The goal of this is to apply the Probabilistic Soft Logic (PSL) framework to a problem with relational data.
The starting point is a database with the raw data. This database contains information about entities (persons and organizations) and certain actions undertaken by these entities. From this data, we can extract certain relations such as "Person P is a member of organization O". Some relations, however, are missing and that's where PSL comes in. We are basically interested in the probability that certain relations exist in the real-world even though they are not in the database.
To this end, we will use the existing PSL framework. This is an important point: we don't have to write our own framework. We can use something that's has already been developed. There are excellent video tutorial available, including one tutorial that explains the intuition and theory behind PSL and another tutorial that demonstrates PSL based on example that is very similar to what we need.
You can either apply with or without experience using PSL. If you don't have experience with PSL, you can still apply under the following conditions:
1) The fact that you are beginner will be reflected in your compensation. You can use this as a learning opportunity, but we cannot pay you to read basic tutorials.
2) You've worked on other data mining/machine learning projects before.
The framework is written in Groovy. Our project where we apply the framework will be written in Kotlin. However, I expect the code basis to be small. Lack of experience with Kotlin development shouldn't be an impediment since it's easy to learn and very similar to other languages that you might be familiar with such as Scala, Ceylon and Groovy. You should, however, be familiar with at least one JVM language other than Java.
I'm looking forward to your message. We will discuss the details of this project via Skype.
We'll have at the very least one other similar project. If the current project works out as expected, we would like to continue working with you on the other tasks.