HYDERABAD: While several startups are offering many artificial intelligence (AI) solutions to the Telangana government, the latter is declining their offers because of the unavailability of big data, which is a must for these solutions to work.
Detecting COVID-19 using AI-powered X-rays was one of the solutions proposed by around three companies amid a delay in test results and shortage of testing kits, an IT Department official said. "Using AI and machine learning algorithms, these X-rays can identify the virus. The companies showcased their technology through videos," the official said. However, the availability of big data turned out to be a hindrance.
That is because the technology requires a huge chunk of data set for the AI to train itself to detect the cases. This is how it works: An X-ray usually shows formations in lungs, which can be associated with a larger number of respiratory problems such as pneumonia.
In COVID-19 patients, experts said, this formation is a hazy opacity that does not obscure underlying bronchial structures or pulmonary vessels. The conundrum lies in the fact that these anomalies are not exclusive to COVID-19 patients.
Thus, there is a need for big data, which will train the AI to fine-tune and detect Covid-19 patients correctly. As of now, different researchers have experimented and come up with results varying from 70 per cent to 90 per cent accuracy. However, they all claim in unison that their results would improve if they have access to larger data sets.
When asked regarding the availability of big data for such exercises, another senior official in the IT Department said, "We do not have such big data sets. We have told all the companies, who are offering AI-based solutions, that it would not be possible. We have also made it clear that they can pursue it for academic purposes, but cannot be put in practice."
Recently, Express reported that a few US-based companies used the State’s health data to predict containment areas and the spread of the infection. However, the data that was used was anonymised — without identification details. 237 new cases were recorded in Telangana on Sunday. The State’s Covid-19 tally has climbed to 4,974
Three new COVID-19 casualties in State
Telangana recorded 237 new cases of COVID-19 on Sunday. While three more patients died of the disease, 25 were discharged. As on Sunday, the total number of Covid-19 cases in Telangana climbed to 4,974, of which 2,377 are discharged cases, 185 are deaths and 2,412 active cases. Of the 237 new cases, 195 were recorded in the GHMC limits and some of the others in Medchal (10), Rangareddy (8), Sangareddy (5), and Mancherial (3) districts
Artificial Intelligence in COVID-19 times
Detecting Covid-19 using AI-powered X-rays one of the solutions proposed by 3 companies
Solution comes handy amid delay in test results, shortage of testing kits
Using AI and machine learning algorithms, the X-rays can identify the virus
How it works
An X-ray shows formations in lungs, which can be associated with larger number of respiratory problems such as pneumonia
In Covid-19 patients, this formation is a hazy opacity that does not obscure underlying bronchial structures/pulmonary vessels
The conundrum lies in the fact that these anomalies are not exclusive to Covid-19 patients
Thus, there is a need for big data, which will train the AI to fine-tune and detect Covid-19 patients correctly