Auch die jüngste Elektroauto-Studie geriet mutlos. Gelingt dennoch die Transformation zum Tech-Konzern?

Analyse von Joachim Becker

Die Schuhe sind Teil der Inszenierung: Wenn der 63-jährige Chef eines Weltunternehmens mit Jeans und Turnschuhen rumläuft, dann befindet er sich für gewöhnlich im Urlaub oder in der Midlife-Crisis. Dieter Zetsche will augenscheinlich nicht zum alten Eisen gehören. Doch sein Problem ist weniger privater als unternehmerischer Natur: Der Daimler-Boss will das Flaggschiff der deutschen Autoindustrie zur Tech-Company umbauen.

Vor einer Gründerzeit im Neckar-Valley muss er einige Altlasten bewältigen. Zum Beispiel den Erfolg des bewährten Geschäftsmodells: Trotz Rekordabsatzzahlen fordert Zetsche ein radikales Umdenken seiner Mitarbeiter. Statt sprudelnde Erlöse zu feiern, sollen sie sich an einer Revolution beteiligen. Ausgang offen.

Bisher stand Mercedes auf der Bremse

Was Zetsche auf dem Pariser Autosalon verkündet, ist eine Revolution von oben: „Wir wollen nicht nur die Verwandlung unserer Produkte vorantreiben, sondern auch die Verwandlung unserer Organisation signifikant beschleunigen.“ Bisher standen die Stuttgarter nicht nur bei alternativen Antrieben auf der Bremse. Kurz nach einer Welttournee mit Wasserstofffahrzeugen wurde 2013 die angekündigte Serienproduktion abgesagt. Im selben Jahr überließ man BMW i die Vorfahrt bei komplett neuen Elektroautos. Die ersten E-Smarts mit Hochtemperaturzellen hatten 2007 bloß Forschungscharakter. Später half Tesla auch bei der Mercedes-B-Klasse-e-cell mit Batterien nach. Trotzdem oder gerade deshalb zögerte der Elektroingenieur Zetsche, Milliarden auf eine ungewisse Elektro-Zukunft zu wetten.

Anzeige

Zetsche ist kein junger Wilder wie Elon Musk, der als New-Age-Guru einer emissionsfreien Zukunft auftritt. Der Erfolg von Tesla und vor allem die Geschwindigkeit, mit der sich das Start-up weiterentwickelt, sorgen im Daimler-Vorstand allerdings für Stirnrunzeln. Der Elektro-Pionier punktet mit Software-Updates, die neue Funktionen ins Auto bringen. Trotz gravierender Rückschläge wie beim Autopiloten will Tesla das erste autonome Auto auf den Markt bringen.

Die Serienversion des Generation EQ kommt 2018

Bei dem halsbrecherischen Technologietempo gibt es eine Reihe von Unwägbarkeiten: „Unser Zielkorridor für den Elektroabsatz im Jahr 2025 liegt zwischen 15 und 25 Prozent. Genauer können wir es einfach nicht prognostizieren“, gesteht Mercedes-Vertriebsvorstand Ola Källenius. Trotzdem legt Daimler jetzt den Schalter für die neue Elektro-Submarke EQ um. Die Serienversion des Pariser Showcars „Generation EQ“ wird ab 2018 zum Preis eines „vernünftig ausgestatteten Mercedes GLC“ (also für rund 60 000 Euro) angeboten. Mindestens neun weitere reine E-Mobile vom Kompaktauto bis zum Supersportler sollen bis 2025 folgen.

Mercedes will bis 2025 Tesla als Marktführer bei Premium-Elektrofahrzeugen ablösen. Die leistungsstarken Stromer werden aber schon Ende dieses Jahrzehnts Standard sein – als Unterscheidungsmerkmal einer Marke taugen sie dann nicht mehr. Deshalb stürzen sich die Blechbieger in weitere Abenteuer: „Viele Autohersteller wollen heute Mobilitätsanbieter werden. Das ist schön und gut. Aber die Transformation der Branche ist noch viel grundlegender“, warnt der Daimler-Boss.

Daimlers Erfolgsgeschichte geschieht zu langsam

130 Jahre lang definierte sich die Autoindustrie über Hardware. Daimler ist das beste Beispiel, wie schwierig nun das Umdenken ist: Die Stuttgarter haben zwar 2007 das flexible Einweg-Carsharing mit vollvernetzten Smarts erfunden. Doch es dauerte zehn Jahre, um Car2go auf zwei Millionen Nutzer zu bringen. Was Daimler als Erfolgsgeschichte verkauft, geschieht letztlich zu langsam, um mit neuen Wettbewerbern zu konkurrieren. Maßgeschneiderte, automatisierte Mobilitäts-Services könnten dem Verkauf von Privat-Pkw in Zukunft mehr und mehr Konkurrenz machen. Niemand weiß aber, wie und wann sich der Wandel genau vollziehen wird.

