Google Launches a Dermatology AI App in EU

Billions of times each year, people turn to Google’s web search box for help figuring out what’s wrong with their skin. Now, Google is preparing to launch an app that uses image recognition algorithms to provide more expert and personalized help. A brief demo at the company’s developer conference last month showed the service suggesting several possible skin conditions based on uploaded photos.

Machines have matched or outperformed expert dermatologists in studies in which algorithms and doctors scrutinize images from past patients. But there’s little evidence from clinical trials deploying such technology, and no AI image analysis tools are approved for dermatologists to use in the US, says Roxana Daneshjou, a Stanford dermatologist and researcher in machine learning and health.

Many don’t pan out in the real world setting,” she says.

Google’s new app isn’t clinically validated yet either, but the company’s AI prowess and recent buildup of its health care division make its AI dermatology app notable. Still, the skin service will start small—and far from its home turf and largest market in the US. The service is not likely to analyze American skin blemishes any time soon.

At the developer conference, Google’s chief health officer, Karen DeSalvo, said the company aims to launch what it calls a dermatology assist tool in the European Union as soon as the end of this year. A video of the app suggesting that a mark on someone’s arm could be a mole featured a caption saying it was an approved medical device in the EU. The same note added a caveat: “Not available in the US.”

Google says its skin app has been approved “CE marked as a Class I medical device in the EU,” meaning it can be sold in the bloc and other countries recognizing that standard. The company would have faced relatively few hurdles to secure that clearance, says Hugh Harvey, managing director at Hardian Health, a digital health consultancy in the UK. “You essentially fill in a form and self-certify,” he says. Google’s conference last month took place a week before tighter EU rules took effect that Harvey says require many health apps, likely including Google’s, to show that an app is effective, among other things. Preexisting apps have until 2025 to comply with the new rules.


AI Makes Gigantic Leap And Heralds A Revolution In Biology

An artificial intelligence (AI) network developed by Google AI offshoot DeepMind has made a gargantuan leap in solving one of biology’s grandest challengesdetermining a protein’s 3D shape from its amino-acid sequence.

DeepMind’s program, called AlphaFold, outperformed around 100 other teams in a biennial protein-structure prediction challenge called CASP, short for Critical Assessment of Structure Prediction. The results were announced on 30 November, at the start of the conference — held virtually this year — that takes stock of the exercise.

A protein’s function is determined by its 3D shape

This is a big deal,” says John Moult, a computational biologist at the University of Maryland in College Park, who co-founded CASP in 1994 to improve computational methods for accurately predicting protein structures. “In some sense the problem is solved.

The ability to accurately predict protein structures from their amino-acid sequence would be a huge boon to life sciences and medicine. It would vastly accelerate efforts to understand the building blocks of cells and enable quicker and more advanced drug discovery.

AlphaFold came top of the table at the last CASP — in 2018, the first year that London-based DeepMind participated. But, this year, the outfit’s deep-learning network was head-and-shoulders above other teams and, say scientists, performed so mind-bogglingly well that it could herald a revolution in biology.

It’s a game changer,” says Andrei Lupas, an evolutionary biologist at the Max Planck Institute for Developmental Biology in Tübingen, Germany, who assessed the performance of different teams in CASP. AlphaFold has already helped him find the structure of a protein that has vexed his lab for a decade, and he expects it will alter how he works and the questions he tackles. “This will change medicine. It will change research. It will change bioengineering. It will change everything,” Lupas adds.


New Powerful Quantum Computer

Honeywell, a company best known for making control systems for homes, businesses and planes, claims to have built the most powerful quantum computer ever. Other researchers are sceptical about its power, but for the company, it is a step towards integrating quantum computing into its everyday operationsHoneywell measured its computer’s capabilities using a metric invented by IBM called quantum volume. It takes into account the number of quantum bits – or qubits – the computer has, their error rate, how long the system can spend calculating before the qubits stop working and a few other key properties.

Measuring quantum volume involves running about 220 different algorithms on the computer”, says Tony Uttley, the president of Honeywell Quantum Solutions. Honeywell’s quantum computer has a volume of 64, twice as high as the next highest quantum volume to be recorded, which was measured in an IBM quantum computer.

