OctoML, a startup founded by the team behind the Apache TVM machine learning compiler stack project, today announced that it has raised a $15 million Series A round led by Amplify, with participation from Madrone Ventures, which led its $3.9 million seed round. The core idea behind OctoML and TVM is to use machine learning to optimize machine learning models so they can more efficiently run on different types of hardware.
“There’s been quite a bit of progress in creating machine learning models,” OctoML CEO and University of Washington professor Luis Ceze told me.” But a lot of the pain has moved to once you have a model, how do you actually make good use of it in the edge and in the clouds?”
That’s where the TVM project comes in, which was launched by Ceze and his collaborators at the University of Washington’s Paul G. Allen School of Computer Science & Engineering. It’s now an Apache incubating project and because it’s seen quite a bit of usage and support from major companies like AWS, ARM, Facebook, Google, Intel, Microsoft, Nvidia, Xilinx and others, the team decided to form a commercial venture around it, which became OctoML. Today, even Amazon Alexa’s wake word detection is powered by TVM.
Ceze described TVM as a modern operating system for machine learning models. “A machine learning model is not code, it doesn’t have instructions, it has numbers that describe its statistical modeling,” he said. “There’s quite a few challenges in making it run efficiently on a given hardware platform because there’s literally billions and billions of ways in which you can map a model to specific hardware targets. Picking the right one that performs well is a significant task that typically requires human intuition.”
And that’s where OctoML and its “Octomizer” SaaS product, which it also announced, today come in. Users can upload their model to the service and it will automatically optimize, benchmark and package it for the hardware you specify and in the format you want. For more advanced users, there’s also the option to add the service’s API to their CI/CD pipelines. These optimized models run significantly faster because they can now fully leverage the hardware they run on, but what many businesses will maybe care about even more is that these more efficient models also cost them less to run in the cloud, or that they are able to use cheaper hardware with less performance to get the same results. For some use cases, TVM already results in 80x performance gains.
Currently, the OctoML team consists of about 20 engineers. With this new funding, the company plans to expand its team. Those hires will mostly be engineers, but Ceze also stressed that he wants to hire an evangelist, which makes sense, given the company’s open-source heritage. He also noted that while the Octomizer is a good start, the real goal here is to build a more fully featured MLOps platform. “OctoML’s mission is to build the world’s best platform that automates MLOps,” he said.
Computer vision techniques used for commercial purposes are turning out to be valuable tools for monitoring people’s behavior during the present pandemic. Zensors, a startup that uses machine learning to track things like restaurant occupancy, lines, and so on, is making its platform available for free to airports and other places desperate to take systematic measures against infection.
The company, founded two years ago but covered by TechCrunch in 2016, was among the early adopters of computer vision as a means to extract value from things like security camera feeds. It may seem obvious now that cameras covering a restaurant can and should count open tables and track that data over time, but a few years ago it wasn’t so easy to come up with or accomplish that.
Since then Zensors has built a suite of tools tailored to specific businesses and spaces, like airports, offices, and retail environments. They can count open and occupied seats, spot trash, estimate lines, and all that kind of thing. Coincidentally, this is exactly the kind of data that managers of these spaces are now very interested in watching closely given the present social distancing measures.
Zensors co-founder Anuraag Jain told Carnegie Mellon University — which the company was spun out of — that it had received a number of inquiries from the likes of airpots regarding applying the technology to public health considerations.
Software that counts how many people are in line can be easily adapted to, for example, estimate how close people are standing and send an alert if too many people are congregating or passing through a small space.
“Rather than profiting off them, we thought we would give our help for free,” said Jain. And so, for the next two months at least, Zensors is providing its platform for free to “selected entities who are on the forefront of responding to this crisis, including our airport clients.”
The system has already been augmented to answer COVID-19-specific questions like whether there are too many people in a given area, when a surface was last cleaned and whether cleaning should be expedited, and how many of a given group are wearing face masks.
Airports surely track some of this information already, but perhaps in a much less structured way. Using a system like this could be helpful for maintaining cleanliness and reducing risk, and no doubt Zensors hopes that having had a taste via what amounts to a free trial, some of these users will become paying clients. Interested parties should get in touch with Zensors via its usual contact page.
Machine learning experts working at Google Health have published a new study in tandem with the University of California San Francisco’s (UCSF) computational health sciences department that describes a machine learning model the researchers built that can anticipate normal physician drug prescribing patterns, using a patient’s electronic health records (EHR) as input. That’s useful because around 2% of patients who end up hospitalized are affected by preventable mistakes in medication prescriptions, some instances of which can even lead to death.
The researchers describe the system as working in a similar manner to automated, machine learning-based fraud detection tools that are commonly used by credit card companies to alert customers of possible fraudulent transactions: They essentially build a baseline of what’s normal consumer behavior based on past transactions, and then alert your bank’s fraud department or freeze access when they detect a behavior that is not in line with an individual’s baseline behavior.
Similarly, the model trained by Google and UCSF worked by identifying any prescriptions that “looked abnormal for the patient and their current situation.” That’s a much more challenging proposition in the case of prescription drugs versus consumer activity — because courses of medication, their interactions with one another and the specific needs, sensitivities and conditions of any given patient all present an incredibly complex web to untangle.
To make it possible, the researchers used electronic health records from de-identified patients that include vital signs, lab results, prior medications and medical procedures, as well as diagnoses and changes over time. They paired this historical data with current state information, and came up with various models to attempt to output an accurate prediction of a course of prescription for a given patient.
