Why the explainable AI market is growing rapidly

Were you unable to attend Transform 2022? Check out all of the summit sessions in our on-demand library now! Watch here.

Powered by digital transformation, there seems to be no ceiling to the heights organizations will reach in the next few years. One of the notable technologies helping enterprises scale these new heights is artificial intelligence (AI). But as AI advances with numerous use cases, there’s been the persistent problem of trust: AI is still not fully trusted by humans. At best, it’s under intense scrutiny and we’re still a long way from the human-AI synergy that’s the dream of data science and AI experts.

One of the underlying factors behind this disjointed reality is the complexity of AI. The other is the opaque approach AI-led projects often take to problem-solving and decision-making. To solve this challenge, several enterprise leaders looking to build trust and confidence in AI have turned their sights to explainable AI (also called XAI) models.

Explainable AI enables IT leaders — especially data scientists and ML engineers — to query, understand and characterize model accuracy and ensure transparency in AI-powered decision-making.   

Why companies are getting on the explainable AI train

With the global explainable AI market size estimated to grow from $3.5 billion in 2020 to $21 billion by 2030, according to a report by ResearchandMarkets, it’s obvious that more companies are now getting on the explainable AI train. Alon Lev, CEO at Israel-based Qwak, a fully-managed platform that unifies machine learning (ML) engineering and data operations, told VentureBeat in an interview that this trend “may be directly related to the new regulations that require specific industries to provide more transparency about the model predictions.” The growth of explainable AI is predicated on the need to build trust in AI models, he said.


MetaBeat 2022

MetaBeat will bring together thought leaders to give guidance on how metaverse technology will transform the way all industries communicate and do business on October 4 in San Francisco, CA.

Register Here

He further noted that another growing trend in explainable AI is the use of SHAP (SHapley Additive exPlanations) values — which is a game theoretic approach to explaining the outcome of ML models.

“We are seeing that our fintech and healthcare customers are more involved in the topic as they are sometimes required by regulation to explain why a model gave a specific prediction, how the prediction came about and what factors were considered. In these specific industries, we are seeing more models with explainable AI built in by default,” he added.

A growing marketplace with tough problems to solve

There’s no dearth of startups in the AI and MLops space, with a long list of startups developing MLops solutions including Comet,, ZenML, Landing AI, Domino Data Lab, Weights and Biases and others. Qwak is another startup in the space that focuses on automating MLops processes and allows companies to manage models the moment they are integrated with their products.  

With the claim to accelerate MLops potential using a different approach, Domino Data Lab is focused on building on-premises systems to integrate with cloud-based GPUs as part of Nexus — its enterprise-facing initiative built in collaboration with Nvidia as a launch partner. ZenML in its own right offers a tooling and infrastructure framework that acts as a standardization layer and allows data scientists to iterate on promising ideas and create production-ready ML pipelines.

Comet prides itself on the ability to provide a self-hosted and cloud-based MLops solution that allows data scientists and engineers to track, compare and optimize experiments and models. The aim is to deliver insights and data to build more accurate AI models while improving productivity, collaboration and explainability across teams.

In the world of AI development, the most perilous journey to take is the one from prototyping to production. Research has shown that the majority of AI projects never make it into production, with an 87% failure rate in a cutthroat market. However, this doesn’t in any way imply that established companies and startups aren’t having any success at riding the wave of AI innovation.

Addressing Qwak’s challenges when deploying its ML and explainable AI solutions to users, Lev said while Qwak doesn’t create its own ML models, it provides the tools that empower its customers to efficiently train, adapt, test, monitor and productionize the models they build. “The challenge we solve in a nutshell is the dependency of the data scientists on engineering tasks,” he said.

By shortening the lifespan of the model buildup via taking away the underlying drudgery, Lev claims Qwak helps both data scientists and engineers deploy ML models continuously and automate the process using its platform.

Qwak’s differentiators

In a tough marketplace with various competitors, Lev claims Qwak is the only MLops/ML engineering platform that covers the full ML workflow from feature creation and data preparation through to deploying models into production.

“Our platform is simple to use for both data scientists and engineers, and the platform deployment is as simple as a single line of code. The build system will standardize your project’s structure and help data scientists and ML engineers generate auditable and retrainable models. It will also automatically version all models’ code, data and parameters, building deployable artifacts. On top of that, its model version tracks disparities between multiple versions, warding off data and concept drift.”

Founded in 2021 by Alon Lev (former VP of data operations at Payoneer), Yuval Fernbach (former ML specialist at Amazon), Ran Romano (former head of data and ML engineering at and Lior Penso (former business development manager at IronSource), the team at Qwak claims to have upended the race and approach to getting the explainable AI market ready.

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.

