Month: May 2019

DefinedCrowd, the Startup Battlefield alumnus that produces and refines data for AI-training purposes, has just debuted iOS and Android apps for its army of human annotators. It should help speed up a process that the company already touts as one of the fastest in the industry.

It’s no secret that AI relies almost totally on data that has been hand-annotated by humans, pointing out objects in photos, analyzing the meaning of sentences or expressions, and so on. Doing this work has become a sort of cottage industry, with many annotators doing it part time or between other jobs.

There’s a limit, however, to what you can do if the interface you must use to do it is only available on certain platforms. Just as others occasionally answer an email or look over a presentation while riding the bus or getting lunch, it’s nice to be able to do work on mobile — essential, really, at this point.

To that end DefinedCrowd has made its own app, which shares the Neevo branding of the company’s annotation community, that lets its annotators work whenever they want, tackling image or speech annotation tasks on the go. It’s available on iOS and Android starting today.

It’s a natural evolution of the market, CEO Daniela Braga told me. There’s a huge demand for this kind of annotation work, and it makes no sense to restrict the schedules or platforms of the people doing it. She suggested everyone in the annotation space would have apps soon, just as every productivity or messaging service does. And why not?

The company is growing quickly, going from a handful of employees to over a hundred, spread over its offices in Lisbon, Porto, Seattle, and Tokyo. The market, likewise, is exploding as more and more companies find that AI is not just applicable to what they do, but not out of their reach.

Source: Tech Crunch

Biofourmis, a Singapore-based startup pioneering a distinctly tech-based approach to the treatment of chronic conditions, has raised a $35 million Series B round for expansion.

The round was led by Sequoia India and MassMutual Ventures, the VC fund from Massachusetts Mutual Life Insurance Company. Other investors who put in include EDBI, the corporate investment arm of Singapore’s Economic Development Board, China-based healthcare platform Jianke and existing investors Openspace Ventures, Aviva Ventures and SGInnovate, a Singapore government initiative for deep tech startups. The round takes Biofourmis to $41.6 million raised to date, according to Crunchbase.

This isn’t your typical TechCrunch funding story.

Biofourmis CEO Kuldeep Singh Rajput moved to Singapore to start a PhD, but he dropped out to start the business with co-founder Wendou Niu in 2015 because he saw the potential to “predict disease before it happens,” he told TechCrunch in an interview.

AI-powered specialist post-discharge care

There are a number of layers to Biofourmis’ work, but essentially it uses a combination of data collected from patients and an AI-based system to customize treatments for post-discharge patients. The company is focused on a range of therapeutics, but its most advanced is cardiac, so patients who have been discharged after heart failure or other heart-related conditions.

With that segment of patients, the Biofourmis platform uses a combination of data from sensors — medical sensors rather than consumer wearables, which are worn 24/7 — and its tech to monitor patient health, detect problems ahead of time and prescribe an optimum treatment course. That information is disseminated through companion mobile apps for patients and caregivers.

That’s to say that medicine works differently on different people, so by collecting and monitoring data and crunching numbers, Biofourmis can provide the best drug to help optimize a patient’s health through what it calls a ‘digital pill.’ That’s not Matrix-style futurology, it’s more like a digital prescription that evolves based on the needs of a patient in real-time. It plans to use a network of medical delivery platforms, including Amazon-owned PillPack, to get the drugs to patients within hours.

Yes, that’s future tense because Biofourmis is waiting on FDA approval to commercialize its service. That’s expected to come by the end of this year, Singh Rajput told TechCrunch. But he’s optimistic given clinical trials, which have covered some 5,000 patients across 20 different sites.

On the tech side, Singh Rajput said Biofourmis has seen impressive results with its predictions. He cited tests in the U.S. which enabled the company to “predict heart failure 14 days in advance” with around 90 percent sensitivity. That was achieved using standard medical wearables at the cost of hundreds of dollars, rather than thousands with advanced kit such as Heartlogic from Boston Scientific — although the latter has a longer window for predictions.

The type of disruption that Biofourmis might appear to upset the applecart for pharma companies, but Singh Rajput maintains that the industry is moving towards a more qualitative approach to healthcare because it has been hard to evaluate the performance of drugs and price them accordingly.

“Today, insurance companies are blinded not having transparency on how to price drugs,” he said. “But there are already 50 drugs in the market paying based on outcomes so the market is moving in that direction.”