Their best-performing model was accurate “three quarters of the time,” Google says, which means that it matched up with what a physician actually decided to prescribe in a large majority of cases. It was also even more accurate (93%) in terms of predicting at least one medication that would fall within a top 10 list of a physician’s most likely medicine choices for a patient — even if its top choice didn’t match the doctor’s.
The researchers are quick to note that though the model thus far has been fairly accurate in predicting a normal course of prescription, that doesn’t mean it’s able to successfully detect deviations from that with any high degree of accuracy. Still, it’s a good first step upon which to build that kind of flagging system.
An undertaking that involved combining massive amounts of graphics processing power could provide key leverage for researchers looking to develop potential cures and treatments for the novel coronavirus behind the current global pandemic. Immunotherapy startup ImmunityBio is working with Microsoft’s Azure to deliver a combined 24 petaflops of GPU computing capability for the purposes of modelling, in a very high degree of detail, the structure o the so-called “spike protein” that allows the SARS-CoV-2 virus that causes COVID-19 to enter human cells.
This new partnership means that they were able to produce a model of the spike protein within just days, instead of the months it would’ve taken previously. That time savings means that the model can get in the virtual hands of researchers and scientists working on potential vaccines and treatments even faster, and that they’ll be able to gear their work towards a detailed replication of the very protein they’re trying to prevent from attaching to the human ACE-2 proteins’ receptor, which is what sets up the viral infection process to begin with.
The main way that scientists working on treatments look to prevent or minimize the spread of the virus within the body is to block the attachment of the virus to these proteins, and the simplest way to do that is to ensure that the spike protein can’t connect with the receptor it targets. Naturally-occurring antibodies in patients who have recovered from the novel coronavirus do exactly that, and the vaccines under development are focused on doing the same thing pre-emptively, while many treatments are looking at lessening the ability of the virus to latch on to new cells as it replicates within the body.
In practical terms, the partnership between the two companies included a complement of 1,250 NVIDIA V100 Tensor Core GPUs designed for use in machine learning applications from a Microsoft Azure cluster, working with ImmunityBio’s existing 320 GPU cluster that is tuned specifically to molecular modeling work. The results of the collaboration will now be made available to researchers working on COVID-19 mitigation and prevention therapies, in the hopes that they will enable them to work more quickly and effectively towards a solution.
In March, the virus gripping the world — COVID-19 — started to spread in Africa. In short order, actors across the continent’s tech ecosystem began to step up to stem the spread.
Early in March, Africa’s COVID-19 cases by country were in the single digits, but by mid-month those numbers had spiked leading the World Health Organization to sound an alarm.
“About 10 days ago we had 5 countries affected, now we’ve got 30,” WHO Regional Director Dr Matshidiso Moeti said at a press conference on March 19. “It has been an extremely rapid…evolution.”
By the World Health Organization’s stats Tuesday there were 3,671 COVID-19 cases in Sub-Saharan Africa and 87 confirmed deaths related to the virus, up from 463 cases and 8 deaths on March 18.
As COVID-19 began to grow in major economies, governments and startups in Africa started measures to shift a greater volume of transactions toward digital payments and away from cash — which the World Health Organization flagged as a conduit for the spread of the coronavirus.
Kenya, Africa’s leader in digital payment adoption, turned to mobile money as a public-health tool.
At the urging of the Central Bank and President Uhuru Kenyatta, the country’s largest telecom, Safaricom, implemented a fee-waiver on East Africa’s leading mobile-money product, M-Pesa, to reduce the physical exchange of currency.
The company announced that all person-to-person (P2P) transactions under 1,000 Kenyan Schillings (≈ $10) would be free for three months.
Kenya has one of the highest rates of digital finance adoption in the world — largely due to the dominance of M-Pesa in the country — with 32 million of its 53 million population subscribed to mobile-money accounts, according to Kenya’s Communications Authority.
On March 20, Ghana’s central bank directed mobile money providers to waive fees on transactions of GH₵100 (≈ $18), with restrictions on transactions to withdraw cash from mobile-wallets.
Ghana’s monetary body also eased KYC requirements on mobile-money, allowing citizens to use existing mobile phone registrations to open accounts with the major digital payment providers, according to a March 18 Bank of Ghana release.
Growth in COVID-19 cases in Nigeria, Africa’s most populous nation of 200 million, prompted one of the country’s largest digital payments startups to act.
Lagos based venture Paga made fee adjustments, allowing merchants to accept payments from Paga customers for free — a measure “aimed to help slow the spread of the coronavirus by reducing cash handling in Nigeria,” according to a company release.
In March, Africa’s largest innovation incubator, CcHub, announced funding and engineering support to tech projects aimed at curbing COVID-19 and its social and economic impact.
The Lagos and Nairobi based organization posted an open application on its website to provide $5,000 to $100,000 funding blocks to companies with COVID-19 related projects.
CcHub’s CEO Bosun Tijani expressed concern for Africa’s ability to combat a coronavirus outbreak. “Quite a number of African countries, if they get to the level of Italy or the UK, I don’t think the system… is resilient enough to provide support to something like that,” Tijani said.
Cape Town based crowdsolving startup Zindi — that uses AI and machine learning to tackle complex problems — opened a challenge to the 12,000 registered engineers on its platform.
The competition, sponsored by AI4D, tasks scientists to create models that can use data to predict the global spread of COVID-19 over the next three months. The challenge is open until April 19, solutions will be evaluated against future numbers and the winner will receive $5,000.
