Prima dell’esplosione della pandemia, la notizia principale sul 5G riguardava il rollout del 5G, un argomento peraltro a volte controverso. In primo luogo, il dibattito sul ban americano alle aziende cinesi era acceso, con un veto che oltre agli Usa coinvolgeva anche Regno Unito e Australia. In seguito, i temi sensibili sul 5G si sono allargati anche anche ad altri fronti, in particolare per quanto riguarda i rischi per la sicurezza delle reti e la privacy degli utenti legata all’avvento dell’IoT.  

Ma l’esplosione del Covid-19 ha costretto diversi analisti ed esperti a rivedere le loro posizioni sul 5G, perché se è vero che le preoccupazioni sul fronte security permangono, è innegabile che il network di nuova generazione è fondamentale per contrastare la pandemia.

Crisi? Quale crisi?

Il Covid-19 domina da settimane il palinsesto dell’informazione e non è poi così strano che alcuni annunci importanti sul 5G siano passati fuori dai radar quasi sotto silenzio. Nonostante la pandemia, in Cina gli operatori a inizio marzo hanno annunciato che entro il 2020 realizzeranno 500mila stazioni base 5G.

Gli Usa stanno facendo altrettanto. Sempre in marzo la Casa Bianca ha pubblicato una strategia nazionale per affrontare le sfide alla sicurezza poste dalla nuova tecnologia, chiamando tutto il mondo ad una collaborazione globale per lo sviluppo dell’infrastruttura 5G.

In qualche modo, questi sviluppi alla fine non sono una sorpresa. La pandemia ha mandato il business a gambe all’aria in molti paesi, però non lo ha interrotto del tutto.   

 Le aziende si stanno continuando a preparare all’avvento del 5G, sviluppando ad esempio software di pagamento e sistemi collaborativi di lavoro e smart working.

E il valore potenziale del 5G in ottica smart working è sotto gli occhi di tutti e sarà sempre più importante una vota che, passata la crisi sanitaria, i modelli di lavoro saranno cambiati per sempre.

5G, big data e la guerra al virus

Ci sono poi dei modi ancor più diretti in cui il 5G sta contribuendo in modo concreto a contrastare la pandemia. Forbes ha raccontato recentemente in che modo l’Intelligenza Artificiale, la Data Science e la Tecnologia hanno contribuito a combattere il virus per tracciare i casi, fornire medicinali e anche per sviluppare nuovi farmaci per combatterlo. Tutto ciò però poggia in primo luogo su connessioni che devono essere ultraveloci e affidabili, come il 5G, in particolare l’intelligenza artificiale.

La grande maggioranza delle aziende che usa questi sistemi per contribuire nella lotta al coronavirus utilizza servizi cloud per gestire le dosi massive di dati con cui lavorano e per ridurre i costi di manutenzione dei siti.

Per questo le connessioni 5G hanno un grande valore, perché consentono lo scambio rapido di enormi quantitativi di dati con i cloud provider.

Analizzando questo fenomeno, emerge che se da un lato la pandemia sta danneggiando una vasta gamma di settori, dall’altro vi sono invece altri titpi di business che ne traggono vantaggio.

Fra questo, ci sono certamente le aziende che già prima della pandemia stavano sviluppando sistemi di tracciamento degli utenti basati sull’Intelligenza Artificiale, quelle che offrono chatbot per la condivisione automatica di informazioni e quelle che rientrano nel mercato della cybersecurity.

Le sfide

Di certo le preoccupazioni sulla sicurezza delle reti 5G non sono state completamente silenziate.L’ecosistema del 5G ha ereditato alcune vulnerabilità già figlie del 4G, e ci sono tuttora significative preoccupazioni sul fatto se la industry della sicurezza sia in grado di gestire il 5G. Ma negli ultimi tempi c’è stato un leggero ma sostanziale cambiamento nel mood in cui gli analisti parlano del 5G: non ci si domanda più se sia possibile realizzare le nuove reti, ma in medo possano essere realizzate in scurezza. Questo cambiamento contiene due elementi principali: da un lato sgombra il campo sulle assurde fake news che sono circolate, secondo cui il 5G avrebbe contribuito alla diffusione del virus. In questo senso, la pubblicazione della strategia nazionale sul 5G della Casa Bianca è una novità assai benvenuta in questo senso e dimostra come gli Usa abbiano la volontà di guidare lo sviluppo di standard di cybersecurity validi per tutti. La seconda linea di pensiero riguarda la necessità di favorire una transizione graduale al 5G.

Perché la necessità di rispondere al Covid-19 può accelerare il 5G

Negli USA è stato creato il termine “Covidiots”,  “Covidioti”, per indicare quella gioventù bruciata che, nonostante gli allarmi ed i divieti dovuti al Coronavirus ed alla relativa malattia, il Covid-19, continuano in pazzie e vacanze spensierate. Tizi quali questo:

“If I get corona, I get corona. At the end of the day, I’m not gonna let it stop me from partying”: Spring breakers are still flocking to Miami, despite coronavirus warnings. https://t.co/KoYKI8zNDH pic.twitter.com/rfPfea1LrC

— CBS News (@CBSNews) March 18, 2020

“Se prendo il Coronavirus, prendo il Coronavirus. Nulla mi impedirà di fare festa” come dice questo COVIDIOT che si è precipitato il Florida per il tradizionale “Sping Break”, le vacanze catenate degli universitari in primavera.

