Hello everybody,
currently topic 1 is coronavirus, so the article proposed by me today will be about coronavirus. More specifically, it's about how AI helps fight a pandemic. Read the article and share your opinion. Below are the article and help questions.
currently topic 1 is coronavirus, so the article proposed by me today will be about coronavirus. More specifically, it's about how AI helps fight a pandemic. Read the article and share your opinion. Below are the article and help questions.
https://fortune.com/2020/03/16/ai-coronavirus-health-technology-pandemic-prediction
Help questions:
- How does AI help in the fight against disease?
- How does AI deal with changing data in real time?
- Is AI able to predict flu incidence fairly effectively?
1. How does AI help in the fight against disease?
ReplyDeleteThe author of this article himself stated that it's only "dry" data analysis. Data to which hardly anyone attaches a sentence. See: sluggishness of the Chinese government, no tests in the US. I don't know if this disease can be compared to any earlier disease. Maybe with the plague 1347 - 1353. See, she lasted six years! Do you think the current one will end until Easter? We must get used to it and learn to live with it!
Back to the question. How has AI contributed? Apart from those who collect and process this data, it's probably nothing. Because no one else has access to them. Even if they make it public, not everyone will benefit from it. It was just like the Germans didn't believe that the enigma had been broken during World War II!
2. How does AI deal with changing data in real time?
Probably good. You can see that he can draw some conclusions or suggest something. Only a long way ahead of us. Let us remember that, however, the basic inference methods today are thresholding or probability theory. The rest remains in a strong testing phase.
3. Is AI able to predict flu incidence fairly effectively?
As I wrote earlier. I think so, the question remains who is able to believe in these predictions and to what extent. Will our or other government invest in 40 million vaccines, unknown as mutated influenza based on AI predictions? I dare to doubt!
Thank you for your answer. I agree with you regarding access to data, as they are unlikely to be made public. In addition, AI prediction about the area of influenza can be helpful in fighting the virus.
Delete1. How does AI help in the fight against disease?
ReplyDeleteIn this article we can see that AI is used to forecasting the outbreak of a disease. After analyzing many information from the Internet researchers try with AI detect potentially risk for people. A big role plays startups. They built systems to early warnings about outbreaks and forecasting the direction of the disease's spread.
2. How does AI deal with changing data in real time?
I think its depends on input data. When we have a good data then AI will react correctly on changing data in realtime. Researchers need to monitoring situation and providing new data to help AI choose correct decision. A wrong data have a big impact on AI. A good example presented in this article is google. They tried to forecast the severity of seasonal flu outbreaks and they failed.
3. Is AI able to predict flu incidence fairly effectively?
I think it is possible but it depends on good input data that we will used to train our model. One startup presented in article showed that this is possible. They used smart thermometers that work with an app to help people decide when to see a doctor. The results are very good. We will see what future will bring.
Thank you for your answer. The selection of appropriate input data is actually crucial as to the correctness of the model's operation. An example with intelligent thermometers is very interesting.
Delete1. How does AI help in the fight against disease?
ReplyDeleteAfter reading the article I agree with the commentators above. AI can help in predicting the outbreak based on statistics, but it seems to me that the benefit will be small. Even if AI provides optimal data, this is unlikely to be used.
2. How does AI deal with changing data in real time?
It all depends on the quality of the data on which the AI will be based. The very quality of predictions is based on probability theory, I am skeptical about this at the moment
3. Is AI able to predict flu incidence fairly effectively?
As the commentator above set an example, the Google company tried but failed. Therefore, I think that in theory it can, but there is no certainty that this will turn out, there are too many factors on which forecasting depends
Thank you for your answer. The performance of artificial intelligence-based models will always depend on the input. Weather forecasting is also based on the huge amount of data that the computer needs to convert. Real artificial intelligence has not yet been discovered.
Delete1. How does AI help in the fight against disease?
ReplyDeleteArtificial intelligence can detect the first symptoms, indicating that new outbreaks are developing in the region. Clinical and historical data on other viruses are analyzed, as well as information from hospitals, airports and other public places such as schools, where the number of absent children is checked to assess the speed at which the virus is spreading. Artificial intelligence and analytics are used at every stage of vaccine development, production and distribution. By analysing data from clinical trials, it is possible to determine whether they are safe and meet specific medical standards. Artificial intelligence also makes it possible to monitor the occurrence of undesired side effects.
2. How does AI deal with changing data in real time?
The best way to add new cases and teach the network. And a second way to use self-coders. A self-coder is a neural network, which is taught so that the output shows the values received at the input. The self-coder learns to compress the data from the input layer into a short code, and then unzip the code into something that is closely matched to the original data. This forces the auto-encoder to engage in dimensional reduction, for example by learning to ignore noise. Some architectures use stacked diluted autoencoder layers to recognize images. This is the case here. But the most important thing is to add new data that is validated by an expert.