Zindi will also sponsor a hackathon in April to find solutions to coronavirus related problems.
Image Credits: Sam Masikini via Zindi
On the digital retail front, Pan-African e-commerce company Jumia announced measures it would take on its network to curb the spread of COVID-19.
The Nigeria headquartered operation — with online goods and services verticals in 11 African countries — said it would donate certified face masks to health ministries in Kenya, Ivory Coast, Morocco, Nigeria and Uganda, drawing on its supply networks outside Africa.
The company has also offered African governments use of of its last-mile delivery network for distribution of supplies to healthcare facilities and workers.
Jumia is reviewing additional assets it can offer the public sector. “If governments find it helpful we’re willing to do it,” CEO Sacha Poignonnec told TechCrunch.
More Africa-related stories @TechCrunch
African tech around the ‘net
Since its inception, Cape Town based crowdsolving startup Zindi has been building a database of data scientists across Africa.
It now has 12,000 registered on its its platform that uses AI and machine learning to tackle complex problems and will offer them cash-prizes to find solutions to curb COVID-19.
Zindi has an open challenge focused on stemming the spread and havoc of coronavirus and will introduce a hackathon in April. The current competition, sponsored by AI4D, tasks scientists to create models that can use data to predict the global spread of COVID-19 over the next three months.
The challenge is open until April 19, solutions will be evaluated against future numbers and the winner will receive $5000.
The competition fits with Zindi’s business model of building a platform that can aggregate pressing private or public-sector challenges and match the solution seekers to problem solvers.
Founded in 2018, the early-stage venture allows companies, NGOs or government institutions to host online competitions around data oriented issues.
Zindi’s model has gained the attention of some notable corporate names in and outside of Africa. Those who have hosted competitions include Microsoft, IBM and Liquid Telecom. Public sector actors — such as the government of South Africa and UNICEF — have also tapped Zindi for challenges as varied as traffic safety and disruptions in agriculture.
Image Credits: Zindi
The startup’s CEO didn’t imagine a COVID-19 situation precisely, but sees it as one of the reasons she co-founded Zindi with South African Megan Yates and Ghanaian Ekow Duker.
The ability to apply Africa’s data science expertise, to solve problems around a complex health crisis such as COVID-19 is what Zindi was meant for, Lee explained to TechCrunch on a call from Cape Town.
“As an online platform, Zindi is well-positioned to mobilize data scientists at scale, across Africa and around the world, from the safety of their homes,” she said.
Lee explained that perception leads many to believe Africa is the victim or source of epidemics and disease. “We wanted to show Africa can actually also contribute to the solution for the globe.”
With COVID-19, Zindi is being employed to alleviate a problem that is also impacting its founder, staff and the world.
Lee spoke to TechCrunch while sheltering in place in Cape Town, as South Africa went into lockdown Friday due to coronavirus. Zindi’s founder explained she also has in-laws in New York and family in San Francisco living under similar circumstances due to the global spread of COVID-19.
Lee believes the startup’s competitions can produce solutions that nations in Africa could tap as the coronavirus spreads. “The government of Kenya just started a task force where they’re including companies from the ICT sector. So I think there could be interest,” she said.
Starting April, Zindi will launch six weekend hackathons focused on COVID-19.
That could be timely given the trend of COVID-19 in Africa. The continent’s cases by country were in the single digits in early March, but those numbers spiked last week — prompting the World Health Organization’s Regional Director Dr Matshidiso Moeti to sound an alarm on the rapid evolution of the virus on the continent.
By the WHO’s stats Wednesday there were 1691 COVID-19 cases in Sub-Saharan Africa and 29 confirmed deaths related to the virus — up from 463 cases and 10 deaths last Wednesday.
The trajectory of the coronavirus in Africa has prompted countries and startups, such as Zindi, to include the continent’s tech sector as part of a broader response. Central banks and fintech companies in Ghana, Nigeria, and Kenya have employed measures to encourage more mobile-money usage, vs. cash — which the World Health Organization flagged as a conduit for the spread of the virus.
The continent’s largest incubator, CcHub, launched a fund and open call for tech projects aimed at curbing COVID-19 and its social and economic impact.
Pan-African e-commerce company Jumia has offered African governments use of its last-mile delivery network for distribution of supplies to healthcare facilities and workers.
Zindi’s CEO Celina Lee anticipates the startup’s COVID-19 related competitions can provide additional means for policy-makers to combat the spread of the virus.
“The one that’s open right now should hopefully go into informing governments to be able to anticipate the spread of the disease and to more accurately predict the high risk areas in a country,” she said.
Espressive, a four-year-old startup from former ServiceNow employees, is working to build a better chatbot to reduce calls to company help desks. Today, the company announced a $30 million Series B investment.
Insight Partners led the round with help from Series A lead investor General Catalyst along with Wing Venture Capital. Under the terms of today’s agreement, Insight founder and managing director Jeff Horing will be joining the Espressive Board. Today’s investment brings the total raised to $53 million, according to the company.
Company founder and CEO Pat Calhoun says that when he was at ServiceNow he observed that, in many companies, employees often got frustrated looking for answers to basic questions. That resulted in a call to a Help Desk requiring human intervention to answer the question.
He believed that there was a way to automate this with AI-driven chatbots, and he founded Espressive to develop a solution. “Our job is to help employees get immediate answers to their questions or solutions or resolutions to their issues, so that they can get back to work,” he said.
