CodRep: Machine Learning on Source Code Competition -- 3
Budget: $30 – $250 AUD
Software engineering/machine learning problem that was launched a few years ago. Full details of the task can be found at https://github.com/KTH/CodRep-competition, but basically the challenge involves predicting where to insert a specific line into a source code file. i.e. the system takes as input a set of pairs (source code line, source code file), and outputs, for each pair, the predicted line number of the line to be replaced by in the initial source code file.
Unfortunately it's too late to enter the actual competition but this still forms an interesting machine learning problem. There are many different ways of tacking this - e.g. one approach might perform some degree of syntactic and semantic analysis of the inputs before using machine learning, whereas another might adopt a deep learning approach and train a network with hundreds of thousands of examples of maintenance changes.
Unfortunately it's too late to enter the actual competition but this still forms an interesting machine learning problem. There are many different ways of tacking this - e.g. one approach might perform some degree of syntactic and semantic analysis of the inputs before using machine learning, whereas another might adopt a deep learning approach and train a network with hundreds of thousands of examples of maintenance changes.