How Deep Learning is Helping Society

Deep Learning has allowed many organizations to tap into a sea of information and neural networks to improve data services and analysis .

As of today, it’s being utilized in image recognition, natural language processing and predictive analysis. We’ve seen how scientists are using deep learning technologies to help a computer tell a story through the images it sees. It can help to detect how a particular word is used through contextual analysis .

While Big Data Analytics and Deep Learning have allowed companies to improve their customer relations management systems and improve internal systems, it can do more.

Deep Learning is helping communities become safer and healthier.

Deep Learning helps in crime prevention

Big Data and Deep Learning is helping in crime prevention and the the City of Chicago and San Francisco  are prime examples of it.

The city governments provided information on transportation, building maintenance records, utility usage and emergency response dispatches. By providing access to these data sets, data analysis providers were able to create visualizations of the city’s crime patterns. With the help of deep learning and machine learning platforms, they were able to pinpoint crime spots within the cities of San Francisco and Chicago. Both cities found out that in spite of the cold, most crimes happened during the winter season. They also discovered that most crimes occurred at midnight, noon and 6PM.

They were also able to geomap possible locations where the police force could enact arrests effectively and pinpoint areas where they needed help the most in making arrests. It was so effective that they reported double-digit drops in crime rates. In the future, we’ll likely to see more countries adopt Big Data Analytics, Data Visualization and Deep Learning technologies to help their crime prevention efforts.

Deep Learning makes medical detection and diagnosis more accurate

Deep Learning has applications for medicine as well. Hospitals worldwide are using deep learning platforms to go through large volumes of medical images, medical records and histories, and other pertinent data.  The goal is to help doctors  make better, faster and more accurate medical diagnoses.

Enlitic, a deep learning healthcare company, is one of the few that is making it possible. They used their algorithm lung CT scans to detect potentially cancerous growths. To prove how accurate their deep learning-aided diagnosis is, they compared their results to a panel of four of the world’s top human radiologists.  The results were not at all surprising. The Enlitic’s platform reported no false negatives while the human radiologists had a 7% false negative rate. Their rate at incorrectly diagnosing cancer was at a staggering 66%. Enlitic’s false positive rate was at 47%. It’s still pretty high but it is noticeably lower than the human radiologists.

That’s one of the reasons this healthcare company managed to raise US$10 million to enhance their deep learning imaging systems. We’ll soon see better diagnosis accuracy and and it’s quite possible that in the very near future, patient diagnosis can be made without a doctor.

Deep Learning helps mitigate risks

lloopp™, a consulting and managed services company, has a Philippine office and it’s been helping telecommunications and media companies, bank institutions, and fast food corporations understand their customers and serve them better. But aside from that, lloopp™ has been supporting its Philippine data scientists, deep learning experts and specialists to come up with innovative new ways to help the less fortunate communities.

lloopp™ currently has two deep learning projects that they have developed.

One is Project #Agos, a collaborative platform that uses mobile and web technologies, deep learning and natural language processing systems to help local government agencies and units and first responder teams, react faster to medical emergencies, typhoons, earthquake and other disasters.

lloopp partnered with Rappler, one of the most recognized social media news sites in the Philippines, to help them coordinate with LGUs and emergency teams. #Agos taps into twitter streams and singles out mentions and semantic use of emergency, fires and other incidents. Once the system detects one, it will send a notification alert to Rappler and through the platform, they will inform agencies either through SMS or social media networks to take appropriate action.

Deep Learning helps Remote Communities stop Diseases

And there’s more. lloopp Philippine data scientists are currently developing a medical diagnostic tool to help prevent the spread of intestinal and stool parasites in remote communities. While it’s not a advanced as the Enlitic system, lloopp has developed an image recognition tool that helps detect parasitic eggs in stools. What’s great about it is that it can work with mobile phone cameras. A person uses their mobile phone camera to take pictures of possibly infected stools. Users can then send the images to  their system. lloopp has developed an  image recognition tool equipped with deep learning algorithms  to help detect presence of parasitic eggs in the stool images sent. Once it’s detected, the information will be relayed to LGUs and health agencies to stop the spread.

According to lloopp data scientists, this will benefit remote communities who don’t have immediate access to medical laboratories and hospitals. The system will also relay possible treatments for treating the infection through natural remedies available in the area, cleaning and hygiene recommendations.

In spite of the availability of deep learning technologies and computing power, there are still challenges that need to be overcome. One of them is access to more information and data. Not all government agencies and corporations are willing to share the information readily. If analytics and deep learning companies have wider and better access to information, they’ll be able to make better and more accurate analyses and provide even better insights and applications. There’s also the matter of converting the data sets for use in data visualizations. Some data sets are still locked in archaic excel sheets, simple text, and even paper documents.

But the development of deep learning and its applications will get more varied and useful. The future is looking bright all thanks to big data and deep learning.

Related Posts
The Who’s Who of Deep Learning – Part Two