They do that by providing a very narrowly focused natural language processing (NLP) engine to understand the question and find answers quickly, while using machine learning to improve on those answers over time.
“We’re not trying to solve every problem that NLP can address. We’re going after a very specific set of use cases which is really around employee language, and as a result, we’ve really tuned our engine to have the highest accuracy possible in the industry,” Calhoun told TechCrunch.
He says what they’ve done to increase accuracy is combine the NLP with image recognition technology. “What we’ve done is we’ve built our NLP engine on top of some image recognition architecture that’s really designed for a high degree of accuracy and essentially breaks down the phrase to understand the true meaning behind the phrase,” he said.
The solution is designed to provide a single immediate answer. If, for some reason, it can’t understand a request, it will open a help ticket automatically and route it to a human to resolve, but they try to keep that to a minimum. He says that when they deploy their solution, they tune it to the individual customers’ buzzwords and terminology.
So far they have been able to reduce help desk calls by 40% to 60% across customers with around 85% employee participation, which shows that they are using the tool and it’s providing the answers they need. In fact, the product understands 750 million employee phrases out of the box.
The company was founded in 2016. It currently has 65 employees and 35 customers, but with the new funding, both of those numbers should increase.
As the world becomes more deeply connected through IoT devices and networks, consumer and business needs and expectations will soon only be sustainable through automation.
Recognizing this, artificial intelligence and machine learning are being rapidly adopted by critical industries such as finance, retail, healthcare, transportation and manufacturing to help them compete in an always-on and on-demand global culture. However, even as AI and ML provide endless benefits — such as increasing productivity while decreasing costs, reducing waste, improving efficiency and fostering innovation in outdated business models — there is tremendous potential for errors that result in unintended, biased results and, worse, abuse by bad actors.
The market for advanced technologies including AI and ML will continue its exponential growth, with market research firm IDC projecting that spending on AI systems will reach $98 billion in 2023, more than two and one-half times the $37.5 billion that was projected to be spent in 2019. Additionally, IDC foresees that retail and banking will drive much of this spending, as the industries invested more than $5 billion in 2019.
These findings underscore the importance for companies that are leveraging or plan to deploy advanced technologies for business operations to understand how and why it’s making certain decisions. Moreover, having a fundamental understanding of how AI and ML operate is even more crucial for conducting proper oversight in order to minimize the risk of undesired results.
Companies often realize AI and ML performance issues after the damage has been done, which in some cases has made headlines. Such instances of AI driving unintentional bias include the Apple Card allowing lower credit limits for women and Google’s AI algorithm for monitoring hate speech on social media being racially biased against African Americans. And there have been far worse examples of AI and ML being used to spread misinformation online through deepfakes, bots and more.
Through real-time monitoring, companies will be given visibility into the “black box” to see exactly how their AI and ML models operate. In other words, explainability will enable data scientists and engineers to know what to look for (a.k.a. transparency) so they can make the right decisions (a.k.a. insight) to improve their models and reduce potential risks (a.k.a. building trust).
But there are complex operational challenges that must first be addressed in order to achieve risk-free and reliable, or trustworthy, outcomes.
The tech industry is mobilizing its considerable resources to attempt to support efforts against the growing global coronavirus pandemic. Over the weekend, the CEOs of Amazon, Apple and Microsoft all shared updates regarding some aspects of their company’s ongoing contributions, which range from donations of medical supplies and personal protective equipment (PPE) for frontline healthcare workers, to software projects that help track and analyze the global spread of infection.
Apple CEO Tim Cook shared on Twitter that the company has been attempting to source necessary supplies that are needed for healthcare workers both in the U.S. and Europe, and that the company is joining “millions of masks” for this use. Apple also detailed some of its other updates via earlier releases, including a $15 million donation, along with two-to-one corporate matching for all employee donations that go towards COVID-19 response.
Amazon founder and CEO Jeff Bezos provided an update on Saturday on the company’s official blog that included details about the change in Amazon’s prioritization for its warehousing and logistics operations, which now focus on essential items including daily household staples, baby and medical supplies. Bezos also reiterated Amazon’s commitment to hiring 100,000 new roles, along with raising hourly wages for fulfilment workers.
Bezos notes that while the company has “placed purchase orders for millions of face masks” that it intends to distribute to its full-time and contract workers who are not able to work from home, “very few of those orders have been filled” to to the global supply shortage. He further notes that these resources are likely to go to frontline healthcare workers first, and that the company will focus on getting them to their staff in order of priority once they become available.
Microsoft CEO Satya Nadella provided a lengthy update about his company’s various efforts in a LinkedIn post on Saturday, publishing an email he sent to all Microsoft employees for external consumption. Nadella describes some of its telehealth platform software work, as well as a number of collaborative data projects, including the John Hopkins University global COVID-19 confirmed case tracker. The Centers for Disease Control and Prevention (CDC) also released a chatbot assessment tool for COVID-19 that uses Microsoft’s health chatbot tech as its underlying framework.
Microsoft is also seeing Teams and Minecraft being used globally for remote learning iniativies designed to supplement in-perosn school closures, and it’s working on machine learning and big data projects to support global research efforts. Earlier this week, Microsoft’s Chief Scientific Officer Eric Horvitz announced that it would be providing an open research data set in partnership with colleagues at academic institutions around the world, as well as the White House Office of Science and Technology Policy and the Chan Zuckerberg initiative. The data set, called the COVID-19 Open Research Data Set, includes more than 29,000 scholarly articles about the virus, and will grow as more are published.
