1.Understanding of the medical industry
Nature there is no difference between artificial intelligence and the Internet, are all tools, these tools for business is infrastructure, like the water in the life of coal, so any “Internet +” or “+” AI cannot leave the understanding of the industry, in terms of the financial industry, such as, trust, private equity, fund, stock, mortgages, business loans how these concepts are related business; Liquidation, risk control, collection, credit investigation and how each link is reflected in the product, these need to have a certain understanding of the industry;
The health care industry, classification, remote consultation, medical diagnosis and treatment, health intervention, gene sequencing is what meaning, medical institutions, doctors, drug companies, pharmacies, insurance, etc. How to understand the mapping relations between the multiple, these all need from the perspective of industry have a holistic cognition, even the study of positive policy, technology and capital.
2.Excellent active learning ability
Is new in the face of new areas, the initial stage, if there is difference, it must be slowly pulled open, due to the different learning ability by learning can improve a person’s cognitive level, like Andrew Ng said that, in fact, in many cases we don’t know what to do next, at this time will need a lot of learning and reading, Talk to experts in your field.
I have a very deep experience, every time I listen to, tongji fudan public class, the professor often has a rich imagination, finally get the feeling, and these public class theory will establish relationships in work and life, so when we read the books, listen to enough enough public class or when talking with enough experts, our brain will receive enough input information, New ideas will naturally follow.
3.Basic data analysis ability
AI field there are quite a proportion of user demand from large data analysis itself, relying on the deep learning for user’s psychology and behavior, like Tmall elves why more with more understand you, explaining the user behavior data path, so the data analysis ability in the process of problem solving, in the process of hands-on practice and exploration, will be of great use.
We need to do is before, during and after the product launch analysis using data mining using scene, find product improvement point, breakthrough and even a tipping point, with rich interaction scenarios encourage innovation in the AI, in terms of data analysis ability itself, we need data from the cognition, the collection, integration and expression, to develop ability of five aspects, such as comprehensive promotion, This is also why Ren Zong always calls on the president to add Statistics and Discrete Mathematics to the basic subjects when he visits universities. The ability of data analysis should have a certain foundation in the university.
4.Some technical knowledge is required
Often see BBS argument product managers want to understand technology, I see is concerned, the era of mobile product manager would draw prototype, logical good, communication can also seems can do, but in the era of AI and big data, don’t understand the technical product manager may have some passive, in AI emerging stage, there will be a phenomenon of technical ability is greater than the ability of the products, This is why the demand for algorithm engineers is far greater than the demand for products.
In the Internet era of 20 big guys, 15 of them have technical and coding background, big guy, especially with you, especially in the AI era, so having technical background is a big advantage to be an AI PM.
What are ai medical products like?
Skin treatment solution: AI+ map recognition + medical science + prescription diagnosis
Identify user’s lesion map through AI, and provide medical services such as disease knowledge popularization, diagnosis and counseling, medication prescription, etc. Several problems solved by AI in the field of skin medicine:
1，Solve the problem of scarce dermatologist resources
20000 + the dermatologist, but nearly 160 million patients with skin disease, so a dermatologist resources shortage, peak, a dermatologist can accepts more than 60 + offline clinic patients a day, dermatologists also has the superiority of natural, experienced doctor can be directly through the diagnosis to upload lesions in patients with atlas, “Internet + medical treatment” can alleviate the problem of resource mismatch among regional dermatologists, but it is still difficult to alleviate the problem of shortage of doctors’ resources. In the era of “AI+ medical treatment”, skin disease types can be identified through lesion maps and diagnostic reports can be output to patients.
2，Solve the problem of mechanized diagnosis of repeated diseases by doctors
Acne skin disease relapse easily, so the return rate is also high, as for dermatology online visits, some doctors diagnose day same disease may be as many as dozens of cases, we traditionally, it is templated Top100 diseases and diagnosis platform database content packaging, doctors identify patients after upload lesions map, give the template to quickly solve, It will bring some problems, such as a doctor too dependent on template cause diagnosis problem of homogeneity, the content of the template itself hard problems, for example, in the long term user expectations also reduces, but the AI can learn by depth according to patients’ rapid repetition of different diagnosis of diseases is different from person to person, the greater the data, the accuracy is higher, until far more than the doctor, The day is drawing nearer.
3，Solve the misdiagnosis rate of doctors with insufficient clinical experience
Covered with skin diseases is tens of thousands of, this kind of skin diseases such as dermatitis, its subclasses diseases have hundreds of, these dermatitis segment disease clinical manifestation and map sample are different, in the face of such a large diseases, rarely a dermatologist across departments in-depth diagnosis of many skin diseases, and some diseases have a certain similarity, Inexperienced dermatologist to distinguish these map also has the certain difficulty and misjudgment, AI can be combined with clinical mapping large data, data feed, through the deep learning condition of accurate positioning and carry on comprehensive judging benign and malignancy, rapid given careful diagnosis of different diseases, thus improve the doctor diagnosis efficiency and accuracy.
4，Solve the problem that the patient does not know the skin disease
Dermatologist teaching time is not long, most of the limitations in the outpatient service, after leaving the hospital, probably didn’t know what the situation, forget the doctor’s advice and cure, also does not have a detailed understanding of his condition, common skin disorder was fine, the key is a lot of diseases are very rare, such as “face disseminated lupus of su li sex”, Tell people what she had skin has the difficulty, not to mention its clinical manifestation and the matters needing attention, the AI can through integration of the map data, in-depth study, not only can let users understand the symptoms of the disease, etiology, diagnosis, treatment, and can be real-time matching with the most similar rehabilitation cases, the patient At the same time, according to the patient’s situation, the latest, suitable for the patient’s attention and treatment plan, etc.
5，Solve the problem of intelligent case management
At present, hospitals rely on medical records or data companies to organize medical records. The medical scientific research of the department and the extraction of medical record feature information are also completed manually. A lot of manpower and money are needed, and accuracy cannot be guaranteed. Through deep study of artificial intelligence technology can realize the intelligent medical record management, automatic feature extraction of medical record information, the data of the hospital, not only do standardized processing speed but also for scientific research, to do it and find the shadow of tencent, well-informed, such as cloud of medical clinical data, outputting structured electronic medical records, make precipitation data real help to the doctor, Just like the example of intelligent case management of Tencent Miying below.
Skin care solution: AI+ Map recognition + accurate skin care solution
Connection precision skin care services, and even accurate medicine makeup service, and the reality of the traditional intelligent skin measurement may be merely a process, from the photograph measuring skin, intelligent analysis, skin care science to product recommendations, if there is no professional phase measuring skin to the deep study on the level of the skin texture, it is difficult to let users pay, only the surface of these parameters must be not enough, For example, skin pigmentation, sensitivity, tolerance, etc., all require AI deep learning to give feedback.
In fact, there is much more to facial feature data than that, and when an authoritative data report can conquer the user and achieve a truly high level of accuracy, the customer’s expectation of paying will be met, and a business model will be born.