Watch 44 million atoms simulated using AI and a supercomputer

 AI Simulating

Watch 44 million atoms simulated using AI and a supercomputer   

This simulation models a huge number of atoms in detail with the help of artificial intelligence

By Alex Wilkins, 16 May 2023  in New Scientist

The most accurate simulation of objects made from tens of millions of atoms has been run on one the world’s top supercomputers with the help of artificial intelligence.

Existing simulations that describe in detail how atoms behave, interact and evolve are limited to small molecules, because of the computational power needed. There are techniques to simulate much larger numbers of atoms through time, but these rely on approximations and aren’t accurate enough to extract many detailed features of the molecule in question.

Now, Boris Kozinsky at Harvard University and his colleagues have developed a tool, called Allegro, that can accurately simulate systems with tens of millions of atoms using artificial intelligence.

Kozinsky and his team used the world’s 8th most powerful supercomputer, Perlmutter, to simulate the 44 million atoms involved in the protein shell of HIV. They also simulated other common biological molecules such as cellulose, a protein missing in people with haemophilia and a widespread tobacco plant virus.

“Anything that’s essentially made out of atoms, you can simulate with these methods at extremely high accuracy, and now also at large scale,” says Kozinsky. “This is one demonstration, but by no means constrained to this domain.” The system could also be used for many problems in materials science, such as investigating batteries, catalysis and semiconductors, he says.

To be able to simulate such large numbers of particles, the researchers used a kind of AI called a neural network to calculate interactions between atoms that were symmetrical from every angle, a principle called equivariance.  ...' 

Can AI save lives? Cancer Detection Study Suggests Yes

Cancer Detection

Can AI save lives? Cancer detection study suggests yes

New research shows algorithm efficient at detecting recurrence in high-risk patients

May 1, 2023 - 5:00 pm

Much of the world may currently be fretting about how to limit the impact (lack of privacy, copyright issues, loss of jobs, world domination, etc.) of artificial intelligence. However, that does not mean that there isn’t enormous potential for AI to improve quality of life on earth. 

One such application is healthcare. With the ability to process big data sets, the deployment of AI could lead to significant advances in predictive diagnostics, including early detection of cancer. While more research is needed, one of the latest studies in the field shows promising results for AI-assisted diagnosis of lung cancer. 

Doctors and researchers at the Royal Marsden NHS foundation trust, the Institute of Cancer Research, and Imperial College London have built an AI algorithm they say can diagnose cancerous growths more efficiently than current methods. 

In the study named OCTAPUS-AI, researchers used imaging and clinical data from over 900 patients from the UK and Netherlands following curative radiotherapy to develop and test ML algorithms to see how accurately the models could predict recurrence. 

Specifically, the study looked at if AI could help identify the risk of cancer returning in non-small cell lung cancer (NSCLC) patients. Researchers used CT scans to develop an AI algorithm using radiomics. This is a quantitative approach which extracts novel data and predictive biomarkers from medical imaging. 

Research algorithm superior to current technology

NSCLC patients make up 85% of lung cancer cases. While the disease is often treatable when caught early, in over a third of patients, the cancer returns. The study found that using the algorithm, clinicians may eventually be able to identify recurrence earlier in high-risk patients. 

The scientists used a measure called area under the curve (AUC) to see how efficient the model was at detecting cancer. A perfect 100% accuracy score would be a 1, whereas a model that was purely guessing 50-50 would get 0.5. In the study, the AI algorithm built by the researchers scored 0.87. This can be compared to the 0.67 score of the technology currently in use. 

“Next, we want to explore more advanced machine learning techniques, such as deep learning, to see if we can get even better results,” Dr Sumeet Hindocha, Clinical Oncology Specialist Registrar at The Royal Marsden NHS Foundation Trust, and Clinical Research Fellow at Imperial College London, said. “We then want to test this model on newly diagnosed NSCLC patients and follow them to see if the model can accurately predict their risk of recurrence.”...' 

Arterys: The future of precision medicine

Brought to my Attention , aimed at healthcare practices...