When Ada Health was founded nine years ago, hardly anyone was talking about combining artificial intelligence and physician care — outside of a handful of futurists.
But the chatbot boom gave way to a powerful combination of AI-augmented health care which others, like Babylon Health in 2013 and KRY in 2015, also capitalized on. The journey Ada was about to take was not an obvious one, so I spoke to Dr. Claire Novorol, Ada’s co-founder and chief medical officer, at the Slush conference last year to unpack their process and strategy.
Co-founded with Daniel Nathrath and Dr. Martin Hirsch, the startup initially set out to be an assistant to doctors rather than something that would have a consumer interface. At the beginning, Novorol said they did not talk about what they were building as an AI so much as it was pure machine learning.
Years later, Ada is a free app, and just like the average chatbot, it asks a series of questions and employs an algorithm to make an initial health assessment. It then proposes next steps, such as making an appointment with a doctor or going to an emergency room. But Ada’s business model is not to supplant doctors but to create partnerships with healthcare providers and encourage patients to use it as an early screening system.
It was Novorol who convinced the company to pivot from creating tools for doctors into a patient-facing app that could save physicians time by providing patients with an initial diagnosis. Since the app launched in 2016, Ada has gone on to raise $69.3 million. In contrast, Babylon Health has raised $635.3 million, while KRY has raised $243.6 million. Ada claims to be the top medical app in 130 countries to date and has completed more than 15 million assessments to date.
One of the more interesting and useful applications of artificial intelligence technology has been in the world of biotechnology and medicine, where now more than 220 startups (not to mention universities and bigger pharma companies) are using AI to accelerate drug discovery by using it to play out the many permutations resulting from drug and chemical combinations, DNA and other factors.
Now, a startup called Turing — which is part of the current cohort at Y Combinator due to present in the next Demo Day on March 22 — is taking a similar principle but applying it to the world of building (and “discovering”) new consumer packaged goods products.
Using machine learning to simulate different combinations of ingredients plus desired outcomes to figure out optimal formulations for different goods (hence the “Turing” name, a reference to Alan Turing’s mathematical model, referred to as the Turing machine), Turing is initially addressing the creation of products in home care (e.g. detergents), beauty and food and beverage.
Turing’s founders claim that it is able to save companies millions of dollars by reducing the average time it takes to formulate and test new products, from an average of 12 to 24 months down to a matter of weeks.
Specifically, the aim is to reduce all the time it takes to test combinations, giving R&D teams more time to be creative.
“Right now, they are spending more time managing experiments than they are innovating,” Manmit Shrimali, Turing’s co-founder and CEO, said.
Turing is in theory coming out of stealth today, but in fact it has already amassed an impressive customer list. It is already generating revenues by working with eight brands owned by one of the world’s biggest CPG companies, and it is also being trialed by another major CPG behemoth (Turing is not disclosing their names publicly, but suffice it to say, they and their brands are household names).
Turing is co-founded by Shrimali and Ajith Govind, two specialists in data science that had worked together on a previous startup called Dextro Analytics. Dextro had set out to help businesses use AI and other kinds of business analytics to help with identifying trends and decision making around marketing, business strategy and other operational areas.
While there, they identified a very specific use case for the same principles that was perhaps even more acute: the research and development divisions of CPG companies, which have (ironically, given their focus on the future) often been behind the curve when it comes to the “digital transformation” that has swept up a lot of other corporate departments.
“We were consulting for product companies and realised that they were struggling,” Shirmali said. Add to that the fact that CPG is precisely the kind of legacy industry that is not natively a tech company but can most definitely benefit from implementing better technology, and that spells out an interesting opportunity for how (and where) to introduce artificial intelligence into the mix.
R&D labs play a specific and critical role in the world of CPG.
Before eventually being shipped into production, this is where products are discovered; tested; tweaked in response to input from customers, marketing, budgetary and manufacturing departments and others; then tested again; then tweaked again; and so on. One of the big clients that Turing works with spends close to $400 million in testing alone.
But R&D is under a lot of pressure these days. While these departments are seeing their budgets getting cut, they continue to have a lot of demands. They are still being expected to meet timelines in producing new products (or often more likely, extensions of products) to keep consumers interested. There are a new host of environmental and health concerns around goods with huge lists of unintelligible ingredients, meaning they have to figure out how to simplify and improve the composition of mass-market products. And smaller direct-to-consumer brands are undercutting their larger competitors by getting to market faster with competitive offerings that have met new consumer tastes and preferences.
“In the CPG world, everyone was focused on marketing, and R&D was a blind spot,” Shrimali said, referring to the extensive investments that CPG have made into figuring out how to use digital to track and connect with users, and also how better to distribute their products. “To address how to use technology better in R&D, people need strong domain knowledge, and we are the first in the market to do that.”
Turing’s focus is to speed up the formulation and testing aspects that go into product creation to cut down on some of the extensive overhead that goes into putting new products into the market.
Part of the reason why it can take upwards of years to create a new product is because of all of the permutations that go into building something and making sure it works consistently as a consumer would expect it to (which still being consistent in production and coming in within budget).
“If just one ingredient is changed in a formulation, it can change everything,” Shirmali noted. And so in the case of something like a laundry detergent, this means running hundreds of tests on hundreds of loads of laundry to make sure that it works as it should.
