Microsoft developing AI tool to diagnose cervical cancer faster

Image Source : news.microsoft.com

World Health Organization (WHO) estimates that cervical cancer is the fourth most frequent cancer among women worldwide. India accounts for 16% of global cervical cancer patients and an increase in demand for cervical cancer demand has been worryingly seen.

SRL Diagnostics is the largest diagnostics laboratory in India and they receive around 100,000 pap smear samples every year. A cytopathologist in their Mumbai laboratory alone screens around 200 slides for cervical cancer every day in addition to 100 other slides for other types of cancer.

To solve the problem of the low proportion of cytopathologists available, SRL Diagnostics partnered with Microsoft to create an AI network that would have an API to enable screening out of normal slides, so that the cytopathologists could concentrate on the abnormal slides. This would give a huge boost in the overall speed of screening as only 2% of all samples turn out to be abnormal and need deeper analysis.

To develop the AI algorithm, cytopathologists studied digitally scanned versions of Whole Slide Imaging (WSI) slides manually and recorded their observations, which in turn was used as training data for the AI model. Initially only one cytopathologist was assigned by SRL Diagnostics but since each WSI consists of 1800 tile images, this task proved to be too burdensome for an individual. The matter of subjectivity was also something the AI algorithm had to adjust for.

A digitally scanned image of a WSI slide, used to train the AI model. | Source : news.microsoft.com

“Different cytopathologists examine different elements in a smear slide in a unique manner even if the overall diagnosis is the same. This is the subjectivity element in the whole process, which many a time is linked to the experience of the expert,” reveals Dr. Arnab Roy, Technical Lead New Initiatives & Knowledge Management, SRL Diagnostics

To address the burden of volume, five cytopathologists were assigned to the task across multiple labs in different locations. This lead to thousands of tile images of cervical smear being annotated. Discordant and concordant notes were created for each sample image. Each sample image with discordant notes from a minimum of three cytopathologists were then sent to senior cytopathologists. This is how the issue of subjectivity was handled.

This is the first such AI-Device-Labs setup in the Histopathology space in this part of the world and impacts the entire spectrum of stakeholders. For patients, it reduces the turnaround time for diagnosis and onset of treatment. It lends better productivity and accuracy to the efforts of cytopathologists. For doctors, it offers insights that inspire more qualitative treatment decisions. 

“With the growing burden of cancer, there is a need to quickly and accurately analyze the samples to help clinicians arrive at a diagnosis faster and with a higher degree of objectivity. The work done by the SRL-Microsoft consortium in developing deep learning-based algorithms as an assistive tool in a relatively short span of time, speaks volumes about the capabilities of both the partners. This particular cervical cancer AI API shall be useful in screening liquid-based cytology slide images, unlocking precious dead-time of the pathologists enabling them to report more cases and/or focus more on complicated cases,” adds Arindam Haldar, CEO, SRL Diagnostic.

In August this year SRL Diagnostics launched an internal preview of the API. In a span of three to six months the AI model will be put through rigorous clinical validation protocols. More than half a million anonymized digital tile images will be used in this exercise, making it one of the largest of its kind. Following internal validation, the API will be used in external cervical cancer diagnostics, including hospitals and other diagnostic centers.

This is one of the latest news about Microsoft using the power of AI to solve real world problems in India. Last week they had announced how their HAMS project was helping to automate driving tests in India. You can read about that project here.

Researchers use Lasers to hijack Google Home, Amazon Alexa and Siri

Image Source : lightcommands.com

Voice assistant devices like Google Home and Amazon Alexa have become really popular devices and are viewed as the next big thing in home automation. The open nature of these platforms and the low cost of these devices have led to a huge number of companies coming out with home automation products compatible with either one or both of them. In fact, it is really hard to find electrical appliance companies who have not forayed into this market with their latest offerings.