Questi personaggi dallo sguardo intelligente e sveglio, il meglio della cultura d’oltreoceano, sono un caso patologico, ma hanno anche messo in luce quanto in profondità si possano seguire, praticamente pedinare, le persone, anche se, in teoria, non si sta violando la loro .  Società come la Tectonix sono in grado di visualizzare in modo quasi puntuale la presenza ed il movimento delle persone utilizzando i loro dati di posizionamento, contenuti nei loro smartphone, pur mantenendoli, teoricamente, anonimi:

Want to see the true potential impact of ignoring social distancing? Through a partnership with @xmodesocial, we analyzed secondary locations of anonymized mobile devices that were active at a single Ft. Lauderdale beach during spring break. This is where they went across the US: pic.twitter.com/3A3ePn9Vin

— Tectonix GEO (@TectonixGEO) March 25, 2020

Come vedete i flussi delle persone sono  seguite in modo perfetto, anche su scala minima. Basta identificare un luogo e, tramite la geolocalizzazione dei telefonini, si possono analizzare i flussi delle persone giunte in quel particolare posto, nel caso specifico i “Covidiot” che sono andati a Fort Lauderdale per lo Spring Break.

Potete pensare di essere protetti dall’anonimia dei dati: sono telefoni, non collegati ad alcuna utenza, ad essere rilevati e tracciati,  quindi nessun pericolo…..  SBAGLIATO. Grazie allo studio dei dati statistici ed alla AI è possibile identificare anche singoli numeri , teoricamente anonimi.  Semplicemente una specifica persona ha delle abitudini particolari che , analizzati nella loro ripetitività,permettono alla AI di identificarlo. Se Giovanni fuma, e nella sua area c’è un telefono che si ferma sempre dal tabaccaio, unendo questo dato con altri sarà possibile identificare il suo specifico telefono, anche se il numero è, teoricamente, anonimo.

Quindi avete guadagnato un telefonino, ma avete perso il libertà e privacy. Ricordatevelo questo, e se fate una corsetta fuori casa, lasciate lo smartphone fra le quattro mura.

Sostieni ScenariEconomici.it

Fra Mass Media ormai completamente uniformati ai bollettini di informazione del Potere, Scenari Economici rimane una voce che presenta opinioni libere ed indipendenti. Aiutaci a crescere ed a farci conoscere, sia condividendo i nostri contenuti, sia, se vuoi, con una donazione. Grazie!

https://scenarieconomici.it/come-i-telefoni-permettono-di-spiarvi-il-caso-covidiots/

As we enter 2020, it’s time to leave behind old marketing strategies and adopt new ones. After all, there is a lot of innovation happening. From AI in customer service to influencers taking over social media, a lot has changed in the last year. Digital marketing campaigns now sprawl across different channels, tools, and processes.

While some strategies might remain partially the same, they will get a facelift in 2020. If you are looking to take your marketing to the next level in 2020, we’ve got some tips for you.

Here are some of the best ways to get more out of your marketing campaigns this year:

1. Create more video content

Videos have transformed the content marketing landscape. They have changed the way brands promote their products and create content. And video content is also becoming more popular. Wondering why?

Because people like watching videos.

In fact, 72% of people said that they would like to watch videos to learn about a product or service.

If you haven’t been using video content, you’re losing out on a lot of opportunities to grow your brand. With the evolution of social media platforms like Instagram, YouTube, and Facebook, the video content game has changed dramatically. Video content is a versatile tool to increase brand awareness and drive engagement.

So, what kind of videos can brands create for promotion?

There are many different options — educational videos, onboarding videos, meet-the-team videos, customer journey videos, event videos, and more.

Looking for inspiration?

Take a look at the Reebok video marketing campaign

They created a video for their #HonourYourDays campaign which shows a woman’s relationship with running in reverse mentioning that the average human life is about 25,915 days.

While it doesn’t directly promote their products, it does highlight the importance of fitness. It’s a great way to spread awareness without being too promotional.

[embedded content]

Source: YouTube

2. Use chatbots to enhance customer experience

With the help of intelligent chatbots, you can also leverage customer data. These virtual assistants are capable of collecting data from customer interactions and giving you insights into how your brand can improve.

For example, if customers are complaining about a specific product, you can dive deep into the issue and understand the problem on a granular level.

Providing a great customer experience can help brands increase lead generation and revenue. What’s more, it can also help you improve your customer experience and lower your customer acquisition cost.

One of the biggest benefits of customer service chatbots is that they are available 24×7 and give quick responses to queries.

Chatbots are also great tools for helping customers keep track of their purchases.

Mastercard’s chatbot on Facebook Messenger

For example, Mastercard’s chatbot on Facebook Messenger allows its users to ask specific questions. On it, a user can ask how much they’ve spent on food or on Uber.