The Turing platform brings in historical data from across a number of past permutations and tests to essentially virtualise all of this: it suggests optimal mixes and outcomes from them without the need to run the costly physical tests, and in turn this teaches the Turing platform to address future tests and formulations. Shrimali said that the Turing platform has already saved one of the brands some $7 million in testing costs.
Turing’s place in working with R&D gives the company some interesting insights into some of the shifts that the wider industry is undergoing. Currently, Shrimali said one of the biggest priorities for CPG giants include addressing the demand for more traceable, natural and organic formulations.
While no single DTC brand will ever fully eat into the market share of any CPG brand, collectively their presence and resonance with consumers is clearly causing a shift. Sometimes that will lead into acquisitions of the smaller brands, but more generally it reflects a change in consumer demands that the CPG companies are trying to meet.
Longer term, the plan is for Turing to apply its platform to other aspects that are touched by R&D beyond the formulations of products. The thinking is that changing consumer preferences will also lead into a demand for better “formulations” for the wider product, including more sustainable production and packaging. And that, in turn, represents two areas into which Turing can expand, introducing potentially other kinds of AI technology (such as computer vision) into the mix to help optimise how companies build their next generation of consumer goods.
I see far more research articles than I could possibly write up. This column collects the most interesting of those papers and advances along with notes on why they may prove important in the world of tech and startups.
This week: advances in rocketry, machine learning, wireless transmission and more.
In some ways, rocketry is not so different from its beginnings around WWII, but as other bottlenecks give way it is becoming feasible to experiment with truly innovative types of rocket engines. One such type is the rotating detonation engine, an alternative to the standard means of controlling and directing the combustion that creates thrust. The process is highly chaotic, however, and not well understood enough to control properly.
University of Washington researchers set up a test rotating detonation engine and studied the combustion patterns inside using an ultra-high-speed camera. The footage was analyzed to produce the first mathematical model simulating the process. It’s still at a very early stage but understanding the mechanism of a new technology like this is necessary before putting it into practice. When packaged in software, this type of simulation can also be licensed to aerospace firms. You can read the full paper here.
Cartesiam, a startup that aims to bring machine learning to edge devices powered by microcontrollers, has launched a new tool for developers who want an easier way to build services for these devices. The new NanoEdge AI Studio is the first IDE specifically designed for enabling machine learning and inferencing on Arm Cortex-M microcontrollers, which power billions of devices already.
As Cartesiam GM Marc Dupaquier, who co-founded the company in 2016, told me, the company works very closely with Arm, given that both have a vested interest in having developers create new features for these devices. He noted that while the first wave of IoT was all about sending data to the cloud, that has now shifted and most companies now want to limit the amount of data they send out and do a lot more on the device itself. And that’s pretty much one of the founding theses of Cartesiam. “It’s just absurd to send all this data — which, by the way, also exposes the device from a security standpoint,” he said. “What if we could do it much closer to the device itself?”
The company first bet on Intel’s short-lived Curie SoC platform. That obviously didn’t work out all that well, given that Intel axed support for Curie in 2017. Since then, Cartesiam has focused on the Cortex-M platform, which worked out for the better, given how ubiquitous it has become. Since we’re talking about low-powered microcontrollers, though, it’s worth noting that we’re not talking about face recognition or natural language understanding here. Instead, using machine learning on these devices is more about making objects a little bit smarter and, especially in an industrial use case, detecting abnormalities or figuring out when it’s time to do preventive maintenance.
Today, Cartesiam already works with many large corporations that build Cortex-M-based devices. The NanoEdge Studio makes this development work far easier, though. “Developing a smart object must be simple, rapid and affordable — and today, it is not, so we are trying to change it,” said Dupaquier. But the company isn’t trying to pitch its product to data scientists, he stressed. “Our target is not the data scientists. We are actually not smart enough for that. But we are unbelievably smart for the embedded designer. We will resolve 99% of their problems.” He argues that Cartesiam reduced time to market by a factor of 20 to 50, “because you can get your solution running in days, not in multiple years.”
One nifty feature of the NanoEdge Studio is that it automatically tries to find the best algorithm for a given combination of sensors and use cases and the libraries it generates are extremely small and use somewhere between 4K to 16K of RAM.
NanoEdge Studio for both Windows and Linux is now generally available. Pricing starts at €690/month for a single user or €2,490/month for teams.
Alkymi, an early stage startup that wants to bring intelligence to highly manual business processes like copying and pasting financial data from emails and attachments, launched today with a $5 million seed investment.
Canaan Partners led the round with participation from previous investor Work-Bench. SimCorp also contributed as a strategic investor. Under the terms of the investment agreement, Joydeep Bhattacharyya from Canaan will become a member of the Alkymi board.
Company founder and CEO Harald Collet says the startup is bringing machine learning to the business analyst’s in-box with the goal of automating many of the tedious manual parts of the job. The company has created a solution that extracts data automatically that these analysts previously had to copy and paste into applications, spreadsheets or databases.
“What we do is we focus on automating tasks in emails and documents and really focusing on helping business analysts in [automating] those tasks where they have been taking and picking out of business data customer and financial data that’s being fed into business processes,” Collet told TechCrunch.
For today, that strictly involves financial services, which is an industry Collet has worked in for two decades, and which could benefit greatly from this approach. He uses an investment asset manager as an example. This person would receive emails with data in them about investments, copy and paste the data into an application or database, and repeat this many times to report on overall investment performance. Alkymi would automatically extract some amount of this data, reducing the overall manual copying and pasting required.