A team of researchers at the University of Michigan and University of Electro-Communications, Tokyo have found a surprising way to 'hijack' voice assistants with light. This is made possible by a physical phenomenon that allows light to create acoustic pressure waves. The field of research that studies this phenomenon is called photoacoustics and the first work in this field was done by none other than Alexander Graham Bell in 1880.

So, to summarize, microphones can be made to react to laser light directed at them in the same way as they react to sound. Since voice assistants use sound as the interaction medium, this phenomenon allows them to be made to do so by light as well. The most scary part of this is that light waves are inaudible and so people will not be able to detect that a command is being sent to their device.

At the start of this article you can see a diagram illustrating how a laser directed from 75 meters away at a downward angle of 21 degrees and through a glass window was able to make a Google Home open the garage door. Below is a video of the same demonstration:

Now, many of you might be thinking that these devices do have voice recognition and that should be enough to protect your devices from following commands that are not your own. However, it seems that this feature is enabled by default only on smartphones and tablets and not on smart speakers. In addition, only the wake-up commands like "OK Google" and "Alexa" are verified by the devices and not the rest of the commands. To compound this problems it has been found that the verification is quite weak and can be fooled by online text to speech synthesis tools that can imitate people's voice. We have also seen demonstrations of technologies using machine learning to imitate people's voices to a high degree of accuracy.

The researchers were able to send commands to Siri one devices like the iPhone XR and iPad 6th Gen as well, so Apple devices are vulnerable to these hijackings as well.

Two drawbacks for hijackers and advantages for us are that this technology only works in Line of sight and there is no way to for the hijacking system to get the feedback about successful execution of commands from the device unless they are in the audible range of the device's speaker. Nevertheless, until companies come up with really secure counter measures, be wary of connecting devices that could compromise your security if hijacked.

Microsoft's Project Silica succeeds in storing Superman film on a piece of glass.

Image Source: The Verge

Cloud storage has now seamlessly entered into our digital lives to such an extent that we neither realize that we are utilizing it in many instances and nor do we comprehend the fact that all this data is being physically stored in hardware whose capacity in increasingly flattening. This problem is compounded by the fact that the amount of data that each of us generate is increasing exponentially.

Even if the infrastructure keeps up with this rising demand, the hardware such as disk drives itself have a lifespan of around five years. This means that to keep the data saved they have to be cyclically written on to newer hardware.

A unique solution thought of, was to use the same ultrashort optical pulses used in LASIK surgeries to store data in glass by permanently changing its structure. This can help keep the data saved for centuries. Quartz glass also doesn’t need energy-intensive air conditioning to keep material at a constant temperature or systems that remove moisture from the air – both of which could lower the environmental footprint of large-scale data storage.

The hard silica glass used by Microsoft can withstand being boiled in hot water, baked in an oven, microwaved, flooded, scoured, demagnetized and other environmental threats that can destroy priceless historic archives or cultural treasures if things go wrong.

The way the technology works is that a laser encodes data in glass by creating layers of three-dimensional nanoscale gratings and deformations at various depths and angles. The system uses machine learning algorithms to read the data back by decoding the images and patterns that are created as polarized light shines through the glass.

Image Source: The Verge

Project Silica’s infrared lasers encode data in “voxels,” the three-dimensional equivalent of the pixels that make up a flat image. Since the information is being stored within the glass itself rather than its surface like that of traditional storage medium, a 2-mm-thick piece of glass, for instance, can contain more than 100 layers of voxels. 

Warner Bros., which approached Microsoft after learning of the research, is always on the hunt for new technologies to safeguard its vast asset library: historic treasures like “Casablanca,” 1940s radio shows, animated shorts, digitally shot theatrical films, television sitcoms, dailies from film sets. For years, they had searched for a storage technology that could last hundreds of years, withstand floods or solar flares and that doesn’t require being kept at a certain temperature or need constant refreshing.

The result of the collaboration was that the 1978 iconic film "Superman", was successfully stored, as a proof of concept, in a piece of glass roughly the size of a drink coaster - 75mmx75mm and just 2mm thick.