Outcome-based payments mean insurance firms reimburse all outcomes based on the performance of the drugs, in other words how well patients recover. The rates vary, but a lack of reduction in remission rates can see insurers lower their payouts because drugs aren’t working as well as expected.

Singh Rajput believes Biofourmis can level the playing field and added more granular transparency in terms of drug performance. He believes pharma companies are keen to show their products perform better than others, so over the long-term that’s the model Biofourmis wants to encourage.

Indeed, the confidence is such that Biofourmis intends to initially go to market via pharma companies, who will sell the package into clinics bundled with their drugs, before moving to work with insurance firms once traction is gained. While the Biofourmis is likely to be bundled with initial medication, the company will take a commission of 5-10 percent on the recommended drugs sold through its digital pill.

Doubling down on the US

With its new money, Biofourmis is doubling down on that imminent commercialization by relocating its headquarters to Boston. It will retain its presence in Singapore, where it has 45 people who handle software and product development, but the new U.S. office is slated to grow from 14 staff right now to up to 120 by the end of the year.

“The U.S. has been a major market focus since day one,” Singh Rajput said. “Being closer to customers and attracting the clinical data science pool is critical.”

While he praised Singapore and said the company remains committed to the country — adding EDBI to its investors is certainly a sign — he admitted that Boston, where he once studied, is a key market for finding “data scientists with core clinical capabilities.”

That expansion is not only to bring the cardio product to market, but also to prepare products to cover other therapeutics. Right now, it has six trials in place that cover pain, orthopedics and oncology. There are also plans to expand in other markets outside of the U.S, and in particular Singapore and China, where Biofourmis plans to lead on Jianke.

Not lacking in confidence, Singh Rajput told TechCrunch that the company is on course to reach a $1 billion valuation when it next raises funding, that’s estimated as 18 months away and the company isn’t saying how much it is worth today.

Singh Rajput did confirm, however, that the round was heavily oversubscribed, and that the startup rebuffed investment offers from pharma companies in order to “avoid a conflict of interest and stay neutral.”

He is also eying a future IPO, which is tentatively set for 2023 — although by then, Singh Rajput said, Biofourmis would need at least two products in the market.

There’s a long way to go before then, but this round has certainly put Biofourmis and its digital pill approach on the map within the tech industry.

Source: Tech Crunch

It’s a bit strange to hear that the world’s leading social network is pursuing research in robotics rather than, say, making search useful, but Facebook is a big organization with many competing priorities. And while these robots aren’t directly going to affect your Facebook experience, what the company learns from them could be impactful in surprising ways.

Though robotics is a new area of research for Facebook, its reliance on and bleeding-edge work in AI are well known. Mechanisms that could be called AI (the definition is quite hazy) govern all sorts of things, from camera effects to automated moderation of restricted content.

AI and robotics are naturally overlapping magisteria — it’s why we have an event covering both — and advances in one often do the same, or open new areas of inquiry, in the other. So really it’s no surprise that Facebook, with its strong interest in using AI for a variety of tasks in the real and social media worlds, might want to dabble in robotics to mine for insights.

What then could be the possible wider applications of the robotics projects it announced today? Let’s take a look.

Learning to walk from scratch

Walking is a surprisingly complex action, or series of actions, especially when you’ve got six legs, like the robot used in this experiment. You can program in how it should move its legs to go forward, turn around, and so on, but doesn’t that feel a bit like cheating? After all, we had to learn on our own, with no instruction manual or settings to import. So the team looked into having the robot teach itself to walk.

This isn’t a new type of research — lots of roboticists and AI researchers are into it. Evolutionary algorithms (different but related) go back a long way, and we’ve already seen interesting papers like this one:

By giving their robot some basic priorities like being “rewarded” for moving forward, but no real clue how to work its legs, the team let it experiment and try out different things, slowly learning and refining the model by which it moves. The goal is to reduce the amount of time it takes for the robot to go from zero to reliable locomotion from weeks to hours.

What could this be used for? Facebook is a vast wilderness of data, complex and dubiously structured. Learning to navigate a network of data is of course very different from learning to navigate an office — but the idea of a system teaching itself the basics on a short timescale given some simple rules and goals is shared.

Learning how AI systems teach themselves, and how to remove roadblocks like mistaken priorities, cheating the rules, weird data-hoarding habits and other stuff is important for agents meant to be set loose in both real and virtual worlds. Perhaps the next time there is a humanitarian crisis that Facebook needs to monitor on its platform, the AI model that helps do so will be informed by the autodidactic efficiencies that turn up here.