For each user, the answer is personalized based on their bank statement. Users can also send submit questions to ask for more information about a particular product.

Image via Mastercard

3. Focus on personalization

Most content marketers believe that when a prospect customer engages with their brand and makes a purchase, it’s a win.

But what’s next?

A customer journey starts right from when a target customer engages with your brand and it continues through conversion. But it also includes providing customer support and elevating their buyers’ journey to turn them into cheerleaders of your brand.

So, it is essential to reach out to your audience and provide valuable services that are not solely revenue-centric. 76% of consumers believe that it’s easy to switch to a different brand to find an experience that will match their expectations.

So, how can you provide a memorable experience to your customers?

That’s where personalization comes into play. Not only can it improve the customer experience you deliver, but it can also boost your sales.

Providing a personalized experience can significantly impact your relationships with your customers. So, how can you create a more personalized experience for your customers?

Tools like Google Analytics and Cortex can help you get insights into your audience’s behavior. By using that data, you can find out what kind of content your audience is looking for and then optimize your content creation strategy accordingly.

4. Leverage influencer marketing

Influencer marketing has grown to become a whopping five to ten billion-dollar industry. It has helped marketers reach their target audiences, generate brand awareness, get more leads, and boost sales.

Collaborating with social media influencers has become a great way for brands to promote their products and services. When an influencer endorses your brand or reviews your product, it can get you a lot of publicity.

How?

Take a leaf from Daniel Wellington’s digital marketing strategy

To create buzz about their brand, they send their watches to multiple influencers. They encourage those influencers to create posts with the hashtag #DanielWellington to help promote their products.

Source: Instagram

Another example of great influencer marketing is Subaru’s #MeetAnOwner campaign

In this campaign, the company reached out to influencers from diverse backgrounds. One of the biggest influencers in their campaign was Devin Graham (@devinsupertramp) with over five million YouTube subscribers at the time. He is known for his daredevilry.

By partnering with Devin, Subaru was able to reach out to outdoor adventure junkies. Graham’s sponsored video with Subaru gained around 1.3 million views and 1.1k comments.

[embedded content]

Source: YouTube

5. Automate your email marketing campaigns

About 73% of marketers consider email marketing to be crucial to their company’s success.

But it can be a time-consuming strategy. If you’re short on time, there is a way out – automate your email marketing campaigns. Not only can it help you save a ton of time, but it can also help you create more personalized content.

With this strategy, you can ensure that your content is delivered to your customers at the right times.

For instance, if a customer has abandoned their cart, you can send them a reminder to complete the transaction.

Here’s an example of how American technology company, FiftyThree, notifies their customers about abandoned carts with an appealing message.

Source: Shopify

Marketers often use these emails to encourage their customers to take another step or get closer to a purchase.

Emails also provide you with an opportunity to upsell your products.

For example, an email that includes a section containing products similar to a user’s previous purchases can trigger their engagement.

Source: Smartrmail

Email automation saves the time and efforts that are otherwise invested into following up with each action of a user. It also reduces the potential for errors such as hitting the send button prematurely when the user might not be ready.

6. Build a community

Building a community around your brand will help you improve your brand awareness, drive higher engagement, and organically boost your SEO.

When users create and share content around your brand, you are likely to get noticed by search engines. Brand mentions, post shares, comments, and likes — all of these help your cause.

Further, having a community of followers also helps strengthen your branding.

When you share valuable content that helps your target audience or aligns with their interests, they tend to associate more with your brand and become loyal customers.

Some companies even have brand ambassadors who regularly create content for them in return for benefits such as free products, invites to product launches or other events.

A great example of a community is GoPro, an action camera company that has solidified its position in the market by building a strong community of sports enthusiasts, photographers, and adventurers across the world.

The brand builds personal relationships with their customers while maintaining a focus on its target audience of sports enthusiasts.

Source: Instagram

Conclusion

Videos, chatbots, influencer marketing, personalization, and more — there is a lot of innovation happening in the digital marketing space. Anyone, from big brands to mom-and-pop stores, can use these strategies to grow their business. What are you waiting for? It’s time to incorporate these strategies into your marketing plan for 2020.

Have you used any of these strategies before? Please share your experiences and insights in the comments section.

Shane Barker is a Digital Strategist, Brand and Influencer Consultant. He can be found on Twitter .

https://www.searchenginewatch.com/2020/03/02/six-great-for-more-out-of-digital-marketing-campaigns-2020/

Le aziende che passano dalla sperimentazione alla piena integrazione dell’Intelligenza Artificiale all’interno della propria strategia, migliorano la propria performance finanziaria del 32%. E’ quanto emerge dallo studio “AI: Built to Scale” realizzato da Accenture, secondo cui  tre manager italiani su quattro affermano che è necessariopassare dalla sperimentazione alla piena integrazione dell’Intelligenza Artificiale (AI) nella propria strategia aziendale, per poter continuare ad essere competitivi nei prossimi cinque anni.

