Artificial Intelligence Systems to prevent epidemics

With the purpose to overcome the gap in between the public health system and the technological innovation, Rainier Mallol has managed to create an Artificial Intelligence based on Big data capable of predicting where the next outbreaks of fatal diseases arose. Discover below his career as a social innovator.

Rainier, you have created AIME, a platform that uses Artificial Intelligence to predict where the next outbreaks of dengue, Zika and Chikungunya will emerge. Where does the idea come from and how does it work?

I am from the Dominican Republic, I was born, and I was educated in Santo Domingo. The reality of my country is very similar to others in the Latin American region: citizens are constantly threatened by governments full of infectious and ordinary corruption, as well as private organizations that monopolize markets and block competition. This type of non-democratic institutions result in a massive uncertainty in the social classes, stripping it of rights that the obvious state, and of an unequal market.

Due to this situation, we not only have a huge rate of unemployment, a bad and deficient electrical system, ruled by the private sector, poor transport systems, and above all, a health system without sufficient funds to shelter and protect the health of our cities. Since I was a child, I was not so affected by the lack of this health system, but I grew up with countless cases and stories of close relatives and friends who fell victims of these failures. Some went to hospitals where there were no medical resources or water, others contracted infectious diseases such as dengue, and others went abroad to receive appropriate medical treatment.

My family always instilled in my medical knowledge. My mother, who was raised because of a deceased father, is a dermatologist. I grew up seeing slides of medical cases, bruises, burns, sunken nails and seborrheic dermatitis. However, it is worth mentioning that, in the Dominican Republic, being a doctor means having a vocation: unlike other countries, doctors in my country are severely affected by our health system, since they receive a salary that is not equivalent to their studies, efforts and works. Reason why, my mother always exhorted me not to study medicine, and so it was.

Actually, I am conscious of the reality of my country. It hurts me to see other Dominicans living in the conditions we have because of the greed of our leaders and businessmen. I suffer when I see statistics of children dying from illnesses, and I hate the consequences, even in our culture, of so many years under rulers who are not leaders. Just two months before finishing my degree in Telematics Engineering at the PUCMM, 4 other students and I created a technology consulting company. In addition to having clients for profit, we also had clients for social impact purposes, where we contributed our knowledge to social entrepreneurs who did not have technological skills. Our idea was, at that time, to contribute by doing what we did best: develop technologies.

With my willing for entrepreneur, with my desire to make a social change and with my knowledge in the area of computing, I applied to NASA, specifically to Singularity University (SU) in 2014, for the class of 2015. In SU, which sought to gather 80 people from around the world to show them state-of-the-art technologies along with major social problems, I met the person who is now my founding partner, a doctor and PhD in Public Health of Malaysia called Dhesi Baha Raja.

On the first day of classes, Dhesi asked me about which disease is common in my country, and I told him that Dengue (just arriving at NASA I received news of a cousin who contracted it), and explained the causes of dengue it still exists today and in many LATAM and Southeast Asian nations: when Public Health officials do not know where or when the next outbreak of dengue will occur, they make decisions without sufficient bases, investing resources almost blindly, resources that always They are limited. After knowing this cause, I almost immediately asked him a question that has marked all the following days of my life: What if we make a system that can predict dengue outbreaks months in advance? The reason why I became interested in the subject was that I was finally able to visualize a clear way in which I could contribute directly to my country, including the world. For the first time I was able to conceptualize something that would act and impact independently of corruption, social classes and past problems.

How does it work?

We create an “Artificial Intelligence” (AI), using Machine Learning or “Machine Learning”, which identifies patterns and creates mathematical models based on data from the past.

This type of intelligence learns just like children do: to test, error and correction. Unlike a human, an AI is not affected by mental fatigue, fatigue, biological factors or information overload. In this way, AI systems are good at automating monotonous, repetitive tasks, and are not capable of making or learning from things outside of the data that their creators provide.

With our first platform, we decided to create an AI capable of predicting dengue outbreaks, in order to assist officials and doctors in the public health sector, so that with this tool they can make decisions based on and with time. The algorithm is able to locate in a radius of 400km the next locations where there will be outbreaks, making it a highly visual and intuitive tool.

For the development of the algorithm, in the initial prototype that my partner and I did, we included more than 500 variables per case or infected person, with an accuracy of up to 84% (out of every 100 predictions, 84 become reality). Then, with the help of a team that we were training in a way that was necessary, we improved the algorithm, and used 276 variables and learned by itself, currently with a precision of up to 88%. Unlike a statistical model, AI allows us to use agnostically a large number of variables, without greatly complicating the development process.

Our AI System uses climatic variables (such as wind direction, temperatures), socioeconomic variables (such as average income, demographics, types of houses), geographical (altitudes, lakes) and epidemiological variables (about mosquito and disease).

What kind of algorithms do you use to achieve an 88% efficiency?

We have created a new algorithm, in which part of the learning uses Neural Networks and in other Bayesian Networks. Having said that, something very new in AI companies and projects is the way we collect all the data needed for predictions. For example, we need climatic data as maximum temperature, and socioeconomic datas related to the demography of an area, to mention two examples of many.

The fundamental problem of any Data Science project (the branch of Computer Science where the AI comes from) is the accessibility of the data, and the continuous access to it. In other words, we can have great ideas that can solve a problem using AI, but if we do not have data to train an AI, we cannot do it. For this reason, we created a sub-system which we call REDINT (Remote Data Input Interface), which is a “bot” for each case (a patient who was diagnosed with Zika, for example) looking at external sources such as the Bank World or Weather.com the data we need. This “searches” are done through “API”, which are accesses that different sources give so that third parties can consume their data. In this way, for every case that arrives in our system, the REDINT bot is in charge of obtaining all the relevant data.