Repost: Original Source and Author Link


AI and computer vision powers growing shop-and-go platform

Were you unable to attend Transform 2022? Check out all of the summit sessions in our on-demand library now! Watch here.

AI and computer vision were not necessarily top-of-mind for Sodexo, a food and facilities management company that runs over 400 university dining programs, which was looking for a future-forward, seamless experience to offer students instead of the usual buffet meal options.

All the company knew is that they wanted something like Amazon Go’s cashierless, shop-and-go stores. That is, where shoppers can walk in, pick items off the shelves, and leave without standing in line at the cashier or suffering through swiping codes at the self-checkout. 

“Students today want things they can partially or fully prepare in their room or apartment, with organic, highly-local options,” said Kevin Rettle, global vice president product development and digital innovation at Sodexo. “We also wanted to remove friction, but many solutions still require the interaction of the guest with a cashier – this generation really doesn’t want to talk to a lot of people in their service interactions.” 

For the University of Denver, Sodexo chose the San Jose-based AiFi, which offers a frictionless and cashierless AI-powered retail solution. Its flexibility (the company says it can deploy two stores per week) and diverse locations (sports stadiums, music festivals, grocery store chains, college campuses and more) make it unique, explained Steve Gu, who cofounded AiFi in 2016 with his wife, Ying Zheng. Both Gu and Zheng have Ph.D.s in computer vision and spent time at Apple and Google.

AiFi, which is powered only by cameras and computer vision technology, announced today that it now boasts a total of 80 checkout-free stores worldwide, partnering with retailers including Carrefour, Aldi, Loop and Verizon. It has also opened 53 Zabka stores in Poland and 2 NFL stores. Gu maintains this is an industry benchmark for how this technology can scale in a way that Amazon Go, which has more than 42 stores, cannot.

Cameras and computer vision, not sensors

Amazon Go’s stores are retrofitted with specialized cameras, sensors, and weighted shelves, Gu explained. “That makes the solution very expensive and hard to scale,” he said. Instead, AiFi uses the “cheapest-possible off-the-shelf cameras,” combined with what he says is the real power: Computer vision. 

AiFi deploys sophisticated AI models through a large number of cameras placed across the ceiling, Gu said, in order to understand everything happening in the shop. Cameras track customers throughout their shopping journey, while computer vision recognizes products and detects different activities, including putting items onto or grabbing items off the shelves.

Beneath the platform’s hood are neural network models specifically developed for people-tracking as well as activity and product recognition. AiFi also developed advanced calibration algorithms that allow the company to re-create the shopping environment in 3D.

AiFi also leverages simulated datasets. “We spend quite a lot of effort building those simulated environments so we can train the AI algorithms and the models inside them,” Gu said. “That really helps us develop those models faster and make them more scalable.” 

In a simulated world, he explained, you can easily adjust human shapes and characteristics, as well as the shelf layout and the look of the product. You can create a cluttered, crowded store environment or one that is neat and orderly. “Things that cannot be done in the real world can be easily done in a simulated world,” he said. “The AI can learn about those scenarios and will then be able to perform or outperform in a real setting.” 

Computer vision that is constantly evolving

AiFi’s system is evolving and will improve over time, Gu continued, citing current challenges including the ability for the platform to recognize small items such as gum or lipstick.

“If they are not placed in the right place, it’s very hard for the computer vision to discern what it is,” he said. There are also issues related to items with similar looks and textures. “If they are placed together in adjacent spaces it sometimes causes confusion for the cameras and computer vision to recognize these products,” he said. “But the good thing is that it’s not purely based on the visual texture – you also have the 3D scene geometry, the location, the context as well.”  

There also are current limitations to the size of the store and the number of people it can track. “The question is can the solution also be scalable to super centers of 100,000 square feet?” he said. “Also, the system is able to track hundreds of people shopping simultaneously in a shop environment. But in order for that to further scale, to track thousands of people, with very complex shopping behavior, that’s something that is still a work in progress.”

To enter an AiFi-powered store, shoppers don’t need a biometric scan or an AiFi app — they can swipe a credit card or use the retailer’s app. At the University of Denver, for example, Sodexo wanted a partner that was agnostic to the front end. “We were able to use our wallet and payment processing, and tie the AiFi technology, the cameras, and the AI into our system,” said Rettle.

Consumer adoption is key

“From a product ownership perspective, you always kind of hold your breath. Is it going to work?” he said. But ultimately, at the University of Denver the students immediately took to the AiFi concept.

“We didn’t have to teach any of the students what to do,” he said. “They get it without having a bunch of prompts.”

Critics in the retail space also predicted the AiFi technology would be a “loss-prevention nightmare — that the students will figure out how to game the system,” Rettle said. Instead, the current accuracy rate for the AiFi solution is 98.3% and the shrink rate (what shoppers walk out without paying for) has actually declined, he said.