Arterys:  The future of precision medicine   that only human + AI can achieve. 

The Arterys platform extracts actionable insights from medical images to add clinical value, improve diagnostic decision making, efficiency and productivity.

Arterys is the medical imaging AI platform allowing you to weave leading AI clinical applications directly into your existing PACS or EHR driven workflow to make it a natural extension of what you already do.

We are making AI real by improving physician experience, accuracy of diagnosis and treatment, financial performance and outcomes that matter to patients and providers.

Accessible anywhere from any validated device via the cloud for faster performance, ease of deployment with no PHI exchange and completely secure.

See how Arterys is transforming healthcare through deep learning and AI

History:  It started at Stanford

Before we founded Arterys, we were graduate students, searching for a way to make medicine better. Faster. More precise.

We pushed the boundaries of our disciplines until they finally overlapped, intersecting cloud computing with cutting edge medical image acquisition.

Soon after, Arterys was founded on our shared beliefs:

Faster, smarter diagnosis where it counts the most

We started by improving diagnoses for newborns and kids with heart defects. At the time, pediatric cardiovascular disease was diagnosed with ultrasound, which only offers a partial view of the heart with no blood flow quantifications. Or an MRI captured over grueling hours on the scanning table.

Without accurate quantification, cardiac medicine was an educated guess with a high error rate.

The solution was obvious: 4D Flow technology to visualize and quantify blood flow in mere minutes.

But image archiving servers in hospitals couldn’t read 4D Flow’s big data files. So we applied a system that could: cloud computer processing.

With cloud computing, we could put those life-saving 4D Flow images into radiologists hands. In mere moments, physicians could diagnose and make accurate treatment decisions.

Anyone with access to a web browser could access Arterys to quantify regurgitant flow and determine if a child with heart defects needed surgery.

We were helping to save kids’ lives with better diagnoses.

But we weren’t satisfied.

Even though we put 4D Flow at radiologists’ fingertips, we still saw physicians manually drawing contours to quantify the size of cardiac ventricles.

Artificial Intelligence gets real

We decided to power Deep Learning AI with cloud computing GPUs to automatically quantify and segment ventricles as accurately as manual measurements by experienced physicians. In 2017, our technology received the first ever US FDA clearance for leveraging cloud computing and deep learning in a clinical setting.

But we were still not satisfied.

Physicians around the world approached us, asking if we could apply the same AI cloud computing approach to image processing and analysis to cancer patients. Then to liver patients. Then lung, breast, brain, and everything in between.

The word was getting out that radiologists could receive automatic accurate measurements through deep learning powered in the cloud.

All of your data, all in one place. Simple.

Despite our breakthroughs, existing radiology workflows put AI image acquisition and analysis out of reach. Working outside of a unified platform, physicians switched interfaces multiple times an hour to access different AI imaging tools. This stole vital time that could be spent with a patient. It was time to transcend silos with one united AI web platform.

Soon after, we received FDA clearance for our Arterys web-based AI imaging platform.

Despite these successes, we are still not satisfied.

We want to push medicine to make it even faster and more precise. We’ve just gotten started  .... ..'

Massive Cell Data sets

Analyzing organisms with minimal computing power. 

Algorithm Uses Online Learning for Massive Cell Datasets

Michigan Medicine, Kelly Malcolm, April 19, 2021

An algorithm developed by University of Michigan (U of M) researchers employs online learning to accelerate the analysis of enormous cell datasets, using the amount of memory found on a standard laptop computer. The algorithm enables new datasets to be added to existing ones without reprocessing the older datasets, and allows researchers to segment datasets into mini-batches so less memory is required for processing. U of M's Joshua Welch said, "Our technique allows anyone with a computer to perform analyses at the scale of an entire organism. That's really what the field is moving towards."