L’88% dei C-level italiani ritiene che far leva sull’intelligenza artificiale (AI) sia fondamentale per raggiungere i propri obiettivi di crescita e quasi tutti la considerano un fattore strategico. Tuttavia, il 77% riconosce di avere difficoltà nel farlo.

L’analisi ha coinvolto 1.500 manager (C-level) di 16 settori in vari Paesi, allo scopo di capire in che modo le aziende stiano implementando l’intelligenza artificiale nelle loro organizzazioni. Dal report emerge che solo il 16% delle aziende nel mondo ha compiuto il passo che dalla fase sperimentale porta alla piena integrazione di solide capability di intelligenza artificiale all’interno dell’azienda. questo gruppo di top performer sta ottenendo il triplo di ritorno dagli investimenti in Intelligenza Artificialerispetto ad aziende simili.

“Integrare l’Intelligenza Artificiale all’interno dei processi aziendali e nell’intera struttura organizzativa – coinvolgendo tutti i principali stakeholder – permette alle aziende di migliorare la propria agilità competitiva – ha commentato Marco Morchio, Responsabile di Accenture Strategy – I dati ce lo dimostrano, le aziende in grado di scalare strategicamente l’AI migliorano le proprie performance del 32% secondo i tre principali indicatori (valore dell’impresa / ricavi; prezzo / utile per Azione; prezzo / vendite). Il segreto del successo di questi top performer – i cosiddetti strategic scaler – si basa su tre elementi chiave: la disponibilità di una solida base di dati rilevanti; l’impegno, in primo luogo dei manager, a implementare strategicamente l’Intelligenza Artificiale in tutte le aree e i livelli dell’organizzazione e il coinvolgimento di team diversificati e multidisciplinari dedicati all’AI.”

Queste caratteristiche consentono agli Strategic Scaler di adottare soluzioni di Intelligenza Artificiale ad un ritmo più sostenuto, con una velocità doppia rispetto alle altre aziende, riuscendo a lanciare il doppio dei progetti pilota e a portarne a scala di più ed in tempi più brevi, entro uno o due anni dal lancio.

Un’ulteriore caratteristica che distingue le aziende più performanti è quella di coinvolgere i giusti talenti. Invece di fare affidamento su un solo esperto di intelligenza artificiale, il 92% degli strategic scaler nel mondo ha integrato tutta la struttura l’aziendale con team multidisciplinari. La multidisciplinarietà e la pluralità di stili individuali è un propulsore di creatività e innovazione che, oltre ad avere benefici tangibili sulla gestione dell’Intelligenza Artificiale permette di interpretare il cambiamento, cavalcare le sfide del mercato e raggiungere alte performance.

“Le aziende devono agire al proprio interno creando una cultura aziendale basata sui dati, per individuare gli insight in grado determinare un cambiamento capace di generare valore – ha dichiarato Massimo Morielli, Responsabile di Accenture Digital – L’utilizzo dell’intelligenza artificiale nelle organizzazioni implica la nascita di nuove forme di intelligenza, frutto della collaborazione tra i talenti delle persone e le opportunità che questa tecnologia può offrire, è per questa ragione che noi parliamo di applied intelligence”.

https://www.key4biz.it/lai-migliora-del-32-la-performance-finanziaria-delle-aziende/283071/

Facebook on Friday removed what it called a global network of more than 900 accounts, pages, and groups from its platform and Instagram that allegedly used deceptive practices to push pro-Trump narratives to about 55 million users. The network used fake accounts, artificial amplification, and, notably, profile photos of fake faces generated using artificial intelligence to spread polarizing, predominantly right-wing content around the Web, including on Twitter and YouTube.

It represents an alarming new development in the information wars, as it appears to be the first large-scale deployment of AI-generated images in a social network. In a report on the influence operation, researchers from disinformation groups Graphika and DFRLab noted that this was the first time they had seen the technology used to support an inauthentic social media campaign.

The images were used to enhance the authenticity of some of the 610 Facebook accounts, 156 groups, 89 pages, and 72 Instagram accounts associated with “The BL,” a digital news outlet that described itself as a “pure mountain spring, moistening the heart of every reader” on Facebook before it was removed. Facebook’s investigation connected The BL to The Epoch Times, a conservative media organization with ties to Chinese spiritual group Falun Gong and a history of aggressive support for Donald Trump.

Ostensibly a US-based media organization, The BL network’s pages were operated by users in Vietnam and the US, who, Facebook says, made widespread use of fake accounts to evade detection and funnel traffic to its own websites. A report by Graphika and DFRLab researchers found that the majority of The BL network’s fake accounts were used to maintain a cluster of over 80 groups and pages promoting President Trump. The accounts acting as administrators for these pages—which boasted names such as “America Needs President Trump,” “Trump for America’s President,” and “WE STAND WITH TRUMP & PENCE!”—were predominantly fake and created in Vietnam.

“Fake accounts served as the administrators of Facebook groups, increased the membership numbers of those groups, liked posts on the Pages, and posted large quantities of content from TheBL-related assets,” the report notes. “This structure constituted a large-scale artificial amplification factory whose only observable function was to boost content from TheBL and, to a lesser extent, from the Epoch Times.”