Bisher sind die Entwicklungsabteilungen der Autohersteller entlang von neuen Produkten aufgestellt. Genauso wichtig werden allerdings innovative Geschäftsmodelle sein. Daimler will den Technologiewandel vom Kundenerlebnis her neu denken: „Wir erwarten, dass sich das Auto von einem Produkt in eine ultimative Plattform verwandelt. Das ist ein fundamentaler Perspektivenwechsel“, sagt Dieter Zetsche. Diese Plattform ruhe auf vier Säulen: Vernetzung (Connected), Autonomes Fahren, Sharing und Elektromobilität. Zusammen ergeben die Anfangsbuchstaben das Wort Case. „Wir haben gerade einen neuen Unternehmensbereich mit diesem Namen gegründet, um diese Themenfelder zusammenbringen“, so Zetsche.

Noch ist unklar, was die Kunden wollen

Den Kunden in den Mittelpunkt zu stellen, ist eine prima Idee. Das Problem ist nur: Kaum ein Mercedes-Käufer hat bisher nach Elektromobilen gefragt. Geschweige denn Interesse an einer Internet-Plattform gezeigt, über die er seine Luxuskarosse mit anderen teilen kann. Genau das will Mercedes mit einer Sharing-Plattform ab November dieses Jahres in Deutschland erproben.

Zetsche stellt in Paris jedoch klar, das keine einzelne Technologie oder Dienstleistung den Unterschied machen werde, sondern ein neuartiges Gesamterlebnis von Mobilität: „Jede der Case-Säulen hat das Potenzial, die gesamte Automobilindustrie auf den Kopf zu stellen. Aber die wahre Revolution ist die Verbindung dieser Aspekte in einem umfassenden, nahtlosen Paket.“

Zetsches ständiger Balanceakt

Alt und neu, analog und digital, Sakko zur verwaschenen Jeans: Als Vordenker balanciert Zetsche ständig zwischen den Gegensätzen. Seine Grundsatzrede auf dem Pariser Autosalon klingt über weite Strekken wie ein Appell an die eigene Belegschaft: Das Schweizer Uhrwerk als Zeichen für Verlässlichkeit und Präzision im mechanischen Zeitalter – „das bleibt auch in Zukunft wichtig!“, beruhigt er seine Mitarbeiter. Schon im nächsten Moment predigt er jedoch das Credo des digitalen Zeitalters: Ihm gefalle die Idee einer „agilen Schwarmorganisation“, verkündet der Manager mit dem grauen Walrossbart: „Case wird als rechtlich getrennte Organisation ein perfekter Startpunkt für diese Vorstellung sein.“

Mehr Silicon Valley wagen, ohne die Stärken der Vergangenheit aufzugeben, lautet die Botschaft. Noch weiß allerdings niemand, wie dieses Autofahren 2.0 wirklich aussieht, geschweige denn, wie man damit Geld verdient. Elektro-Studien wie der Mercedes Generation EQ und der VW I. D. zeigen in Paris jedenfalls das genaue Gegenteil einer Design-Revolution. Mit ihren mutlos-monolithischen Grundformen pendeln sie irgendwo zwischen Van und Crossover. Bloß nicht auffallen!

Der Fluch der großen Reichweite

Damit sich die Hoffnungsträger wenigstens ein bisschen vom Mainstream unterscheiden, wurde ihnen das Dach tief ins Gesicht gedrückt. Doch der Trick funktioniert nur auf geschickt fotografierten Bildern. Wer versucht hat, auf den Rücksitzen des VW I. D. zu sitzen oder sich unter dem Dachholm des Mercedes EQ durchzuschlängeln, erkennt den Schwindel: In der Serie werden aus halbwegs schnittigen Showcars bleischwere Hochdachautos. Das ist der Fluch der großen Reichweite.

Die Physik lässt sich auch im digitalen Zeitalter nicht überlisten: Stromer mit 500 Kilometer Radius benötigen riesige Unterflur-Batteriepakete, auf denen die Passagiere thronen. Tesla kann dieses hochgebockte Kutschendesign mit einem flachen Batterieformat recht gut kaschieren. Weil kein anderer Hersteller die schmalen Rundzellen von Panasonic verwendet, werden sich die Designer mit ihren Tesla-Fightern mächtig anstrengen müssen.