Autonomous Cars Delivering Meds in London

More attempts at automated delivery:

Autonomous Green Robot Cars to Deliver Medicine Around London

Interesting Engineering By Chris Young

A fleet of autonomous, electrically-powered green robot vehicles has started delivering medicine to care homes in London's Hounslow borough, as part of a public trial. The Kar-go, from U.K. startup Academy of Robotics, will be the first custom-built autonomous delivery vehicle to conduct last-mile deliveries on public roads in Britain. The robot car can travel at 96 kilometers (60 miles) per hour, carry a maximum load of 48 parcels, and use artificial intelligence to sort out parcels and calculate the speediest delivery route. The initial trials will have human operators sitting inside the Kar-gos, before they eventually transition to fully autonomous driving. The vehicle will drive itself to and from the sender and recipient's address, with a smartphone application alerting the recipient upon arrival. A robotic conveyor within the Kar-go enables contact-free parcel handover  ... " 

Who Owns Smartphone Data

Continuing look at data and ownership and security.

Who Has Access to Your Smartphone Data?    By Keith Kirkpatrick

Communications of the ACM, October 2020, Vol. 63 No. 10, Pages 15-17 10.1145/3416078

In a world that is defined by the generation and collection of data by technology and communications companies, personal information—including where people go, with whom they associate, what they purchase, and what they read, listen to, and even eat—it is quite a simple task to create a detailed profile of an individual based solely on the data captured in his or her phone.

The right to access and use the cache of personal information stored in each person's smartphone has become a major question about balancing personal privacy rights against governments' desire to monitor and retrieve data about its citizens' activities for law enforcement, public safety, and health issues. While much of the attention over the past several years has focused on demands from law enforcement to access this data to aid in criminal investigations, the COVID-19 pandemic of 2020 has refocused the debate on the government's right to access location data during health or other public safety emergencies.

Within the U.S., the primary communications privacy law that regulates the disclosure of and access to electronic data held by communication services providers, including wireless carriers, Internet Service Providers (ISPs), social media platforms, and search companies, among others, is the Electronic Communications Privacy Act of 1986 (ECPA) which, along with the Uniting and Strengthening America by Providing Appropriate Tools Required to Intercept and Obstruct Terrorism (USA PATRIOT) Act OF 2001, protects wire, oral, and electronic communications while those communications are being made, are in transit, and when they are stored on computers. As the Act explicitly states, "Some information can be obtained from providers with a subpoena; other information requires a special court order; and still other information requires a search warrant."

Andrew Crocker, senior staff attorney on the Electronic Frontier Foundation's civil liberties team, says the ECPA generally "requires the government to use legal process to get data about users," rather than simply allowing them to request and receive information from service providers.  .... " 

Ethical Questions about AI Surgery

Wondered at first where the ethical questions were coming from.  Access to advanced methods?  No, it seems privacy was once again involved.

Researchers examine the ethical implications of AI in surgical settings
Kyle Wiggers

A new whitepaper   https://arxiv.org/pdf/2007.14302.pdf   coauthored by researchers at the Vector Institute for Artificial Intelligence examines the ethics of AI in surgery, making the case that surgery and AI carry similar expectations but diverge with respect to ethical understanding. Surgeons are faced with moral and ethical dilemmas as a matter of course, the paper points out, whereas ethical frameworks in AI have arguably only begun to take shape.

In surgery, AI applications are largely confined to machines performing tasks controlled entirely by surgeons. AI might also be used in a clinical decision support system, and in these circumstances, the burden of responsibility falls on the human designers of the machine or AI system, the coauthors argue.  .... " 

Internet of Medical Things

From the ACM, Vint Cerf talks about the current and near future state of digital medicine.

On the Internet of Medical Things
By Vinton G. Cerf
Communications of the ACM, August 2020, Vol. 63 No. 8, Page 5  10.1145/3406779

Google Vice President and Chief Internet Evangelist Vinton G. Cerf

In my last column (June 2020), I wrote about my experience with COVID-19 and the challenges involved with getting medical attention. The problem is still with us, even with the improved availability of personal protection equipment and masks. The experience of calling for a doctor's appointment and being told I could not come into the doctor's office was unsettling to say the least. A "video consultation" was all that was offered. My reaction was "Wait, you won't be getting any vital signs or other medical information that way!" This led to the natural conclusion that remote detection would be helpful in these conditions. Telemedicine has long been of interest, especially for treating patients in rural or isolated areas where physicians and hospitals may be in short supply or absent entirely. Wearable sensors have become popular items for people who want to track their daily exercise or challenge themselves to exceed past performance with new records. ... "

AR Simulation for First Responders

An example of where AR can work best, where the specific examples are not rare, yet are relatively easy to construct in a form of simulation.