Many of these accounts used profile pictures that researchers said appeared to be created using neural networks trained on images of real human faces to create photos of people who did not actually exist. Researchers from Graphika and DFRLab found 18 images of faces in profile photos used by The BL network that they said had likely been generated using generative adversarial networks.

While the technology is capable of creating passable fake faces, it has a few obvious tells that tipped the researchers off. It works largely by mimicking visual patterns but doesn’t understand human anatomy and struggles with backgrounds and symmetrical features, like glasses or earrings.

The above image, from the Graphika report, was taken from the profile photo for “Alfonzo Macias,” an admin for one of the groups in The BL network. “Note the asymmetry in the glasses and also the poorly defined background,” the report says. “The authors checked this image with experts at the University Federico II of Naples, who assessed that this image was GAN-generated with 99.9 percent certainty.”

This story originally appeared on wired.com.

https://arstechnica.com/?p=1636319

More than 13,000 artificial intelligence mavens flocked to Vancouver this week for the world’s leading academic AI conference, NeurIPS. The venue included a maze of colorful corporate booths aiming to lure recruits for projects like software that plays doctor. Google handed out free luggage scales and socks depicting the colorful bikes employees ride on its campus while IBM offered hats emblazoned with “I ❤️A👁.”

Tuesday night, Google and Uber hosted well-lubricated, over-subscribed parties. At a bleary 8:30 the next morning, one of Google’s top researchers gave a keynote with a sobering message about AI’s future.

Blaise Aguera y Arcas praised the revolutionary technique known as deep learning that has seen teams like his get phones to recognize faces and voices. He also lamented the limitations of that technology, which involves designing software called artificial neural networks that can get better at a specific task by experience or seeing labeled examples of correct answers.

“We’re kind of like the dog who caught the car,” Aguera y Arcas said. Deep learning has rapidly knocked down some longstanding challenges in AI—but doesn’t immediately seem well suited to many that remain. Problems that involve reasoning or social intelligence, such as weighing up a potential hire in the way a human would, are still out of reach, he said. “All of the models that we have learned how to train are about passing a test or winning a game with a score [but] so many things that intelligences do aren’t covered by that rubric at all,” he said.

Hours later, one of the three researchers seen as the godfathers of deep learning also pointed to the limitations of the technology he had helped bring into the world. Yoshua Bengio, director of Mila, an AI institute in Montreal, recently shared the highest prize in computing with two other researchers for starting the deep learning revolution. But he noted that the technique yields highly specialized results; a system trained to show superhuman performance at one videogame is incapable of playing any other. “We have machines that learn in a very narrow way,” Bengio said. “They need much more data to learn a task than human examples of intelligence, and they still make stupid mistakes.”

Bengio and Aguera y Arcas both urged NeurIPS attendees to think more about the biological roots of natural intelligence. Aguera y Arcas showed results from experiments in which simulated bacteria adapted to seek food and communicate through a form of artificial evolution. Bengio discussed early work on making deep learning systems flexible enough to handle situations very different from those they were trained on, and made an analogy to how humans can handle new scenarios like driving in a different city or country.

https://arstechnica.com/?p=1632485

In an earlier deep learning article, we talked about how inference workloads—the use of already-trained neural networks to analyze data—can run on fairly cheap hardware, but running the training workload that the neural network “learns” on is orders of magnitude more expensive.

In particular, the more potential inputs you have to an algorithm, the more out of control your scaling problem gets when analyzing its problem space. This is where MACH, a research project authored by Rice University’s Tharun Medini and Anshumali Shrivastava, comes in. MACH is an acronym for Merged Average Classifiers via Hashing, and according to lead researcher Shrivastava, “[its] training times are about 7-10 times faster, and… memory footprints are 2-4 times smaller” than those of previous large-scale deep learning techniques.

In describing the scale of extreme classification problems, Medini refers to online shopping search queries, noting that “there are easily more than 100 million products online.” This is, if anything, conservative—one data company claimed Amazon US alone sold 606 million separate products, with the entire company offering more than three billion products worldwide. Another company reckons the US product count at 353 million. Medini continues, “a neural network that takes search input and predicts from 100 million outputs, or products, will typically end up with about 2,000 parameters per product. So you multiply those, and the final layer of the neural network is 200 billion parameters … [and] I’m talking about a very, very dead simple neural network model.”

At this scale, a supercomputer would likely need terabytes of working memory just to store the model. The memory problem gets even worse when you bring GPUs into the picture. GPUs can process neural network workloads orders of magnitude faster than general purpose CPUs can, but each GPU has a relatively small amount of RAM—even the most expensive Nvidia Tesla GPUs only have 32GB of RAM. Medini says, “training such a model is prohibitive due to massive inter-GPU communication.”

Instead of training on the entire 100 million outcomes—product purchases, in this example—Mach divides them into three “buckets,” each containing 33.3 million randomly selected outcomes. Now, MACH creates another “world,” and in that world, the 100 million outcomes are again randomly sorted into three buckets. Crucially, the random sorting is separate in World One and World Two—they each have the same 100 million outcomes, but their random distribution into buckets is different for each world.