3. Is AI able to predict flu incidence fairly effectively?
Alibaba, who developed the model, claims that the new algorithm they developed can detect coronovirus and is 96% reliable. This is possible thanks to SI scans of patients' chest. The model was learned from 5000 confirmed cases of this virus. A study of its effectiveness was carried out in all hospitals in China, and more than 100 facilities decided to implement the AI Alibaba method. Fortunately, not only this company was involved in the development of this health problem, but also its competitors. They also base their research on artificial intelligence cognitive methods and machine learning.
Thank you for such a comprehensive answer. I agree with you regarding the selection of appropriate input data checked by an expert. Currently, we have a lot of data and they are not always verified, which affects the operation of the model.
Delete1.How does AI help in the fight against disease?
ReplyDeleteDue to the scale of corona virus epidemic, amount of data related to it that can by use to fight with it is too big to be analyzed by human. So this is perfect field for big data tools to by used. Thera are many different approaches being checked for example the combined analysis of personal, clinical, travel and social data including family history and lifestyle habits obtained from sources like social media would enable more accurate and precise predictions of individual risk profiles and healthcare results.
Beside predictions, AI has leveraged its natural language processing capabilities to build a multi-lingual virtual healthcare agent that can answer questions related to COVID-19, provide reliable information and clear guidelines, recommend protection measures, check and monitor symptoms, and advise individuals whether they need hospital screening or self-isolation at their homes.
2.How does AI deal with changing data in real time?
There are many ways to update data recurrently, of course assuming that data are available online. An API can be used to update the data with required frequency for example every 5 minutes or 30 s depend on what is needed. If API is not an option there is a shortcut to for example scrape the content from www directly.
3.Is AI able to predict flu incidence fairly effectively?
It will depend on goodness of the data set that will be used, additionally the confirmation of which parameters are crucial for the infection. Consider if the dozens of parameters like age, sex, ethnicity, mask type, mask use, sanitizer frequency use, transit mode, crowd interactions impact the results. If we will now answers for this questions, and train model on confirmed data I think we will be able to predict the possibility of infection with high certainty.
Thank you very much for your answer. Of course, the selection of data is crucial for the operation of artificial intelligence. If it was possible to obtain quite a lot of detailed data about a specific person, individual profiles could be created for patients or people who need specific medical support.
DeleteAI can be used for a forecasting of a spreading, some models can be used for a diagnosis as there are many examples of AI systems differentiating between pneumonia and covid based on RTG images (unfortunately due to the lack of medical data most of the examples are merely demos that cannot by applied in practice). Me and my colleague we've created during a hackathon an AI system intended to be used for monitoring if people respect safe distance rule in public, it even was mentioned in media :) https://www.sztucznainteligencja.org.pl/inteligentny-monitoring-w-gdyni-wykryje-tlum/
ReplyDeleteRegarding the realtime data, in the forecasting systems current data should be a problem as models rarely would need them. If there was a model predicting outbreaks or number of infected people in a let's say 5 day window than adding newer data to it could be seen simply as updating the model to always have predictions for the next 5 days everyday.
I haven't seen any evaluations when it comes to the flu forecasting. It seems such models are being used so at least there should be models that are accurate enough so someone uses them.
Thank you very much for your answer. I am very glad that you managed to create such a support system that is very much needed at the present time. I agree with you about the selection of ready and tested models that can be successfully used at least partly to suit your needs.
Delete1. How does AI help in the fight against disease?
ReplyDeleteAI can help locate epidemic sources earlier, as well as model its spread. AI algorithms can also be used in more complex simulations of virus behavior in the human body.
2. How does AI deal with changing data in real time?
AI was created to change and teach new behavior based on new cases. I think that if the data entered is correct, the more of them will be, the AI will be able to react faster.
3. Is AI able to predict flu incidence fairly effectively?
At this point, the attempts made do not yet give the results we would expect. However, I think that we should not give up and that sooner or later we will manage to model such events correctly.
Thank you for your answer. I agree with you about perseverance in pursuing your goal. I think that it is worth trying and improving the operation of artificial intelligence.
Delete1. AI is able to predict new outbreaks based on collected and analyzed data. AI BlueDot predicted epidemics well before the outbreak of the largest number of cases by analyzing information from social media, articles, etc. AI can analyze government data, from hospitals, the number of cases, the direction of the spread of viruses comparing it with the data of previous epidemics. Thanks to this, it can predict which regions should be quarantined to minimize the risk of large numbers of infected. Artificial intelligence is also used to simulate the effects of vaccines.
ReplyDelete2. Adding data in real time is a difficult and time consuming process. Much depends on the quality of the data, when it comes to coronavirus we know that many people have mild symptoms and are not even tested. It is difficult to base the updated set on data only from positive test results, because the overall picture of the spread of the epidemic may be skewered.