AR Tool Shown to Help Surgeons Remotely Guide First Responders in Battlefield-Like Scenarios
By Purdue University News  via ACM

Researchers from Purdue University and the Indiana University School of Medicine have developed and tested an augmented reality (AR) headset that helps surgeons remotely guide medics through performing surgery in simulated war zones.

The System for Telementoring with Augmented Reality (STAR) transmits a recorded view of the operating environment to the surgeon, who employs a touchscreen to annotate the recording with drawings of how to complete a surgical procedure.

First responders can see the surgeon's instructions directly on their view through the headset.

Tests evaluated first responders performing cricothyroidotomies to open a blocked airway on a patient simulator in both indoor and outdoor settings, with battlefield-like visual and audio distractions; even those with little to no experience beforehand operated successfully after receiving instructions from surgeons through the system.  ... 

From Purdue University News
View Full Article 

Bio-Ink for 3D Printing Inside the Body

Quite a remarkable possibility when linked to medical robotics.

Bio-Ink for 3D Printing Inside the Body
IEEE Spectrum
Charles Q. Choi
July 1, 2020

Researchers at The Ohio State University (OSU) have developed a bio-ink that can be three-dimensionally (3D) printed at human body temperature, and solidified using visible light. Bio-inks are composed of living cells suspended in a gel and are safe for use inside people, potentially paving the way for 3D printing inside the body. Using a 3D printing nozzle affixed to robotic machinery to dispense bio-ink in a controlled manner, the researchers were able to bio-print onto soft materials, with interlocking knobs left beneath the surface anchoring the printed structure to the body like surgical staples. OSU's David Hoelzle said the goal is not to bio-print an entire organ, but to "augment a standard surgery by delivering a biomaterial with a tethered growth factor to jumpstart healing, or a tethered drug to prevent infection. ... We envision a biomaterial bio-ink printing tool as another tool in the surgeon's toolset."  ... ' 

Biomimicry for Regenerative Medicine

Could the cells in these simple worms point towards a mechanism for regeneration of organs in humans?  Considerable,  non technical article with links to more .....

Flatworms muscle new eyes' wiring into their brains

Peter Reddien's lab at the Whitehead Institute takes a step forward in understanding how neural circuits could be regenerated in adults.

Eva Frederick | Whitehead Institute

If anything happens to the eyes of the tiny, freshwater-dwelling planarian Schmidtea mediterranea, they can grow them back within just a few days. How they do this is a scientific conundrum — one that Peter Reddien's lab at Whitehead Institute has been studying for years.

The lab's latest project offers some insight: in a paper published in Science June 26, researchers in Reddien's lab have identified a new type of cell that likely serves as a guidepost to help route axons from the eyes to the brain as the worms complete the difficult task of regrowing their neural circuitry.

Schmidtea mediterranea's eyes are composed of light-capturing photoreceptor neurons connected to the brain with long, spindly processes called axons. They use their eyes to respond to light to help navigate their environment.  .... 

This study is a step forward in a body of work that aims to expand the capabilities of regenerative medicine. “Imagine a scenario where someone experiences a spinal cord injury or an eye injury or stroke that leads to the loss of a neural circuit,” says Atabay. “The reason we can't fully cure these cases today is that we lack fundamental information regarding how these systems can regenerate. Looking at regenerative organisms provides a lot of insights. From this case, we see that regenerating the lost system may not be enough; you may also need to regenerate systems that are properly patterning that system.” ... ' 

Virtual Metabolic Human Models

Considerable complexity in such models, but they at least will point to the kinds and amounts of data and predictions that will be necessary to provide computational results.   A kind of Digital Twin.