With each world instantiated, a search is fed to both a “world one” classifier and a “world two” classifier, with only three possible outcomes apiece. “What is this person thinking about?” asks Shrivastava. “The most probable class is something that is common between these two buckets.”

At this point, there are nine possible outcomes—three buckets in World One times three buckets in World Two. But MACH only needed to create six classes—World One’s three buckets plus World Two’s three buckets—to model that nine-outcome search space. This advantage improves as more “worlds” are created; a three-world approach produces 27 outcomes from only nine created classes, a four-world setup gives 81 outcomes from 12 classes, and so forth. “I am paying a cost linearly, and I am getting an exponential improvement,” Shrivastava says.

Better yet, MACH lends itself better to distributed computing on smaller individual instances. The worlds “don’t even have to talk to one another,” Medini says. “In principle, you could train each [world] on a single GPU, which is something you could never do with a non-independent approach.” In the real world, the researchers applied MACH to a 49 million product Amazon training database, randomly sorting it into 10,000 buckets in each of 32 separate worlds. That reduced the required parameters in the model more than an order of magnitude—and according to Medini, training the model required both less time and less memory than some of the best reported training times on models with comparable parameters.

Of course, this wouldn’t be an Ars article on deep learning if we didn’t close it out with a cynical reminder about unintended consequences. The unspoken reality is that the neural network isn’t actually learning to show shoppers what they asked for. Instead, it’s learning how to turn queries into purchases. The neural network doesn’t know or care what the human was actually searching for; it just has an idea what that human is most likely to buy—and without sufficient oversight, systems trained to increase outcome probabilities this way can end up suggesting baby products to women who’ve suffered miscarriages, or worse.

https://arstechnica.com/?p=1632445

Imagine you’re trying to make a left turn onto a busy road. Car after car rolls past, keeping you trapped as your frustration rises. Finally, a generous driver decelerates enough to create a gap. A check of the traffic from the opposite direction, a quick bit of acceleration, and you’re successfully merged into traffic.

This same scene plays out across the world countless times a day. And it’s a situation where inferring both the physics and the motives of your fellow drivers is difficult, as evidenced by the fact that the United States sees 1.4 million accidents each year from drivers in the process of turning. Now imagine throwing autonomous vehicles into the mix. These are typically limited to evaluating only the physics and to make conservative decisions in situations where information is ambiguous.

Now, a group of computer scientists has figured out how to improve autonomous vehicle (AV) performance in these circumstances. The scientists have essentially given their AVs a limited theory of mind, allowing the vehicles to better interpret what the behaviors of their nearby human drives are telling them.

Mind the theory

Theory of mind comes so easily to us that it’s difficult to recognize how rare it is outside of our species. We’re easily able to recognize that our fellow humans have minds like our own, and we use that recognition to infer things like the state of their knowledge and their likely motivations. These inferences are essential to most of our social activities, driving included. While a friendly wave can make for an unambiguous signal that your fellow driver is offering you space in their lane, we can often make inferences based simply on the behavior of their car.

And, critically, autonomous vehicles aren’t especially good at this. In many cases, their own behavior doesn’t send signals back to other drivers. A study of accidents involving AVs in California indicated that over half of them involved the AV being rear-ended because a human driver couldn’t figure out what in the world it was doing. (Volvo, among others, is working to change that.)

It’s unrealistic to think that we’ll give AVs a full-blown theory of mind any time soon. AIs are simply not that advanced, and it would be excessive for cars, which only have to deal with a limited range of human behaviors. But a group of researchers at MIT and Delft University of Technology has decided that putting an extremely limited theory of mind in place for certain driving decisions, including turns and merges, should be possible.

The idea behind the researchers’ work, described in a new paper in PNAS, involves a concept called social value orientation, which is a way of measuring how selfish or community-oriented an individual’s actions are. While there are undoubtedly detailed surveys that can provide a meticulous description of a person’s social value orientation, autonomous vehicles generally won’t have the time to be giving their fellow drivers surveys.

So the researchers distilled social value orientation into four categories: altruists, who try to maximize the enjoyment of their fellow drivers; prosocial drivers, who try to take actions that allow all other drivers to benefit (which may occasionally involve selfishly flooring it); individualists, who maximize their own driving experience; and competitive drivers, who only care about having a better driving experience than those around them.

Value-oriented

The researchers developed a formula that would let them calculate the expected driving trajectory for each of these categories given the starting position of other cars. The autonomous vehicle was programmed to compare the trajectories of actual drivers to the calculated version and use that to determine which of the four categories the drivers were likely to be in. Given that classification, the vehicle could then project what their future actions would be. As the researchers wrote, “we extend the ability of AVs’ reasoning by incorporating estimates of the other drivers’ personality and driving style from social cues.”