Some products don’t quite work yet with AiFi’s solution, Rettle admits, including college and fan “swag.” “The platform still has to understand consumer behavior around that, which will certainly evolve with the technology,” he said.

Rettle also said he doesn’t envision a campus or stadium that could shift to 100% autonomous retail. “For us it’s something that complements,” he said. “But I see a strong future in terms of being able to continue to deploy and drive ubiquity with the solution based on consumer acceptance.”

For Gu, AiFi’s potential is “huge,” with over a dozen new stores in the works and a growing partnership with Microsoft as an independent software vendor partner (AiFi runs its solution on Azure). “You’re going to see a lot of autonomous retail in a variety of verticals — not just stadiums, festivals and universities, but offices, movie theaters and other spaces,” he said.

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.

Repost: Original Source and Author Link


Google Play Pass adds Dead Cells, KOTOR to its growing list

When Apple launched its Arcade subscription service, it was expected that Google would also soon follow suit. Unlike Apple Arcade, Google Play Pass, in theory, includes not just games but also apps, though the focus, unsurprisingly, has been on the entertainment side of mobile software. So when Google announced the new members of its growing Google Play Pass family, it’s no surprise that they are almost all made up of games, ranging from some obscure titles to popular hits like Dead Cells.

The idea behind Google Play Pass is that you simply pay a $5 fee per month to get access to some apps and games that would normally cost a lot more than that when purchased individually. For games, it also brings the advantage of having no ads, at least for games that offer that paid option, and also no need for in-app purchases. Of course, that only works if the list includes highly-rated and desired titles and, for better or worse, the almost 800-strong roster is quite the mixed bag.

On the one hand, you have titles like the popular platformer Dead Cells that would have cost $9 plus IAPs without Google Play Pass. March also saw the addition of the classic Star Wars: Knights of the Old Republic, more popularly known as KOTOR, a $10 port of a PC RPG, Football Manager 2021, the latest installment in the simulation series, normally costing $9.

There are, however, also some lesser-known games, like EVO ISLAND or Funky Karts. Since February, Google Play Pass has also added quite a selection of education apps for kids, though parents might want to first check how Google Play Pass apps mix with parental controls.

Many will probably point out the sometimes questionable screening process of which apps and games go into the Google Play Pass collection. Not too long ago, that list included an app that was discovered to be carrying malware. At the same time, the sheer volume of apps and games available in the subscription suggests you won’t get tired from testing them all out without having to pay for them first.

Repost: Original Source and Author Link


Microsoft’s Surface sales near $2 billion, but its Azure cloud business is growing faster

Cloud, cloud, cloud, followed by Office, gaming, Surface and then Windows: By now, Microsoft has laid out its priorities, and the financial results Microsoft reported Wednesday afternoon reflected this.

For consumers, the high point of Microsoft’s fourth-quarter calendar 2019 results was the fact that Microsoft reported $1.98 billion in Surface sales alone—almost, but not quite, making Surface a $2 billion-dollar business. Interestingly, chief financial officer Amy Hood implied that sales could have been higher, referring to unexplained “execution challenges” in the consumer portion of the Surface business.

Overall, Microsoft reported profits of $11.6 billion during the second quarter of its fiscal 2020 calendar, up 38 percent from a year ago, from revenue of $36.9 billion, up 14 percent from the same period. 

The More Personal Computing business that includes Windows, Xbox, and Surface still generates the most revenue at $13.2 billion. However, it reported just two-percent growth—and that was better than expected, Hood said, due to strong Windows PC sales. 

Microsoft’s growth is primarily being driven by the Intelligent Cloud business, and specifically Microsoft Azure. The Intelligent Cloud business recorded $11.9 billion in revenue, growing by a whopping 27 percent. Within that business, Azure grew by 62 percent just by itself. In all, Intelligent Cloud once again edged out Microsoft’s Productivity and Business Processes ($11.8 billion) in terms of revenue. 

Microsoft’s Productivity and Business Processes—Office and Microsoft Dynamics, plus related services—saw 17-percent growth, fueled by 20-percent growth in Office 365 consumer revenue and 37.2 million consumer subscribers.

Uncertainty in consumer business

Microsoft’s MPC business faces more uncertainty than it has in the past. Revenue is expected to drop fairly substantially during the first quarter, Microsoft said, due to a variety of factors. The first, of course, is seasonality: Consumers buy far less during the first calendar quarter than they do during the holidays. Then there are flagging parts of the business, such as search advertising revenue, which dropped from 14-percent growth a year ago to just 6 percent.