Virtual Metabolic Humans—Harvey and Harvetta, Novel Computational Models for Personalized Medicine
Science Foundation Ireland
June 3, 2020

Researchers at National University of Ireland and the Netherlands' Leiden University have created whole-body human computational models for personalized medicine called Harvey and Harvetta. The models can simulate individual metabolisms, physiologies, diets, and gut microbiomes; they also can predict known biomarkers of inherited metabolic diseases, and facilitate investigation of potential metabolic interactions between humans and their gut microbiomes. Harvey and Harvetta are anatomically interconnected whole-body virtual male and female models incorporating more than 80,000 biochemical reactions distributed across 26 organs and six blood-cell types. Their development required the creation of novel algorithms and software for constraint-based simulation of high-dimensional biochemical networks. Leiden University's Ines Thiele said, “Harvey and Harvetta will usher in a new era for research into causal host-microbiome relationships and greatly accelerate the development of targeted dietary and microbial intervention strategies.”

Virtual Care Services

Experienced this recently, was nicely done.   Though I would require at least some face to face and visual analysis as part of the care.   Might this be done by using images and sharing these ahead of the appointment?

Forrester writes about the quicker adoption.  Part 2 with links to the first part.

Virtual Care Is A Requirement — Not A “Nice-To-Have”
Arielle Trzcinski, Senior Analyst  Forrester
Virtual Care Blog Series — Part 2

The pandemic has pushed virtual care technologies to make the leap across the chasm of adoption as even the pragmatists and conservatives have started deploying these services. Forrester made the call that 2020 would be the tipping point in virtual care adoption, and the pandemic accelerated this shift as barriers that inhibited faster adoption have been removed, such as lack of consumer awareness, cost and reimbursement hurdles, and the ability for patients to connect with their existing provider that they trust.

To better understand how adoption has played out in the market, what areas are seeing the greatest amount of growth, and to establish a baseline in virtual care, Forrester has been connecting with the supply side — the vendors in the virtual care space. These vendors have reported a significant rise in enrollment, adoption, and new implementations by healthcare organizations (HCOs). As part of our weekly series on virtual care, we are continuing to highlight how virtual care is transforming the future of healthcare. Missed the previous blog? Check it out here.  .... " 

Organs on a Chip

Had seem examples of this some years ago, to what degree has it been predictively accurate?

Using “organs-on-a-chip” to model complicated diseases
A new approach reveals how different tissues contribute to inflammatory diseases such as ulcerative colitis.

Anne Trafton | MIT News Office

Press Inquiries
MIT biological engineers have created a multitissue model that lets them study the relationships between different organs and the immune system, on a specialized microfluidic platform seeded with human cells.

Using this type of model, sometimes called “organs-on-a-chip” or “physiome on-a-chip,” the research team was able to explore the role of circulating immune cells in ulcerative colitis and other inflammatory diseases. They also discovered that a metabolic byproduct generated by bacteria living in the human gut plays an important role under these inflammatory conditions.

“We’ve shown that now you can start to attack some of these really thorny, chronic inflammatory diseases by designing experiments in these organs on chips,” says Linda Griffith, the School of Engineering Professor of Teaching Innovation, a professor of biological engineering and mechanical engineering, and the senior author of the study.

This approach, described today in the journal Cell Systems, could also lend itself to studying many other complex diseases, the researchers say.

“Now we have options to really decrease or increase the level of disease complexity, under controlled and systematic conditions,” says Martin Trapecar, an MIT postdoc and the lead author of the paper.

Complex models

Nearly 20 years ago, Griffith’s lab first began working on a model of the human liver known as the “liver chip.” This system, which consists of engineered human liver tissue grown on a specialized scaffold, can be used to test drug toxicity. More recently, she has been working on small-scale replicas of many interconnected organs, also known as microphysiological systems (MPS). In 2018, she reported the development of a platform that could be used to model interactions between up to 10 organs at a time. .... ' 

Thinking about the Laws of Voice

Note that this article points to an interesting book on the topic.  Have not read, but plan to.