It takes some time to train the underlying machine model, from hours to days, depending on the size and complexity of the operation, but once that’s done, Collet says the software can deal with what it knows, setting aside what it can’t figure out for a human to intervene, and then learn from that in a typical machine learning loop. Over time, it should allow business analysts to do more analysis, instead of spending time on data entry to get to the analysis part. For now, they are looking at rates starting at 40-50% automation, or more for less complex data sets.
While the company is concentrating on financial services today, the long-term plan is to expand into other verticals over time. For now, it is growing quickly with paying financial services customers. It has also partnered with investor SimCorp, which will offer the service on its platform aimed at financial services professionals.
The company launched in 2017, and Collet spent time talking to potential customers before building the product. It offers an on-prem and cloud version, and bills by the workflow. Today, it has 7 employees based in New York City with plans to double that this year.
Artificial intelligence is one of the most important fields in technology right now, which makes it ripe for buzzword-savvy startups to leverage for attention. But while machine learning and related technologies are now frequently employed, it’s less common that it’s central to a company’s strategy and IP.
It’s important to note that this sort of posturing doesn’t necessarily mean a company is bad — it’s entirely possible they have an overzealous communications department or PR firm. Just consider the following points warning signs — if you hear these terms, dig a little deeper to find out exactly what the company does.
There are innumerable variations on this particular line, which is a red flag that the company is trying to paint itself with the AI brush rather than differentiate by other means.
“Our machine-learning powered ___,” “our proprietary AI,” “leverages machine learning…” all basically mean the same thing: AI is involved somewhere along the line.
Apps that purport to connect users (“our unique AI-powered matching engine…”) with the right people or resources based on AI recommendations are also a common offender
But machine learning algorithms have been deeply embedded in computing for many years. They can be simple or complex, tried and true or novel and used for highly visible or completely unknown purposes. There are off-the-shelf algorithms developers can buy to help sort images, parse noisy data and perform many other tasks. Recommendation engines are a dime a dozen. Does using one of these make a product “powered by AI”?
Oxx, a European venture capital firm co-founded by Richard Anton and Mikael Johnsson, this month announced the closing of its debut fund of $133 million to back “Europe’s most promising SaaS companies” at Series A and beyond.
Launched in 2017 and headquartered in London and Stockholm, Oxx pitches itself as one of only a few European funds focused solely on SaaS, and says it will invest broadly across software applications and infrastructure, highlighting five key themes: “data convergence & refinery,” “future of work,” “financial services infrastructure,” “user empowerment” and “sustainable business.”
However, its standout USP is that the firm says it wants to be a more patient form of capital than investors who have a rigid Silicon Valley SaaS mindset, which, it says, often places growth ahead of building long-lasting businesses.
I caught up with Oxx’s co-founders to dig deeper into their thinking, both with regards to the firm’s remit and investment thesis, and to learn more about the pair’s criticism of the prevailing venture capital model they say often pushes SaaS companies to prioritize “grow at all costs.”
TechCrunch: Oxx is described as a B2B software investor investing in SaaS companies across Europe from Series A and beyond. Can you be more specific regarding the size of check you write and the types of companies, geographies, technologies and business models you are focusing on?
Richard Anton: We will lead funding rounds anywhere in the range $5-20 million in SaaS companies. Some themes we’re especially excited about include data convergence and the refining and usage of data (think applications of machine learning, for example), the future of work, financial services infrastructure, end-user empowerment and sustainable business.
TechCrunch Sessions: Robotics + AI brings together a wide group of the ecosystem’s leading minds on March 3 at UC Berkeley. Over 1000+ attendees are expected from all facets of the robotics and artificial intelligence space – investors, students, engineerings, C-levels, technologists, and researchers. We’ve compiled a small list of highlights of attendees’ companies and job titles attending this year’s event below.
STUDENTS & RESEARCHERS FROM:
Did you know that TechCrunch provides a white-glove networking app at all our events called CrunchMatch? You can connect and match with people who meet your specific requirements, message them, and connect right at the conference. How cool is that!?
Want to get in on networking with this caliber of people? Book your $345 General Admission ticket today and save $50 before prices go up at the door. But no one likes going to events alone. Why not bring the whole team? Groups of four or more save 15% on tickets when you book here.
We’ve been dropping into the Australian startup scene increasingly over the years as the ecosystem has been building at an increasingly faster pace, most notably at our own TechCrunch Battlefield Australia in 2017. Further evidence that the scene is growing has come recently in the shape of the Pause Fest conference in Melbourne. This event has gone from strength to strength in recent years and is fast becoming a must-attend for Aussie startups aiming for both national international attention.
I was able to drop in ‘virtually’ to interview a number of those showcased in the Startup Pitch Competition, so here’s a run-down of some of the stand-out companies.