3. I think so. As it was already mentioned, there are systems that have very good results in the diagnosis of infectious diseases on the basis of specialized tests, such as RTG. As far as I know, white spots appear on the x-ray of the lungs of coronavirus patients. Artificial intelligence can analyze x-rays to determine if someone is sick or not. However, you need to expand the data set with other parameters such as the environment in which the person was, whether they could have contact with the infected, or traveled to places where the virus occurred. If we take into account many factors and properly collect data, I think that artificial intelligence would be able to predict the occurrence of subsequent cases with high probability.
Thank you for your answer. I agree with you regarding the choice of data, because as you have noticed, people infected with coronavirus and passing it gently are not always taken into account, which changes the algorithm's operation and causes inconsistent results. In addition, updating data in real time requires the use of BigData tools for the selection and pre-selection of such data.
Delete1. How does AI help in the fight against disease?
ReplyDeleteIn article AI is used to forecasting the outbreak of a disease. Based on statistics this algorithm trying to precitc in which area the disease will start. That knowledge will be very useful if this algorithm has a big accuracy. Silimar articles show that this algorithm works but not very well (yet).
2. How does AI deal with changing data in real time?
It depends on framework/topology which you use. Some need time to precess and them you have to build module which will select which data are more important and only process this data by your algorithm.
3. Is AI able to predict flu incidence fairly effectively?
I didn't see any algorithm which will predict similar information with big accuracy. Moreover if algorithm works perfectly and will predict that in Warsaw flu incidence will be one hundred a day then it's quite hard to check this information. Not all people go with flu to doctors. Some of them (if they works from home, feel not so bad) wouldn't come to doctor and then the statistics will be much lower than predicted value.
Thank you for your answer. I agree with you about the forecast data. It is difficult to build such an accurate system that predicts the spread of the virus. The basic problem is human behavior, because not everyone who is infected feels bad enough not to do anything. Some people don't even suspect they have a virus and yet they feel very good.
DeleteFor some reason the article became pay-walled for me, but short javascript snippet revealed it all.
ReplyDelete1 How does AI help in the fight against disease?
They write about BlueDot startup, that supposedly predicted the outbreak. But then, also about HealthMap that was faster, and its representative states that intel gathered among physicians in Wuhan was also very important.
2 How does AI deal with changing data in real time?
Some propose systems that ignore social-media chatter and go straight into direct syphoning of medical data from devices into AI systems
3 Is AI able to predict flu incidence fairly effectively?
Looks like it can, buy have to stay quiet about it, because it's too easy to trigger positive feedback loops when people start to look up what's going on.
Thank you for your answer. I do not know why suddenly paid access to the content of the article. When I shared it, everything could be read without any payment. Artificial intelligence requires good data, and unfortunately social media cannot always provide reliable data. It would be worth building a system of initial data verification consisting for example in checking the amount of the same information with the acceptance of the threshold accepting the data as reliable and unreliable.
ReplyDelete1. How does AI help in the fight against disease?
ReplyDeleteNowadays, artificial intelligence (AI) is becoming the essential part of our life. AI can help to fight coronavirus as well. AI can help with analyzing huge amount of data which is very complicated and time-consuming for human-being. There are millions gigabytes of data which are produced every day. We can hear about how many new cases are or how many people are recovered in the world, Poland or even in our voivodeship every hour. There are a lot of data not only dedicated to new or recovered cases but also, for example, the data about the travelling, buying products or unemployment. A lot of people ask when the pandemic disappear. Artificial intelligence can try to answer this questions. People are not able to analyze such big data without AI.
2. How does AI deal with changing data in real time?
Artificial Intelligence needs data to create models to predict or estimate something. Not only AI has a problem to integrate the data in a real time, it is the problem which have all the devices working in real time with data. There are programs which allow to gets data from other programs or sites and synchronize then within defined time. The data are the basis of AI operation.
3. Is AI able to predict flu incidence fairly effectively?
I think that AI is able to predict flu with high accuracy. It depends how accurate model is so if the model is created on a basis of parameters which described a flu - why not.
1. How does AI help in the fight against disease?
ReplyDeleteOh, this is an extensive question, but in the case of this disease, this model can detect the first symptoms. It uses that for an indication that new outbreaks are developing in the region, thus it can raise some red flags.
2. How does AI deal with changing data in real-time?
I think if you "feed" the model regularly with data, it should deal with a good-enough accuracy.
3. Is AI able to predict flu incidence fairly effectively?
As mentioned, there was a project Google Flu Trends (GFT), which was a web service operated by Google. Yet, it failed because it consistently overestimated flu prevalence. I think that AI could be an excellent supporting tool while used with additional methods and devices.
I can’t read the article becouse it is writen as paid content.
ReplyDeletehttps://zapodaj.net/e8f96bac4aa04.png.html
DeleteThis comment has been removed by the author.
ReplyDelete