Talk 2: in commercial clone&own settings, it is about customization, thus, code deviates from its origin. How do you cope with this considering FL & CD?

Talk 1: if we look for X-lang similar applications, would clone detection be the best way to go? Chen et al (What is Spain's Paris? SANER16) argued for word-vec

Talk 2: could you use the feature detection to identify malicious behaviour in software (e.g. malicious user profiling, etc.) ?

all speakers: what obstacles remain before clone detection can be used at code review time at Facebook, like Sapienz?

Talk 3: did you use the fact that a given clone may be returned by more than one clone detection system as a feature of machine learning?

talk4: what are the number of clones in DSLs vs conventional languages like java?

Talk2: are the results presented as chunks of source code, or in a higher level model of feature similarity?

Talk 2: how far do you think the automation of the feature extraction should go?

How did you decide which feature location approach to use?

Talk 2: how far do you think the automation of the feature extraction should go?