The researcher Jorge Amaya, from the Center of Mathematical Modeling of Universidad de Chile, created a large algorithm, able to run hundreds of thousands of equations to answer a single question: How to improve the Chilean public education? His answers load from how, how much and where to invest, to what schools close and which ones open.
What goes falling, by the sides of the white screen, is a thread of numbers. Small values, ones and zeros mostly, a three, a five, some algebraic symbol, and the same again, or likely. Paula Uribe, 33 years of age, curly haired and glasses on, watch them fall: she has spent a decade doing it and is the only one who understands it. At her back, in a small office on the sixth floor of the Center for Mathematical Modeling of Universidad de Chile, a board is saturated with numbers that dictate orders. Laws are not written with words, but with equations, which in turn are composed by variables.
In the chaos, we can distinguish some obvious ones, such as the quantity of schools and students, but on the screen are all: in this case, a simulation of the whole public system of Peñalolen and Maipú, 50 thousand variables feed a hundred thousand equations, and that load the numbers of a narrow thread. The engineer knows what they say: a school has to be closed, another needs to be open by a certain corner, fuse other two.
In front of her, 64 year old mathematics engineer and doctor in applied sciences, Jorge Amaya, white shirt and glasses on, looks enthusiastic. After five years of pilot testing, and correcting and improving his enormous algorithm time after time, things now seem to be accelerating: last month he managed a meet up with technicians of the Ministry of Education to show them their technology, and since January Corfo is contributing 150 million CLP to complete its development. Before, they had presented it to the previous government, municipalities and educational corporations, but the enthusiasm had never reached concrete terms. Now, at midday of a rainy Tuesday, gets closer to the screen on which the data keeps falling, and tries to explain with words what the algorithm does which took six months writing on a paper, and would now take half an hour to write on a board.
There’s thousands of variables! A big case, like all the schools in Santiago, can get to have a million of variables. Hours of teaching, classrooms, quantity of courses, if its technic or professional quantity of teachers, teaching hours: logical variables, geographical, demographical. All these things the algorithm analyses to get a solution that satisfies the educational demand of the children at the same cost. All the alternatives. To close a school, to open one, where to do so, primary or secondary, for how many children, with how many hours, everything.
What they are running now is not everything. Are 40 schools, from Peñalolén and Maipú, and the algorithm will take a little bit more than five minutes to reach an answer. The information will go nine floors up from the underground level in which the only head able to process to that speed is located: Leftraru —Lautaro, in mapudungun language—, a supercomputer of more than a million dollars, the most powerful in South America, financed by Conicyt and managed by the Center for Mathematical Modeling for all the universities of the country. In it, the fifty researchers of the center try to find novel mathematical solutions to solve concrete problems of the society. And in it, Jorge Amaya searches the mathematical keys to one of the big questions of the educational reform: how to strengthen again, at an assumable cost, all the Chilean public education.
For that, his algorithm, that because of the lack of a good idea is momentarily called MORE–Modeling of Educational Resources–, is able to give many answers: according to the size of data that upload to their system, that can reach from a suburb to a whole country, hundreds of thousands of equations deliver the best distribution possible of the resources, with a forced rule that no child must go to a school that is located more than 10 minutes from their home. Apart from that, the algorithm doesn’t have mayor complications: can suggest either to open a school or close two or fire ten teachers, even when outside of the computer that is not as easy.
—Municipalities have all the data, but one thing is to have millions of data, and other thing is to have information —says the mathematician—, we do mathematics to objectivate things up. But of course there’s subjectivities: you only need to close a school and the parents will chain themselves at the gates. But we are not technocrats, not all the computer says is word of God. If there´s schools that can’t be closed, because they receive vulnerable children or because they have an important role in the community, our algorithm allow us to signal, for instance, that that one can´t be touched. In this there´s political decisions, but we can give an economical orientation upon which things can be decided. Our goal is the ministry to run this every year, before meaning a decision.
To read further please visit the link below,
Source: Revista Qué Pasa