This is what Carla Krivet, a senior at Philips, is responsible for, among other things, the fields of big data and artificial intelligence in the company, in an interview with the science website. Krivet came to Israel to visit the Philips development center in Haifa, and to promote research collaborations with hospitals
"Artificial intelligence will not replace the doctor but will greatly improve the results of his work." This is what Carla Krivet, a senior at Philips, is responsible for, among other things, the fields of big data and artificial intelligence in the company, in an interview with the science website. Philips employs over a thousand workers in the field of medical devices at the development center in Haifa.
Krivat came to Israel to check the possibility of cooperation and joint research with the Hadassah and Sheba hospitals.
Philips Healthcare for the presence in Israel where today it is the most important arm of the Royal Dutch company known to us as a large manufacturer of consumer products but today its focus is on the field of medicine. In recent years, Philips has been using the new strategy, which has brought it to collaborations with academia and research institutions, and with companies in the software and Internet fields (such as Salesforce and Amazon), while it dismantles previous partnerships and withdraws from its traditional markets.
"We don't push technology just for the sake of technology. Artificial intelligence is not intended to replace the doctor, but it will complement the picture of reality that the doctor sees in front of him. The doctors have to learn the material about each patient from scratch and sometimes find data such as photographs from other hospitals."
"It's like today it seems natural to us that we arrive at a certain place, for example when I landed in Israel, that the phone already knows where we are, allows us to know what the transportation options are, and shows us the weather, as well as many other location-based options. There is no equivalent for this in the medical service. True, there are developments, but still a human coder has to go through the x-ray imaging. Doctors still only see the medical file of the specific patient, but to get the overall picture they have to examine each medical index separately, and rely on the patient's complaints. Imagine what will happen when all the devices in the hospital, as well as other medical documents, will be processed by artificial intelligence, which will present the doctor with everything he needs to know in that context, and he will also be able to receive recommendations on treatments given to patients in a similar situation, as well as a comparison with historical events of the patient - for example previous EKG tests with more precision than the doctor's human eye is capable of."
"In recent years, we have mainly promoted artificial intelligence in two fields - the field of radiology and the field of cardiology."
"In the field of radiology, we at Philips teach the computer to read X-rays or CT imaging, MRI and more. Compare with previous photographs and analyze the situation to present it to the doctor. The computer analyzed a huge number of samples and will know how to identify markers of diseases. Today if you go to a radiologist who has just started his work, the accuracy of the diagnosis may be less than that of a university professor, who has seen thousands of cases, if you were lucky enough to reach him."
"In the field of cardiology - heart medicine, the idea is to install many sensors that will be connected to the patient, and let the computer analyze his condition at any given moment, and alert the nurse in the cardiac intensive care unit that one of the patients is not feeling well. Today such a nurse runs from patient to patient according to the calls, and may miss people who are in a difficult situation. Instead, she will receive an indication from the computer about the deterioration in the condition of patients and what is recommended to be done to stop it in time. I have spent nights with nurses on duty and know how tired they are and if they see someone behaving a little strange, they don't know if he is sleeping or if his illness has worsened. The technology in this case can save human lives. The goals of the early warning system in cardiac intensive care units - decrease the mortality rate, reduce treatment costs, streamline the work of the nurse, and increase the satisfaction of the medical staff and patients."
Will such systems be able to monitor the sick and elderly in their homes?
Krivet: "The technology already exists today, but there is no entity capable of receiving the alerts and reacting to them. When such centers are established, the artificial intelligence will be able to alert and tell what exactly happened, so that they can provide the best treatment. Not to mention that it will be possible to save unnecessary hospitalizations, and with them also exposure to bacteria resistant to antibiotics."
"By the way, when it comes to antibiotic-resistant bacteria, it will be possible to trace the route of infection back to the original infectious agent from test data that is already being done today, but there is no one to make the comparison. This is important from a legal point of view because hospitals are exposed to lawsuits if it is proven that the infection was carried out by medical staff members. Such surveillance could allow them to defend themselves."
What is the uniqueness of your activity in Israel?
"We have a large development center in Haifa with about eight thousand employees, and we are also partners in the chief scientist's SANARA incubator together with Teva. Israelis are entrepreneurs by nature, in other places people are more cautious. This may have been due to necessity. The people are very smart, eager to get things done quickly and learn to adapt.
By the way, most of the employees of the research and development center in Israel are software developers. We believe that the next differentiation in medical devices will no longer be about functionality but about the wisdom of the devices - machine learning capability.
In conclusion, Krivet says that at Philips, the development of intelligent software is seen as the beginning of a period in which, more than ever, artificial intelligence will become the cutting edge of the medical system. More intelligence, and less hardware. And Israel is the center. Israel, according to Phillips, is a country with a great value of intellectual energy, therefore it is at the center of the development of artificial intelligence.
See more on the subject on the science website:
2 תגובות
In my opinion Phillips are not aimed really low. They entered the field developed by Professor Michael Elad and Professor Yonina Elder
From the Technion called compressed sensing, which means that an MRI, regardless of artificial intelligence, will take 10 minutes, not an hour.
But it is indirectly related to intelligence. Michael Elad is a mathematical wizard for artificial intelligence who turns out doctoral students who become professors like someone else makes donuts. I know two. And shortening the time of the MRI means shortening the calculations of the MRI analysis, and performing the MRI analysis, by artificial intelligence with a higher complexity.
The statement is perhaps to mitigate objection.
The aforementioned statement a. Refines the introduction of artificial intelligence into the field of medicine in order to prevent resistance.
B. mainly reflects the difficulty of designing a deep network - but not its capabilities once it has been designed correctly.
In my opinion, if Philips aims low, Google aims high. They will try to go completely on an expert MRI decoding doctor who is a brain. From an economic point of view - it will be decisive: it is cheaper to buy MRI decoding software, than to hire a senior doctor.
third. It is a field open to research = they have not yet discovered the end of their ability. I have never met a field other than quantum physics
where the mathematics is so fascinating, difficult, I estimate it will take me a year or two to master it, and includes engineering knowledge
In addition, a parallel that also requires control. Mathematics is statistics of random variables at its highest possible level. There is a very promising direction that combines Claude Shannon's information theory with statistics. It's fascinating.
We arrive at numerical communication (Viterbi code). And there is another direction of thermodynamics. It is enough if we look at the called function
sigmoid and we see a Fermi-Dirac distribution there. The networks imitate it. and explain that this is how they create unit vectors of features of the analyzed information. That is, one feature in each layer is enhanced, and all the others are degenerated. I never thought of thermodynamic distribution that way.
I have found a niche where the number of intelligent researchers is now small. And in addition, when I master engineering and mathematics, I will return to Israel
To say something in the field of artificial intelligence development is more advanced than today.
I am currently executing some software code. Today I happened to poke through XNUMXD MRI code available in open source of course in the GOOGLE software code database. Such an ability to develop at home, when those who are smarter, have no less chances of success than a resource-rich company, did not exist before.
The exposed database is free, and if you pay $7 per month you get an even bigger pool of software code.