It is important to mention, how we obtain the data of the cases of the disease, and this is where we have to leave the technological, and go to the most human part: where we talk with Managers, Health Ministers, Presidents and Premiers, not only for implement the technology and obtain access to data that can revolutionize how Public Health is treated for its constituents, but also to change the attitude of public servants about AI. In these ways, our Artificial Intelligence models get the data they need, making predictions and improving with each of them.

Indeed, it is revolutionary and can save many lives. How are viruses mapped?

Each disease is different. Initially, we focus on vector-borne diseases (a mosquito is a vector, for example), and these have an extra agent unlike diseases such as tuberculosis, which is transmitted from human to human. It is for this fundamental reason that our team consists not only of engineers and scientists of data, but also of doctors and specialists in Public Health.

In order to really make a change in our health systems, it is essential to know them, and when we talk about Public Health, it is extremely important to know the diseases, the agents that spread it, and the speed at which they do it, to mention some points of interest.

Thanks to this inclusion, we can proceed with security and knowledge, creating platforms that do fit the needs, but with positive repercussions in other areas. For example, when we implemented in Malaysia, the health authorities had obtained funds from the federal government for a project that consisted of mapping all cases of dengue 3 years ago, in 2014. They had already worked for 4 months on the project when we arrived, and with a platform tool they were able to map all the cases in a matter of minutes.

When they really know a system, engineers can create solutions in multiple aspects of the system, so our team of doctors and engineers is extremely aware with the knowledge to identify points of failure, and to solve them.

Usually, how much can a National Health system save with this system?

We are currently in the research process to precisely determine the percentage of correct savings thanks to our implementations, both saved lives, hours of work not lost and saving money.
Empirically, two months after our implementation in Malaysia, outbreaks of dengue were reduced by 52%, thus preventing children and adults from being infected, which leads to fewer deaths, fewer lost work hours, and less investment in the future.

Governments have adopted technology at the national level in their health programs. Can you share some data of success stories and how many people already benefit from the platform?

We have implemented the technology in the entire national territory of Malaysia, in the state of Rio de Janeiro in Brazil, and in the capital of the Philippines, Manila.

Currently, more than 30 million people benefit annually, and it is a number we have projected to double in the next 9 months. With reductions of outbreaks sometimes of more than 50%, with government personnel much more efficient and productive, and with healthier communities, we are really in rejoicing that what started with a question, today is creating a real change.

You are a young entrepreneur. Have you found many obstacles from a country like Dominican Republic to develop your research?

The reality is that the road has been hard, with ups and downs and difficult. The most arduous part is not technical, engineering or scientific, but the part in which we have to deal with bureaucracies of the highest level, with implementations that take years to begin.

Something for which I feel very fortunate, is that this entrepreneurship has allowed me to know very different cultures. This has allowed me to break many stereotypes, strongly confirm others,

and know the past that led to the present of different nations and regions. In the first quarter of the year I can be in negotiations with Japanese investors in Singapore, and in the subsequent one giving a talk to the Inter-American Development Bank in Barranquilla.

In each city and country, living the local culture, eating the dishes of their cuisine, adapting to their ways of working, and above all, leading my project, trying to positively impact the lives of millions of other people.

What has it meant for you to be awarded by MIT as one of the 35 Innovators in Latin America under the age of 35?

When you are a young person, and especially of a region not recognized in technology, there is double discrimination. People do not believe in you first because your face looks very young, and second because they do not know how you got where you are. We planned for this to happen, since I have even experienced youth discrimination in my own country, but in an exponential way, much more accentuated.

Achievements like the MIT Innovators under 35, help us enormously, as our names and our brand go from something little known to something that an extremely valuable brand supports. In our presentations, we put our achievements and acknowledgments at the beginning, where we have specified that it really changes the perspective of our viewers, and thus the discrimination against our youth or origin.

It is even thanks to recognitions such as MIT, Harvard, the Clinton Foundation, the UN, the OAS and Forbes that people trusted us in the initial moments of the project.

In your opinion, what challenges do new technologies face in the field of innovation in health and medicine?

The fundamental challenge, based on my experience, is not technological, it is human.

In the area of health, it is a global phenomenon to find doctors with a blocking attitude towards innovations outside the scope of medical devices.
This attitude, which must be handled appropriately, limits the scientific advance, tests and pilots of technologies, and discourages those who really want to make a change.

After this attitude is improved, then other challenges will come to light, but first we must recognize that others have the capacity to make a positive change, which, although they do not make them so expert, makes them agents of change.

What are your research’s goals and next steps?

Currently, we are in the process of expanding our skills, to diversify ourselves to other sectors for the social good. We plan to create a new company, which is the parent of AIME and others that focus on each of the Sustainable Development Goals. We already have innovations in R&D in the fields of agriculture, citizen security, and impact management (for NGOs), with a view to future solutions related to energy, justice and smart cities.

Our idea is to call that parent company “AI4GOOD”, or “Artificial Intelligence for the [Social] Good”, and we want it to be the “Google” of impact solutions with technology.

Define the essence of your work and your research in a sentence.

Our aspiration is to eliminate the great gap between Technological Innovation and Public Health.

Leave a Reply