This article is part 2 of 3 in the series MachineLearning

Introduction

Intelligent machines (don’t get it literal) are everywhere. Machine learning has developed, somewhat maturely, pretty fast in the last two decades. The applications of Machine Learning are virtually everywhere. From computers playing and winning in highly competitive games to self-driving cars to surveillance to drones and most of all in medical science, financial sector, security sector, we see a rising trend of Machine Learning applicability.

“My personal challenge for 2016 is to build a simple AI — like Jarvis from Iron Man — to help run my home and help me with work.

Artificial intelligence may seem like something out of science fiction, but most of us already use tools and services every day that relies on AI. When you do a voice search on your phone, put a check into an ATM, or use a fitness tracker to count your steps, you’re using basic forms of pattern recognition and artificial intelligence. More sophisticated AI systems can already diagnose diseases, drive cars and search the skies for planets better than people. This is why AI is such an exciting field — it opens up so many new possibilities for enhancing humanity’s capabilities.”

Mark Zuckerberg, Facebook Founder, wrote in a Facebook post on January 17, 2016.

Machine Learning has grown so much so that it has embedded itself into another form called Deep Learning. Deep Learning focuses on helping machines identify patterns that human being would otherwise be oblivious to. Deep Learning in combination with Machine Learning has its tentacles in almost all the fields e.g. in IoT (Internet of Things).

In this episode of the series of articles on Machine Learning, let us look at some amazing applications of Machine Learning for our benefit.

Applications in Medicine and Medical Science

The benefits of Machine Learning in Medical Science are massive. From early detection of developing disease to the 3D printing of human tissues, Machine Learning has played a big role in helping cure the pain of mankind or making it better.

Diagnosis and Early Disease Detection

Companies like Google’s DeepMind Health and IBM are using Machine Learning for early detection of diseases including oncology research. Research is underway for early detection in order to stop a disease from spreading. Cancer is one of the biggest enemies of human kind, Biopharma Company Berg is using Machine Learning to research and develop diagnostics and therapeutic treatments.

Future research including curing of terminal diseases like Cancer and AIDS as well as macular degeneration like in aging eyes using Machine Learning.

Many companies are using Open or Closed data to do early detection of Dementia from Clinical Data. Dementia is the most prevalent degenerative disease, and its early detection has been a long standing issue. This way the quality of life gets an increase in many patients (which is one of the supreme objectives of any technological research and implementation). Since life expectancy has increased, due to which we have seen a rise in the population of seniors. It is of paramount importance that early detection can be done, so as to cater for the increased senior population.

Machine Learning is also being used in the diagnosis of biotic stress in plants. This is required for precision crop protection. Using this one can predict the plant diseases or weeds at an early stage.

Radiotherapy

Dr. Ziad Obermeyer, an assistant professor at Harvard Medical School says that in about 20 years, there will be no more humans working as radiologists. Rather there will be a group of cyborgs working the job, detecting, helping and operating the radiology equipment and reading thousands of reports by the minute. Trials are already underway for radiotherapy programs for early detection and prevention of diseases.

Research and Clinical Trials

Clinical research Trials have improved massively thanks to Machine Learning and smart algorithms in sorting population based on genetics to find the candidates for clinical trials. Research is underway to support this with data sorting algorithms and it is getting better and smarter every day.

Automated Electronic Record Keeping

Gone are the days when it used to take hours for doctors, nurses and clerks to file papers and papers of data of patients. Searching, storing and then rehashing the patients’ files used to be a mammoth task. Nowadays, it takes seconds to do that. These digitized records have becomes the datasets for machine learning consumption.

Machine Learning has furthered this cause by using optical character recognition since long to translate handwritten notes by the doctors into electronic records. Facial recognition and biometrics are another way to save patients’ history and data into a safe and secure storage format. Further to this, there are now standards to create common data interchange format which can be used with wide variety of algorithms in machine learning.

Automated bots are being trialed which can be used to reply patient queries and in turn sorting through requests at a much higher rate.

Predictions of Epidemic Outbreak

It has become easy to predict the outbreak of common diseases like malaria, typhoid etc. thanks to Machine Learning and AI. The algorithms use weather conditions, atmosphere density, water samples and other factors to predict such outbreak thus helping mankind to fortify themselves against such diseases.

Applications in Finance

In the wild highly immersive field of finance, machine learning has started to play an important role. More and more financial institutions are either using their own data to understand it or they are opening it to the world using open APIs for others to analyze. In this space, most of the innovation is going on in Fintech sector but banks and other traditional corporations are also catching up. We have cover one aspect in the past e.g. how blockchains can revamp some of the markets.

Online Trading

The simplest yet most powerful application of ML is in online stock trading where at one time, currency trading was done painfully. Now, ML based applications are making predictions and forecasting in the bullish or bearish trends of the stock market. Paypal, Mobile Financial Services, and Bitcoin are all making use of ML.

Online forex trading and portfolio management are impacted as well. In the space of currency exchange prediction and security price pattern prediction are now been tackled through machine learning algorithms. Previously we were only able to predict using few intervals in the recent prices but now using machine learning we can easily do the same on millions of intervals of prices changes.

Machine Learning along with Natural Language Processing is being used to analyze a multitude of financial news articles to get insights. Then these insights are being used as an additional vector in the stock predictions.

Blockchains & Automation

So far it has been very difficult to connect Blockchains with Machine Learning. But it is now getting possible. For example, a company named LendingRobot has launched a hedge fund that uses blockchain to automate the returns on investment using Machine Learning. It is one of the dimensions of robo-advisor strategy that has been prepping for the last 5 years. As the information in the investment space gets updated at a high rate, it is best that Machine Learning can be used because Machine Learning gets maturity as more and more information gets in.

Smart Contracts can be issues through Blockchains (or other variants). The frequency of issuing these contracts will increase as the transactions get increased. But which contract should land where, and how, will be determined by Machine Learning. This space is still active in development (or in the startups’ corridors).

Applications in Security

There are hundreds of applications of ML in security. For example, ML applications have grown exponentially in helping curb terrorism in parts of the world. Let’s discuss some.

Facial Recognition & Crime Prediction

Thousands of criminals have been caught using this technology. Facial recognition capability has grown so much lately that a man walking through the street, disguised in a different attire can be stripped down facially to show his real faced based on the distance between various features on his face (sorry for being a bit technical). FBI has a huge database of criminals and potential criminals and can narrow down to the correct ID even with a highly blurred image of a CCTV in a locality or street. This huge record coupled with the diverse records on the citizens, it is now possible to predict many things. E.g. in some cities police can predict whether some crime is going to happen at a certain place or not.

Fraud Detection

Paypal is using ML for fraud detection such as money laundering using an algorithm that compares millions of transactions to identify the fraudulent transaction. Other intelligent banking software is also helping detect monetary frauds. It is a very difficult problem to solve. It would become increasingly difficult when Blockchains will be implemented at scale. So machine learning algorithms have to further evolve in this space.

Conclusion

Although we have got some very powerful algorithms in Machine Learning, we are still in the beginning of getting the full potential from this approach. Some of the industries that it has touched are still susceptible to false positives. To bring the quality to the analytics, it is important that this path is tread cautiously. There are other blazingly hot applications as well, but to cover them we would need to write a book. New innovations are coming in all sectors. Most of the innovations now touch some aspects of Machine Learning in some way. New markets will disrupt old markets because most of the human element is now getting out of the way.

Series Navigation<< Part 1 – Machine Learning – A History that Every ‘Being’ Should KnowPart 3 – Computational Power, Deep Learning & Machine Learning >>