Microsoft also acknowledged that the PC market supply chain is changing to adjust for reduced demand. The corporate PC market hastily bought new Windows 10 PCs in advance of the January 14 deadline that just passed, when Microsoft stopped supporting Windows 7. Windows OEM revenue in the current quarter might grow just by low to mid-single digits, Hood said. Overall, the MPC revenue outlook ranges between $10.75 billion and $12.05 billion, a wider spread than normal. Hood called out the uncertainty of the public coronavirus health scare in China, too.

Repost: Original Source and Author Link


For Pokemon’s 25th anniversary, maybe the games should do a little growing of their own

Pokemon is celebrating a big milestone this year, as the franchise is celebrating the 25th anniversary of the release of Pokemon Red and Green in Japan. It’s a big anniversary, to be sure, and it’s prompting not only fans but also Game Freak and The Pokemon Company themselves to take a look back at the history of Pokemon. While the franchise has experienced popularity that’s waxed and waned throughout the years, today Pokemon is as popular as it has ever been, if not more. It also seems to be stuck in the rut of iteration, at least as far as the games are concerned.

Pokemon fans will no doubt remember all of the controversy surrounding Pokemon Sword and Shield. To my eye, there were two different sets of complaints with Sword and Shield. The first revolved around Game Freak’s decision to the cull the Pokedex, marking the first time in franchise history that a main series release wouldn’t feature all of the Pokemon release up to that point. People were angry about that decision, and I think a lot of resentment toward Sword and Shield still persists because of it.

Some of those people who were angry about the Pokedex probably skipped Sword and Shield, but then we have the people who did buy the games and walked away feeling underwhelmed by them. Those feelings may very well have been sparked by the culled Pokedex, but even though I didn’t particularly mind the missing Pokemon, I too was one of those people who just wound up feeling underwhelmed by the whole experience.

I will say that Pokemon Sword and Shield were still fun enough for what they were, and they weren’t awful games by really any stretch. But even with new introductions like Dynamaxing and the Wild Area, it was impossible for me to shake this feeling that I’ve played this game many times before. Pokemon is not a franchise known for taking risks – especially not in recent years – and nowhere was that more apparent than in Sword and Shield.

To be clear, Sword and Shield were not the first Pokemon games that players took issue with, though because of the Pokedex controversy, the pushback we saw from fans was a lot more intense than it usually is. Even though I really enjoyed Sun and Moon and later Ultra Sun and Ultra Moon, I think a lot of people would argue that the Pokemon series started treading water when they made the transition to 3D with Pokemon X and Y. Ever since those titles, it seems that Game Freak is content to introduce new battle mechanics or slightly tweak the Pokemon formula in one way or another (EXP share for the whole party, the Island challenge in Sun and Moon, etc) as their way of “changing” Pokemon from generation to generation.

Sometimes, people really like those changes. Ask a Pokemon fan what they think of Mega Evolution and they’ll probably talk your ear off about how those were way better than Z-Moves or Dynamaxing and they just don’t understand why Game Freak had to get rid of them. But even when these battle mechanics and small tweaks are well-received, it never really feels like Pokemon fans are completely satisfied with modern games in the same way they were with, say, Pokemon HeartGold and SoulSilver or Black 2 and White 2.

When compared to Nintendo’s other marquee franchises, namely Zelda and Mario, Pokemon definitely doesn’t stack up on a game-by-game basis. That’s particularly true now that we’re living in a post-Breath of the Wild world, where Nintendo managed to flip the Zelda franchise on its head but still deliver an excellent game that’s arguably the best title in the series. And when it comes to Mario, Nintendo has always found a way to make each game feel distinct while keeping that familiar core gameplay loop that hooked us in the first place.

I don’t think it would take that much to change the Pokemon series for the better. One thing that I think would go a long way toward satisfying veteran players while keeping the games accessible to newcomers or young players would be the introduction of difficulty levels. Pokemon fans have been making up their own challenge modes since the Pokemon games were introduced, so why not give those who want a more challenging experience the option of picking a harder difficulty?

There are countless small yet impactful changes that Game Freak could make to help Pokemon feel fresh while not changing the core gameplay loop at all. Pokemon fans have wishlists that are miles long and packed with the features they’d love to see, and while I’m not saying that I think Pokemon games should suddenly be developed by committee, players have come up with some really good ideas throughout the years. Listening to fan feedback could help vastly improve the Pokemon series, but for years now, Game Freak has never felt like it cares to listen to fans.

Then again, maybe the joke’s on me here, because despite feeling underwhelmed with Sword and Shield, those games still went on to sell more than 20 million units life-to-date, placing them among the best-selling games on the Switch. I’ll admit that with sales numbers like that, Game Freak doesn’t have much incentive to change up the Pokemon formula, but I don’t think it would take a whole lot to make the series much more appealing to returning players as well as newcomers. That’s why, for Pokemon’s 25th anniversary, it would be nice to see the series do a little growing of its own.

Repost: Original Source and Author Link