This is substantially different from some game-theory work that’s been done in the area. That work has assumed that every driver is always maximizing their own gain; if altruism emerges, it’s only incidental to this maximization. This new work, in contrast, bakes altruistic behavior into its calculations and recognizes that drivers are complicated and may change their tendencies as situations evolve. In fact, previous studies had indicated that in contexts other than driving, about half of the people tested showed prosocial behavior, with another 40% being selfish.

With the system in place, the researchers obtained data on vehicle locations and trajectories as drivers merged onto a highway, a situation that often requires the generosity of fellow drivers. With the social value orientation system in place, the autonomous driver was able make more accurate predictions of its fellow drivers’ trajectories than it could without—prediction errors dropped by 25%. The system also worked on lane changes on crowded freeways, as well as turns into traffic.

Using these evaluations, the researchers could also make some inferences using the traffic patterns they had. For example, they found that a highway driver may start out selfishly following the car in front of them, shift to altruistic as they decelerate to allow a driver to merge, then switch right back to a selfish approach. Similarly, drivers facing a merge onto a freeway typically ended up being competitive—something you see every time a vehicle pulls out and slows down everyone who was stuck in the lane behind it.

While we’re still a long way off from giving autonomous vehicles a general AI or a full theory of mind, the research shows that you can get significant benefits from giving AVs a very limited one. And it’s a nice demonstration that if we want any autonomous system to integrate with something that’s currently a social activity, then paying attention to what social scientists have figured out about those activities can be incredibly valuable.

PNAS, 2019. DOI: 10.1073/pnas.1820676116  (About DOIs).

https://arstechnica.com/?p=1603679

Saturday afternoon (Nov. 16) at Supercomputing 2019, Intel launched a new programming model called oneAPI. Intel describes the necessity of tightly coupling middleware and frameworks directly to specific hardware as one of the largest pain points of AI/Machine Learning development. The oneAPI model is intended to abstract that tight coupling away, allowing developers to focus on their actual project and re-use the same code when the underlying hardware changes.

This sort of “write once, run anywhere” mantra is reminiscent of Sun’s early pitches for the Java language. However, Bill Savage, general manager of compute performance for Intel, told Ars that’s not an accurate characterization. Although each approach addresses the same basic problem—tight coupling to machine hardware making developers’ lives more difficult and getting in the way of code re-use—the approaches are very different.

When a developer writes Java code, the source is compiled to bytecode, and a Java Virtual Machine tailored to the local hardware executes that bytecode. Although many optimizations have improved Java’s performance in the 20+ years since it was introduced, it’s still significantly slower than C++ code in most applications—typically, anywhere from half to one-tenth as fast. By contrast, oneAPI is intended to produce direct object code with no or negligible performance penalties.

When we questioned Savage about oneAPI’s design and performance expectations, he distanced it firmly from Java, pointing out that there is no bytecode involved. Instead, oneAPI is a set of libraries that tie hardware-agnostic API calls directly to heavily optimized, low-level code that drives the actual hardware available in the local environment. So instead of “Java for Artificial Intelligence,” the high-level takeaway is more along the lines of “OpenGL/DirectX for Artificial Intelligence.”

For even higher-performance coding inside tight loops, oneAPI also introduces a new language variant called “Data Parallel C++” allowing even very low-level optimized code to target multiple architectures. Data Parallel C++ leverages and extends SYCL, a “single source” abstraction layer for OpenCL programming.

In its current version, a oneAPI developer still needs to target the basic hardware type he or she is coding for—for example, CPUs, GPUs, or FPGAs. Beyond that basic targeting, oneAPI keeps the code optimized for any supported hardware variant. This would, for example, allow users of a oneAPI-developed project to run the same code on either Nvidia’s Tesla v100 or Intel’s own newly released Ponte Vecchio GPU.

Ponte Vecchio is the first actual product in Intel’s new Xe GPU line and is targeted specifically at HPC supercomputing and data center use. Although neither Savage nor other Intel execs Ars spoke to had timelines or would speak to concrete products, one slide from Intel’s Supercomputing 2019 presentation clearly shows the Xe architecture as encompassing workstation, mobile, and gaming use—so there may be interesting times ahead for rivals in those spaces.

Savage told Ars that although the current version of oneAPI does still require developers to code for a particular architecture family—CPU, GPU, FPGA, etc—Intel plans for a future release to also allow automatic selection of the most optimal hardware type available.

The oneAPI toolkit is available for use and testing now at Intel Devcloud.

https://arstechnica.com/?p=1603199

Microsoft rose to dominance during the ’80s and ’90s thanks to the success of its Windows operating system running on Intel’s processors, a cosy relationship nicknamed “Wintel”.

Now Microsoft hopes that another another hardware–software combo will help it recapture that success—and catch rivals Amazon and Google in the race to provide cutting-edge artificial intelligence through the cloud.

Microsoft hopes to extend the popularity of its Azure cloud platform with a new kind of computer chip designed for the age of AI. Starting today, Microsoft is providing Azure customers with access to chips made by the British startup Graphcore.

Graphcore, founded in Bristol, UK, in 2016, has attracted considerable attention among AI researchers—and several hundred million dollars in investment—on the promise that its chips will accelerate the computations required to make AI work. Until now it has not made the chips publicly available or shown the results of trials involving early testers.