Voice Assistants in Pre and Post-Operative Care and the Duty to Warn Patients of Remote Risks – A Legal Discussion       By Eric Hal Schwartz

This guest post is an edited extract from Voice Technology in Healthcare, Chapter 14, The Laws of Voice, by Bianca Phillips, an officer of the Supreme Court of Victoria, Australia and Heather B. Deixler, a senior associate at Latham & Watkins LLP.

Voice assistants will increasingly provide patients with pre- and post-surgery information. The information may be of a general nature, or the skill could be more personalized. If a potential risk of surgery is so remote that it is rarely seen in practice, and only appears in archaic medical literature, does the voice assistant need to advise of that risk?

THE SCENARIO
Imagine the following hypothetical scenario. A patient, Stacey, is about to undergo cataract surgery. In the pre-op appointment, Stacey’s surgeon highlights the risks of surgery and, in her concluding remarks, advises Stacey that “you may also consult your voice assistant to learn more about your surgery, and ask questions about the risks and post-op process.”

That evening Stacey goes home and says to her voice assistant (VA) “[VA name], what are the risks of cataract surgery?” The VA responds “the risks of cataract surgery include: Posterior capsule opacity (PCO), Intraocular lens dislocation, Eye inflammation, Light sensitivity, Photopsia (perceived flashes of light), Macular edema (swelling of the central retina), Ptosis (droopy eyelid) and Ocular hypertension (elevated eye pressure).” Stacey continues to ask a range of questions about post-operative care and the recovery process.

Surgery seems to have gone well. However, two months after the surgery Stacey advises her doctor of considerably decreased vision in her left eye, which the surgeon determines is sympathetic ophthalmia. The estimated post-operative occurrence is between 0.01%–0.05%. Attempts to treat the sympathetic ophthalmia fail and the patient sustains a permanent loss of vision in her left eye. The surgeon failed to inform Stacey of the risk of sympathetic ophthalmia during the pre-op consultation, and Stacey’s VA also did not inform her of this risk. Stacey wants to know whether the surgeon and/or the VA had a duty to inform her of the remote risk. .... " 

Robot Arm for Frontline Medicine

Taking up front information, saving it in context, sterilizing equipment, could save lots of time when dealing with large groups.  Training in context, as it is feasible, could be useful as needs change.  Standardizing methods also very useful.

Robotic Arm Designed in China Could Help Save Lives on Medical Frontline
Reuters
by Martin Pollard

Researchers at Tsinghua University in China have developed a robot that can perform ultrasounds, take mouth swabs, and listen to sounds made by a patient's organs. The system, which consists of a robotic arm on wheels, could help save lives during the coronavirus outbreak. The researchers converted two mechanized robotic arms with the same technology used on space stations and lunar explorers. The robots were almost entirely automated, and could disinfect themselves after performing actions involving contact with patients. Doctors in China are currently training on the robots—one of which is at the team's lab at Tsinghua University and the other at the Wuhan Union Hospital. Said Tsinghua's Zheng Gangtie, “Doctors are all very brave. But this virus is just too contagious ... We can use robots to perform the most dangerous tasks."

Man Diagnosed with Coronavirus Treated Largely by Robot

Achieving isolation for treatment appears to the primary goal.

Man Diagnosed with Wuhan Coronavirus Is Being Treated Largely by a Robot
CNN  by Nicole Chavez

The first person diagnosed with the Wuhan coronavirus in the U.S. is being treated with the help of a robot at Providence Regional Medical Center in Everett, WA. The robot is equipped with a stethoscope to help doctors take the man's vital signs; it communicates with him via a large screen. "The nursing staff in the room move the robot around so we can see the patient in the screen, talk to him," said the medical center’s Dr. George Diaz, adding that the use of the robot minimizes exposure of medical staff to the infected man, who remains in isolation ... "

Will AI Save Lives?

More on the healthcare uses pf AI methods for accurate diagnosis:

It's too soon to tell if DeepMind's medical AI will save any lives
Artificial intelligence trained on health records can now detect kidney injury up to two days before it occurs. The idea is that an advance warning could help doctors intervene earlier to prevent irreversible damage to the kidneys.

