In many branches of science (e.g., Artificial Intelligence, Engineering etc.), the modelling of the problem is done through the use of functions (e.g., f(x) = y). On a very high-level, we can think of Machine Learning as the problem of approximating function f from the pair of measurements (x,y), and Optimization as the problem of finding the value of input x that maximizes the output y given function f. However in many real-world problems, it is difficult to know or even evaluate this function f. For example in Chemistry, it is typical for a pharmaceutical company to run expensive wet lab experiments for months before the value of output y is measured for some given input x. This creates an exciting challenge for AI researchers to develop smart algorithms that can find the optimal value of input x that maximizes the output y given function f when the evaluation of such function f is expensive.
The purpose of this exciting project is to design and visualize an algorithm that can perform what is known as 'Blackbox Optimization for Unknown Functions', which will help the users (e.g., scientists) to explore the input space of their experiments (i.e., x) that maximizes the unknown function of interest (i.e., f(x)) by performing minimal number of function evaluations (e.g., running minimum number of wet lab experiments).
Proficiency in Python Programming, Machine Learning and/or Optimization.