Medinet Australia is a health tech startup aiming to make healthcare more convenient and accessible to Australians by allowing doctors to do consultations with patients via an app. Somewhat similar to apps like Babylon Health, Medinet’s telehealth app allows patients to obtain clinical advice from a GP remotely; access prescriptions and have medications delivered; access pathology results; directly email their medical certificate to their employer; and access specialist referrals along with upfront information about specialists such as their fees, waitlist, and patient experience. They’ve raised $3M in Angel financing and are looking for institutional funding in due course. Given Australia’s vast distances, Medinet is well-placed to capitalize on the shift of the population towards much more convenient telehealth apps. (1st Place Winner)
Everty allows companies to easily manage, monitor and monetize Electric Vehicle charging stations. But this isn’t about infrastructure. Instead, they link up workplaces and accounting systems to the EV charging network, thus making it more like a “Salesforce for EV charging”. It’s available for both commercial and home charging tracking. It’s also raised an Angel round and is poised to raise further funding. (2nd Place Winner)
AI On Spectrum
It’s a sad fact that people with Autism statistically tend to die younger, and unfortunately, the suicide rate is much higher for Autistic people. “Ai on Spectrum” takes an accessible approach in helping autistic kids and their families find supportive environments and feel empowered. The game encourages Autism sufferers to explore their emotional side and arms them with coping strategies when times get tough, applying AI and machine learning in the process to assist the user. (3rd Place Winner)
Professional bee-keepers need a fast, reliable, easy-to-use record keeper for their bees and this startup does just that. But it’s also developing a software+sensor technology to give beekeepers more accurate analytics, allowing them to get an early-warning about issues and problems. Their technology could even, in the future, be used to alert for coming bushfires by sensing the changed behavior of the bees. (Hacker Exchange Additional Winner)
Rechargeable batteries for things like cars can be re-used again, but the key to employing them is being able to extend their lives. Relectrify says its battery control software can unlock the full performance from every cell, increasing battery cycle life. It will also reduce storage costs by providing AC output without needing a battery inverter for both new and 2nd-life batteries. Its advanced battery management system combines power and electric monitoring to rapidly the check which are stronger cells and which are weaker making it possible to get as much as 30% more battery life, as well as deploying “2nd life storage”. So far, they have a project with Nissan and American Electric Power and have raised a Series A of $4.5M. (SingularityU Additional Winner)
Sadly, seniors and patients can contract bedsores if left too long. People can even die from bedsores. Furthermore, hospitals can end up in litigation over the issue. What’s needed is a technology that can prevent this, as well as predicting where on a patient’s body might be worst affected. That’s what Gabriel has come up with: using multi-modal technology to prevent and detect both falls and bedsores. Its passive monitoring technology is for the home or use in hospitals and consists of a resistive sheet with sensors connecting to a system which can understand the pressure on a bed. It has FDA approval, is patent-pending and is already working in some Hawaiin hospitals. It’s so far raised $2m in Angel and is now raising money.
Here’s a taste of Pause Fest:
Six months ago or thereabouts, a group of engineers and developers with backgrounds from the National Security Agency, Google and Amazon Web Services had an idea.
Data is valuable for helping developers and engineers to build new features and better innovate. But that data is often highly sensitive and out of reach, kept under lock and key by red tape and compliance, which can take weeks to get approval. So, the engineers started Gretel, an early-stage startup that aims to help developers safely share and collaborate with sensitive data in real time.
It’s not as niche of a problem as you might think, said Alex Watson, one of the co-founders. Developers can face this problem at any company, he said. Often, developers don’t need full access to a bank of user data — they just need a portion or a sample to work with. In many cases, developers could suffice with data that looks like real user data.
“It starts with making data safe to share,” Watson said. “There’s all these really cool use cases that people have been able to do with data.” He said companies like GitHub, a widely used source code sharing platform, helped to make source code accessible and collaboration easy. “But there’s no GitHub equivalent for data,” he said.
And that’s how Watson and his co-founders, John Myers, Ali Golshan and Laszlo Bock came up with Gretel.
“We’re building right now software that enables developers to automatically check out an anonymized version of the data set,” said Watson. This so-called “synthetic data” is essentially artificial data that looks and works just like regular sensitive user data. Gretel uses machine learning to categorize the data — like names, addresses and other customer identifiers — and classify as many labels to the data as possible. Once that data is labeled, it can be applied access policies. Then, the platform applies differential privacy — a technique used to anonymize vast amounts of data — so that it’s no longer tied to customer information. “It’s an entirely fake data set that was generated by machine learning,” said Watson.
It’s a pitch that’s already gathering attention. The startup has raised $3.5 million in seed funding to get the platform off the ground, led by Greylock Partners, and with participation from Moonshots Capital, Village Global and several angel investors.
“At Google, we had to build our own tools to enable our developers to safely access data, because the tools that we needed didn’t exist,” said Sridhar Ramaswamy, a former Google executive, and now a partner at Greylock.
Gretel said it will charge customers based on consumption — a similar structure to how Amazon prices access to its cloud computing services.
“Right now, it’s very heads-down and building,” said Watson. The startup plans to ramp up its engagement with the developer community in the coming weeks, with an eye on making Gretel available in the next six months, he said.
Yellow, the accelerator program launched by Snap in 2018, has selected ten companies to join its latest cohort.
The new batch of startups coming from across the U.S. and international cities like London, Mexico City, Seoul and Vilnius are building professional social networks for black professionals and blue collar workers, fashion labels, educational tools in augmented reality, kids entertainment, and an interactive entertainment production company.
The list of new companies include:
The latest cohort from Snap’s Yellow accelerator
Since launching the platform in 2018, startups from the Snap accelerator have gone on to acquisition (like Stop, Breathe, and Think, which was bought by Meredith Corp.) and to raise bigger rounds of funding (like the voiceover video production toolkit, MuzeTV, and the animation studio Toonstar).
Every company in the Yellow portfolio will receive $150,000 mentorship from industry veterans in and out of Snap, creative office space in Los Angeles and commercial support and partnerships — including Snapchat distribution.