Microsoft, which put its own money into Graphcore last December as part of a $200 million funding round, is keen to find hardware that will make its cloud services more attractive to the growing number of customers for AI applications.

Unlike most chips used for AI, Graphcore’s processors were designed from scratch to support the calculations that help machines to recognize faces, understand speech, parse language, drive cars, and train robots. Graphcore expects it will appeal to companies running business-critical operations on AI, such as self-driving-car startups, trading firms, and operations that process large quantities of video and audio. Those working on next-generation AI algorithms may also be keen to explore the platform’s advantages.

Microsoft and Graphcore today published benchmarks that suggest the chip matches or exceeds the performance of the top AI chips from Nvidia and Google using algorithms written for those rival platforms. Code written specifically for Graphcore’s hardware may be even more efficient.

The companies claim that certain image-processing tasks work many times faster on Graphcore’s chips, for example, than on its rivals using existing code. They also say they were able to train a popular AI model for language processing, called BERT, at rates matching those of any other existing hardware.

BERT has become hugely important for AI applications involving language. Google recently said that it is using BERT to power its core search business. Microsoft says it is now using Graphcore’s chips for internal AI research projects involving natural language processing.

Karl Freund, who tracks the AI chip market at Moor Insights, says the results show the chip is cutting-edge but still flexible. A highly-specialized chip could outperform one from Nvidia or Google but would not be programmable enough for engineers to develop new applications. “They’ve done a good job making it programmable, he says. “Good performance in both training and inference is something they’ve always said they would do, but it is really, really hard.”

Freund adds that the deal with Microsoft is crucial for Graphcore’s business, because it provides an on-ramp for customers to try the new hardware. The chip may well be superior to existing hardware for some applications, but it takes a lot of effort to redevelop AI code for a new platform. With a couple of exceptions, Freund says, the chip’s benchmarks are not eye-popping enough to lure companies and researchers away from the hardware and software they are already comfortable using.

Graphcore has created a software framework called Poplar, which allows existing AI programs to be ported to its hardware. Plenty of existing algorithms may still be better-suited to software that runs on top of rival hardware, though. Google’s Tensorflow AI software framework has become the de facto standard for AI programs in recent years, and it was written specifically for Nvidia and Google chips. Nvidia is also expected to release a new AI chip next year, which is likely to have better performance.

Nigel Toon, cofounder and CEO of Graphcore, says the companies began working together a year after his company’s launch, through Microsoft Research Cambridge in the UK. His company’s chips are especially well-suited to tasks that involve very large AI models or temporal data, he says. One customer in finance supposedly saw a 26-fold performance boost in an algorithm used to analyze market data thanks to Graphcore’s hardware.

A handful of other, smaller companies also announced today that they are working with Graphcore chips through Azure. This includes Citadel, which will use the chips to analyze financial data, and Qwant, a European search engine that wants the hardware to run an image-recognition algorithm known as ResNext.

The AI boom has already shaken up the market for computer chips in recent years. The best algorithms perform parallel mathematical computations, which can be done more effectively on a graphics chips (or GPUs) that have hundreds of simple processing cores as opposed to conventional chips (CPUs) that have a few complex processing cores.

The GPU-maker Nvidia has ridden the AI wave to riches, and Google announced in 2017 that it would develop its own chip, the Tensor Processing Unit, which is architecturally similar to a GPU but optimized for Tensorflow.

Graphcore’s chips, which it calls intelligence processing units (IPUs), have many more cores than GPUs or TPUs. They also feature memory on the chip itself, which removes a bottleneck that comes with moving data onto a chip for processing and off again.

Facebook is also working on its own AI chips. Microsoft has previously touted reconfigurable chips made by Intel and customized by its engineers for AI applications. A year ago, Amazon revealed it was also getting into chipmaking, but with a more general-purpose processor optimized for Amazon’s cloud services.

More recently, the AI boom has sparked a flurry of startup hardware companies to develop more specialized chips. Some of these are optimized for specific applications such as autonomous driving or surveillance cameras. Graphcore and a few others offer much more flexible chips, which are crucial for developing AI applications but also much more challenging to produce. The company’s last investment round gave the company a valuation of $1.7 billion.

Graphcore’s chips might first find traction with top AI experts who are able to write the code needed to exploit their benefits. Several prominent AI researchers have invested in Graphcore, including Demis Hassabis, cofounder of DeepMind, Zoubin Ghahramani, a professor at the University of Cambridge and the head of Uber’s AI lab, and Peiter Abbeel, a professor at UC Berkeley who specializes in AI and robotics. In an interview with WIREDlast December, AI visionary Geoffrey Hinton discussed the potential for Graphcore chips to advance fundamental research.

Before long, companies may be tempted to try out the latest thing, too. As Graphcore’s CEO Toon says, “Everybody’s trying to innovate, trying to find an advantage.”

This story originally appeared on wired.com.

Listing image by Graphcore

https://arstechnica.com/?p=1602809