Leveraging “curiosity”

This work is a little less visual, but more relatable. After all, everyone feels curiosity to a certain degree, and while we understand that sometimes it kills the cat, most times it’s a drive that leads us to learn more effectively. Facebook applied the concept of curiosity to a robot arm being asked to perform various ordinary tasks.

Now, it may seem odd that they could imbue a robot arm with “curiosity,” but what’s meant by that term in this context is simply that the AI in charge of the arm — whether it’s seeing or deciding how to grip, or how fast to move — is given motivation to reduce uncertainty about that action.

That could mean lots of things — perhaps twisting the camera a little while identifying an object gives it a little bit of a better view, improving its confidence in identifying it. Maybe it looks at the target area first to double check the distance and make sure there’s no obstacle. Whatever the case, giving the AI latitude to find actions that increase confidence could eventually let it complete tasks faster, even though at the beginning it may be slowed by the “curious” acts.

What could this be used for? Facebook is big on computer vision, as we’ve seen both in its camera and image work and in devices like Portal, which (some would say creepily) follows you around the room with its “face.” Learning about the environment is critical for both these applications and for any others that require context about what they’re seeing or sensing in order to function.

Any camera operating in an app or device like those from Facebook is constantly analyzing the images it sees for usable information. When a face enters the frame, that’s the cue for a dozen new algorithms to spin up and start working. If someone holds up an object, does it have text? Does it need to be translated? Is there a QR code? What about the background, how far away is it? If the user is applying AR effects or filters, where does the face or hair stop and the trees behind begin?

If the camera, or gadget, or robot, left these tasks to be accomplished “just in time,” they will produce CPU usage spikes, visible latency in the image, and all kinds of stuff the user or system engineer doesn’t want. But if it’s doing it all the time, that’s just as bad. If instead the AI agent is exerting curiosity to check these things when it senses too much uncertainty about the scene, that’s a happy medium. This is just one way it could be used, but given Facebook’s priorities it seems like an important one.

Seeing by touching

Although vision is important, it’s not the only way that we, or robots, perceive the world. Many robots are equipped with sensors for motion, sound, and other modalities, but actual touch is relatively rare. Chalk it up to a lack of good tactile interfaces (though we’re getting there). Nevertheless, Facebook’s researchers wanted to look into the possibility of using tactile data as a surrogate for visual data.

If you think about it, that’s perfectly normal — people with visual impairments use touch to navigate their surroundings or acquire fine details about objects. It’s not exactly that they’re “seeing” via touch, but there’s a meaningful overlap between the concepts. So Facebook’s researchers deployed an AI model that decides what actions to take based on video, but instead of actual video data, fed it high-resolution touch data.

Turns out the algorithm doesn’t really care whether it’s looking at an image of the world as we’d see it or not — as long as the data is presented visually, for instance as a map of pressure on a tactile sensor, it can be analyzed for patterns just like a photographic image.

What could this be used for? It’s doubtful Facebook is super interested in reaching out and touching its users. But this isn’t just about touch — it’s about applying learning across modalities.

Think about how, if you were presented with two distinct objects for the first time, it would be trivial to tell them apart with your eyes closed, by touch alone. Why can you do that? Because when you see something, you don’t just understand what it looks like, you develop an internal model representing it that encompasses multiple senses and perspectives.

Similarly, an AI agent may need to transfer its learning from one domain to another — auditory data telling a grip sensor how hard to hold an object, or visual data telling the microphone how to separate voices. The real world is a complicated place and data is noisier here — but voluminous. Being able to leverage that data regardless of its type is important to reliably being able to understand and interact with reality.

So you see that while this research is interesting in its own right, and can in fact be explained on that simpler premise, it is also important to recognize the context in which it is being conducted. As the blog post describing the research concludes:

We are focused on using robotics work that will not only lead to more capable robots but will also push the limits of AI over the years and decades to come. If we want to move closer to machines that can think, plan, and reason the way people do, then we need to build AI systems that can learn for themselves in a multitude of scenarios — beyond the digital world.

As Facebook continually works on expanding its influence from its walled garden of apps and services into the rich but unstructured world of your living room, kitchen, and office, its AI agents require more and more sophistication. Sure, you won’t see a “Facebook robot” any time soon… unless you count the one they already sell, or the one in your pocket right now.

Source: Tech Crunch