Like other quantum computers, this one may eventually be useful for calculations that deal with huge amounts of data. “There are three classes of problems that we are focused on right now: optimization, machine learning, and chemistry and material science,” says Uttley. “We can do those problems shrunk down to a size that fits our quantum computer today and then, as we increase the quantum volume, we’ll be able to do those problems on bigger scales.” However, this quantum computer isn’t yet able to perform calculations that would give a classical computer trouble, a feat called quantum supremacy, which was first claimed by Google in October. “While it’s cool that the company that made my thermostat is now building quantum computers, claiming it’s the most powerful one isn’t really substantiated,” says Ciarán Gilligan-Lee at University College London.

“Google’s Sycamore quantum computer used 53 qubits to achieve quantum supremacy, while Honeywell’s machine only has six qubits so far. “We know that anything less than around 50 or 60 qubits can be simulated on a classical computer relatively easily,” says Gilligan-Lee. “A six-qubit quantum computer can probably be simulated by your laptop, and a supercomputer could definitely do it.” Having the highest quantum volume may mean that Honeywell’s qubits are remarkably accurate and can calculate for a long time, but it doesn’t necessarily make it the most powerful quantum computer out there, he says.

Scott Aaronson at the University of Texas at Austin  agrees. “Quantum volume is not the worst measure, but what I personally care about, much more than that or any other invented measure, is what you can actually do with the device that’s hard for a classical computer to simulate,” he says. “By the latter measure, the Honeywell device is not even close to the best out there.”


Toyota To Build A Smart City Powered By Hydrogen


Japanese carmaker Toyota has announced plans to create a 175-acre smart city in Japan where it will test driverless cars and artificial intelligence. The project, announced at the Consumer Electronics Show in Las Vegas, will break ground at the base of Mount Fuji in 2021. Woven City will initially be home to 2,000 people who will test technologies including robots and smart homesToyota said in a press release that only driverless and electric vehicles will be allowed on the main streets of Woven CityStreets will be split into three types of thoroughfare: roads for fast vehicles, lanes which are a mixture of personal vehicles and pedestrians, and pedestrian footpaths.

Danish architect Bjarke Ingels has been commissioned to design the new city. His business previously worked on projects including Google’s London and US headquartersToyota said the city will be powered by hydrogen fuel cells and solar panels fitted to the roofs of housesBuildings in Woven City will mostly be made of wood and assembled using “robotised production methods,” Toyota said. 

 “Building a complete city from the ground up, even on a small scale like this, is a unique opportunity to develop future technologies, including a digital operating system for the infrastructure.
“With people, buildings and vehicles all connected and communicating with each other through data and sensors, we will be able to test connected AI technology, in both the virtual and physical realms, maximising its potential,” said Akio Toyoda, Toyota’s president.

Google has also experimented with the creation of its own smart city through its Sidewalk Labs division. The company is hoping to transform a 12-acre plot in Toronto’s waterfront district into a smart city, with the first homes due to appear in 2023.


Artificial Intelligence Outperforms Humans In Prediction Of Breast Cancer

An artificial intelligence (AI) system can reduce false positives and false negatives in prediction of breast cancer and outperforms human readers, according to a study published online Jan. 1 in Nature.

Scott Mayer McKinney, from Google Health in Palo Alto, California, and colleagues examined the performance of an AI system for breast cancer prediction in a clinical setting. Data were curated from a large representative dataset from the United Kingdom and a large enriched dataset from the United States.

The researchers observed an absolute reduction of 5.7 and 1.2 percent in false positives in the U.S. and U.K. datasets, respectively, and 9.4 and 2.7 percent, respectively, in false negatives. The system was also able to generalize from the United Kingdom to the United States. The AI system outperformed six human readers in an independent study involving radiologists; the area under the receiver operating characteristic curve was greater for the AI system than the average radiologist (absolute margin, 11.5 percent). The AI system maintained noninferior performance in a simulation in which the AI system participated in the double-reading process that is used in the United Kingdom and reduced the workload of the second reader by 88 percent.

“These analyses highlight the potential of this technology to deliver screening results in a sustainable manner despite workforce shortages in countries such as the United Kingdom,” the authors write.

Several authors disclosed financial ties to technology companies, including Google, which funded the study.


Quantum Supremacy

Researchers in UC Santa Barbara/Google scientist John Martinis’ group have made good on their claim to quantum supremacy. Using 53 entangled quantum bits (“qubits”), their Sycamore computer has taken on — and solved — a problem considered intractable for classical computers.