AIs are already touted as rivals to doctors when it comes to detecting medical conditions such as certain cancers and childhood illnesses. But few undergo rigorous clinical trials, so it’s still too early to know whether they are effective in practice.

Nenad Tomašev at DeepMind and colleagues trained an algorithm to predict the likelihood that a person who was admitted to hospital would go on to develop acute kidney injury (AKI).

They trained the AI using de-identified electronic health records from 703,782 US veterans aged between 18 and 90, who were admitted to hospital between October 2011 and September 2015.

Read more: AIs that diagnose diseases are starting to assist and replace doctors

AKI results in a dramatic drop in the rate at which the kidneys filter blood. This causes a decrease in urine production and a build-up of waste products in the blood, such as creatinine, a by-product of muscle breakdown. Both of these are used as measures for diagnosis.

Based on creatinine levels from a patient’s medical records, at a given time point the AI predicted whether a kidney injury would occur within the next 48 hours. Its accuracy was confirmed by comparing the prediction to whether the patient was later diagnosed.

The algorithm was fairly accurate at predicting the most severe forms of AKI. It correctly predicted 90 per cent of the cases in which the patient’s kidney function deteriorated so severely that they eventually required long-term dialysis.

It is difficult for doctors to anticipate kidney injury, so that level of accuracy is significant given the consequences of a severe injury, which include death or the need for a kidney transplant, says Eric Topol at Scripps Research in the US, who was not involved in the research.

However, the algorithm was far less accurate for all forms of AKI, correctly predicting only 55.8 of all episodes, with a ratio of two false alerts for one correct prediction. .... "

Google DeepMind Links ID with Decision Process

Quite an interesting claim.   The use of deep learning methods to identify problems using data and then applying process embedded solutions.   Here in the area of medicine:  Diagnosing 3D retinal scans.  The method being more transparent than simple deep learning methods.   And much closer to addressing process models and applications.  Will this solve the 'black box' (non transparency) problem of neural AI?  To be seen, but I like the idea.

Google DeepMind might have just solved the “Black Box” problem in medical AI

Deep Mind’s study published last week in Nature Medicine, presenting their Artificial Intelligence (AI) product capable of diagnosing 50 ophthalmic conditions from 3D retinal OCT scans. Its performance is on par with the best retinal specialists and superior to some human experts.

This AI product’s accuracy and range of diagnoses are certainly impressive. It is also the first AI model to reach expert level performance with 3D diagnostic scans. From a clinical point-of-view, however, what is even more groundbreaking is the ingenious way in which this AI system operates and mimics the real-life clinical decision process. It addresses the “Black Box” issue which has been one of the biggest barriers to the integration of AI technologies in healthcare.

DeepMind’s AI system addressed the “Black Box” by creating a framework with two separate neural networks. Instead of training one single neural network to identify pathologies from medical images, which would require a lot of labelled data per pathology, their framework decouples the process into two: 1) Segmentation (identify structures on the images) 2) Classification (analyze the segmentation and come up with diagnoses and referral suggestions)  .... "

Robotic Blood Vessel Instrument Guidance

Seen this movie.  Pgressin this area continues to evolve

Instrument Guidance Through Deep, Convoluted Blood Vessel Networks
Polytechnique Montreal
December 3, 2019

Researchers at the Polytechnique Montreal Nanorobotics Laboratory in Canada have developed a robotic platform that can guide endovascular surgery through deeper, difficult-to-access blood vessel networks than previously possible. The Fringe Field Navigation (FFN) method taps the magnetic field that the superconducting magnet of a clinical magnetic resonance imaging (MRI) scanner produces. The platform utilizes a robotic table positioned within the fringe field by the scanner; this table moves on all axes to position and orient the patient, based on the direction in which a surgical instrument must be guided through the body. The table automatically shifts direction and orientation to position the subject for successive stages of the instrument's journey, with FFN mapping the directional forces of the scanner's magnetic field.  .... "