Artificial intelligence (AI) works with human diagnostic experts and can do more than they do independently. Qum.ai, based in Mumbai, is using deep learning to quickly diagnose and recommend personalized treatment options.
It is estimated that artificial intelligence and automation will replace half of jobs in the next 10 years. However, global technology leaders seem to have differences over the potential long-term impact.
SpaceX and Tesla's CEO Elon Musk believe that from the current speed, artificial intelligence has the potential to bring a catastrophic future and requires active regulation. On the other hand, Facebook CEO Mark Zuckerberg called Musk a singer and accused his "doomsday fear" as unnecessary negative emotions.
Tom Gruber was the co-founder of Siri (later acquired by Apple) and he saw a middle road. In a recent TED talk, he mentioned that "humanized artificial intelligence" (human and artificial intelligence work together) can actually accomplish more tasks than humans and artificial intelligence independently. He gave an example of how artificial intelligence classifiers and human pathologists can achieve 99.5% accuracy.
Founded in January 2016, Qure.ai is a Mumbai-based startup that is trying to achieve similar feats by integrating the power of machine learning and human diagnostics experts. The startup uses deep learning techniques to help diagnose diseases and recommend personalized treatments from medical imaging data such as CT, MRI scans and X-rays.
In addition, Qure.ai is one of the 30 most promising Indian technology startups selected by YourStory's Tech in 2017. Here are their stories:
From left to right - Pooja Rao (Director of Research and Development) and Prashant Warrier (CEO)
Founded by Pooja Rao (head of R&D department) and Prashant Warrier (CEO), Qure.ai has a strong technical background in medicine and machine learning. Pooja is a highly trained doctor with a Ph.D. in International Max Planck Institute for Neuroscience, with a focus on neuroscience and machine learning. She previously worked at the National Genomics Institute in the United States and worked as a data scientist at GoDataDriven in Amsterdam.
Prashant holds a bachelor's degree in technology from the Indian Institute of Technology in Delhi and a Ph.D. from the Georgia Institute of Technology, and has established a personalized digital marketing company called Imagna AnalyTIcs, which was acquired by Fractal AnalyTIcs in 2015.
At the beginning of Qure.ai, a small team of engineers was established in Mumbai in January 2016. Now, more than a year later, Qure.ai is still a streamlined team of 14 employees.
In an interview with YourStory, Pooja explained the difficulties faced by doctors and patients. The volume and complexity of medical imaging has increased dramatically over the years. For doctors, this means an ever-increasing workload and therefore a higher chance of misdiagnosis and error. In some cases, a small mistake is enough to bring life to the patient.
Based on their knowledge of the field, Pooja and Prashant know that deep learning techniques can process large amounts of medical data and detect anomalies based on a fixed pattern. Therefore, they decided to co-found Qure.ai and spent about a year researching and developing and constantly training their algorithms.
Pooja explained that with regard to how machines and humans work together to reduce the possibility of misdiagnosis, "our algorithm reads X-ray or MRI scans to detect abnormalities. If an abnormality occurs, it will point out to the doctor, the doctor will Passing this information is equivalent to getting an enhanced version of the scan results, which makes their work easier and more efficient. This also prevents them from missing information they may not see."
How does Qure.ai work?Radiologists use sophisticated software to view the scan results of these professions. Currently, the products offered by Qure.ai are either developed independently or based on existing technology.
Qure.ai's artificial intelligence is working
Pooja explained that one of the biggest uses of its technology is to prevent medical accidents such as accidents or strokes. she says,
"If properly diagnosed and treated in time, it can effectively treat stroke." It can help detect bleeding and fractures more quickly, thus preventing medical accidents.
Deep learning requires a lot of high quality data. According to Pooja, Qure.ai relies on partners and open medical data to train its algorithms. They currently have a large CT scan database and an MRI database. “Our chest X-ray products have processed more than 1.5 million X-rays.â€
Regarding accuracy, Pooja pointed out that its chest X-ray solution is close to human precision, and its quantitative solution has become more accurate than humans.
Qure.ai's solution has been deployed to four clinics in Mumbai. Pooja said their early clients include radiology specialists, hospitals and medical device manufacturers. The startup adopts a SaaS-based subscription model based on customer size and customization needs.
According to Pooja, deploying artificial intelligence in hospitals is an exciting task and a challenging task. Because of the large amount of data, Pooja believes that “India is one of the best places to deploy artificial intelligence in the real world. We have collected feedback on the usefulness of artificial intelligence to diagnostic experts.â€
Industry overview and future plansAccording to a report by U.S. technology company Nvidia, medical records, such as doctor reports, test results and medical images, are gold mines for health information. This reaffirms the belief that using GPU-accelerated deep learning to process and study a patient's condition and compare the patient to more other cases can provide a better treatment for the doctor.
Big companies and startups want to do business in this area. Google researchers have designed an eye-scanning algorithm to detect a common blindness, showing the potential of artificial intelligence to change medicine. The algorithm can observe retinal images and find diabetic retinopathy.
Photo: Pooja at TechSparks (2017)
New York-based Behold.ai and San Francisco-based Arterys can use common radiographic data and leverage to identify diseases and visualize and quantify the flow of the human body using any MRI machine.
San Francisco-based EnliTIc is another company that analyzes medical images to identify tumors and potential fracture risks. Qure.ai's work seems to be very similar to Behold.ai and EnliTIc.
Currently, Qure.ai is seeking to establish operations in Mumbai and Delhi, but their long-term goal is to expand into other parts of India. When talking about the possibility of overseas expansion, Pooja said, “When we have all the regulatory frameworks, we plan to expand beyond India.â€
·Basic precautions
Do not put expensive oil into low-quality cartridges to avoid wastage. Most pre-filled oil cartridges have the so-called 510 thread. The oil cartridge screws onto a rechargeable battery. Some of these batteries have buttons and some heat up automatically when you pump the oil. Some batteries have multiple temperature settings and some heat up to a preset temperature; these features need to be known in advance.
·Cleaning notes
Use a suitable cleaning tool to clean them, such as activated charcoal or dried tea leaves in a used pipe to absorb the oil. It is important not to use alcohol or other boiling water to clean the pipe, and to wait until it has cooled down completely before cleaning. Otherwise, the hot stem will come into contact with the watery liquid and cause the mouthpiece tenon to loosen, thus shortening the life of the 510 cartridges.
This is what you should be aware of when using 510 cartridges. At the same time, when using 510 cartridges, there are still some vaping tips, for example, when using them, be careful not to suck too hard, will not produce smoke. When you inhale too hard, the smoke is sucked directly into your mouth and not atomized by the atomizer, so gently inhaling is more powerful and gives you a better vaping experience.
510 Cartridge Oem,Leakproof 510 Cartridge,510 Battery And Cartridges Oem,510 Cartridge
Shenzhen MASON VAP Technology Co., Ltd. , https://www.e-cigarettefactory.com