Google’s quantum supreme cryostat with Sycamore inside

A computation that would take 10,000 years on a classical supercomputer took 200 seconds on our quantum computer,” said Brooks Foxen, a graduate student researcher in the Martinis Group. “It is likely that the classical simulation time, currently estimated at 10,000 years, will be reduced by improved classical hardware and algorithms, but, since we are currently 1.5 billion times faster, we feel comfortable laying claim to this achievement.

The feat is outlined in a paper in the journal Nature.

The milestone comes after roughly two decades of quantum computing research conducted by Martinis and his group, from the development of a single superconducting qubit to systems including architectures of 72 and, with Sycamore, 54 qubits (one didn’t perform) that take advantage of the both awe-inspiring and bizarre properties of quantum mechanics.

The algorithm was chosen to emphasize the strengths of the quantum computer by leveraging the natural dynamics of the device,” said Ben Chiaro, another graduate student researcher in the Martinis Group. That is, the researchers wanted to test the computer’s ability to hold and rapidly manipulate a vast amount of complex, unstructured data.

We basically wanted to produce an entangled state involving all of our qubits as quickly as we can,” Foxen said, “and so we settled on a sequence of operations that produced a complicated superposition state that, when measured, returned output (“bitstring”) with a probability determined by the specific sequence of operations used to prepare that particular superposition.” The exercise, which was to verify that the circuit’s output correspond to the sequence used to prepare the state, sampled the quantum circuit a million times in just a few minutes, exploring all possibilities — before the system could lose its quantum coherence. “We performed a fixed set of operations that entangles 53 qubits into a complex superposition state,” Chiaro explained. “This superposition state encodes the probability distribution. For the quantum computer, preparing this superposition state is accomplished by applying a sequence of tens of control pulses to each qubit in a matter of microseconds. We can prepare and then sample from this distribution by measuring the qubits a million times in 200 seconds.” “For classical computers, it is much more difficult to compute the outcome of these operations because it requires computing the probability of being in any one of the 2^53 possible states, where the 53 comes from the number of qubits — the exponential scaling is why people are interested in quantum computing to begin with,” Foxen said. “This is done by matrix multiplication, which is expensive for classical computers as the matrices become large.”

According to the new paper, the researchers used a method called cross-entropy benchmarking to compare the quantum circuit’s bitstring to its “corresponding ideal probability computed via simulation on a classical computer” to ascertain that the quantum computer was working correctly. “We made a lot of design choices in the development of our processor that are really advantageous,” said Chiaro. Among these advantages, he said, are the ability to experimentally tune the parameters of the individual qubits as well as their interactions.


Driverless Taxi Service in US and France By The End Of The Year

The City of Lyon in France, will operate a regular cab service by the end of this year, using driverless electric vehicles from the french company Navya. As a pioneer and specialist in the autonomous vehicle market, Navya has conceived, developed and produced the Autonom Cab, the very first autonomous, personalized and shared mobility solution. The cab was designed from the outset to be autonomous, just like all the vehicles in the Autonom range, meaning that there is no cockpit, steering wheel nor pedals.


At the heart of the smart city, Autonom Cab provides an intelligent transport service for individual trips in urban centers. Able to carry 1 to 6 passengers, The driverless taxi is a fluid, continuous and effective solution that answers user expectations in terms of service before, during and after their trip. Available as either a private or shared service, Autonom Cab places an emphasis on conviviality and comfort. On board, passengers can for example choose to work, benefiting from fully connected technology, or partake in an interactive cultural visit of the city. They can also choose a playlist, or buy their cinema or museum tickets.

As well,  the american company Waymo says that their self-driving car service will begin operations by the end of the year in Phoenix, Arizona. Waymo is a subsidiary of Alphabet, the parent company of Google, and the CEO John Krafcik said engineers at both companies were hard at work on the AI backing their self-driving cars.

People will be able to download a Waymo app and secure rides on autonomous vehicles through it, with no driver present, Krafcik said. Waymo has been operating autonomous vehicles on the roads of Phoenix since October, and is one of the first companies to do so in the US. Originally, Waymo was a part of Google before it was spun off into its own company under the Alphabet umbrella. Despite the separation, members of Google‘s Brain team have helped Waymo engineers by beefing up the neural networks underpinning the AI operating the vehicles.