Auch Matthias Müller beschwört ein „neues Zeitalter“

Dass die meisten Kunden 500 Kilometer Batteriereichweite gar nicht brauchen, ist die Ironie dieses Technologiewandels. Bisher hat kaum jemand die Stromer als Erstauto für die ganze Familie verwendet, geschweige denn Urlaubsfahrten damit geplant. Das Wettrennen um den größten Batterieradius wendet sich also nicht an die umweltbewussten Pioniere, sondern an den komfortorientierten Otto-Normalverbraucher: Einmal pro Woche Tanken ist gelernt. Bloß nicht umgewöhnen!

Auch Matthias Müller beschwört in Paris ein „neues Zeitalter“: „Die Elektromobilität und digitale Vernetzung werden zu Game Changern“. Welche Spielregeln für eine neue Generation von Kunden gelten werden, weiß aber auch das Oberhaupt des Volkswagen-Konzerns nicht sicher zu sagen. Vielleicht sind es digital animierte Innenwelten, die ein neues Markenerlebnis schaffen. Auf dem Mercedes-Stand ließ sich Müller lange die Bedienphilosophie der EQ-Studie erklären. Die hochauflösenden 3-D-Landschaften auf dem Riesenbildschirm sollten ihn wohl von der Tristesse im VW I. D. ablenken.

Quelle: http://www.sueddeutsche.de/auto/die-zukunft-von-daimler-dieter-zetsche-beschwoert-den-wandel-der-autobranche-1.3193256

Mercedes‘ Tesla killer is coming in 2019

Kommentar verfassen

mercedes eqScreenshot

Mercedes-Benz just made a huge move to take on Tesla.

The German automaker unveiled its all-electric SUV concept at the Paris Motor Show Thursday, and with a competitive price tag and solid range potential, it’s poised to become a big competitor in the EV space.

Called Generation EQ, the SUV concept is expected to have a range up to 310 miles.

Called Generation EQ, the SUV concept is expected to have a range up to 310 miles.

Mercedes-Benz

The production version of the SUV is expected to hit the roads in 2019, Dieter Zetsche, the head of the Mercedes-Benz car division, said at the Paris Motor Show press event. Mercedes is calling the SUV unveiled today a „close-to-production concept vehicle.“

The production version of the SUV is expected to hit the roads in 2019, Dieter Zetsche, the head of the Mercedes-Benz car division, said at the Paris Motor Show press event. Mercedes is calling the SUV unveiled today a "close-to-production concept vehicle."

Mercedes-Benz

The car will fall in the same price range as the GLC Crossover, which currently starts at $39,150. That’s a very competitive price for an electric SUV, considering Chevy Bolt’s all-electric SUV crossover will start at $37,500 when it hits showrooms at the end of 2016.

The car will fall in the same price range as the GLC Crossover, which currently starts at $39,150. That's a very competitive price for an electric SUV, considering Chevy Bolt's all-electric SUV crossover will start at $37,500 when it hits showrooms at the end of 2016.

Mercedes-Benz

The interior comes with a massive, 24-inch display that shows speed, range, driving data, and navigation information. The display will alert the driver if the car is running low on battery and of nearby charging stations. The steering wheel also comes with touch controls.

The interior comes with a massive, 24-inch display that shows speed, range, driving data, and navigation information. The display will alert the driver if the car is running low on battery and of nearby charging stations. The steering wheel also comes with touch controls.

Mercedes-Benz

The SUV comes with some autonomous features, but not many. Mercedes says the car can automatically adjust the speed and driving dynamics when approaching curves.

The SUV comes with some autonomous features, but not many. Mercedes says the car can automatically adjust the speed and driving dynamics when approaching curves.

Mercedes-Benz

The car is part of Mercedes‘ efforts to ramp up its electric-car offerings. Daimler’s chief development officer, Thomas Weber, said in May that Mercedes-Benz was aiming to add four new electric cars to its model range by 2020.

The car is part of Mercedes' efforts to ramp up its electric-car offerings. Daimler's chief development officer, Thomas Weber, said in May that Mercedes-Benz was aiming to add four new electric cars to its model range by 2020.

Mercedes-Benz

 

http://www.businessinsider.de/mercedes-electric-suv-production-in-2019-photos-2016-9?op=1