I was recently asked if machine learning requires evaluation metrics. The answer is no, and the fact that people are asking such questions suggests that some newcomers to the field may have a limited perspective on NLP research methodology.
Some comments on how different components in the NLG pipeline can “add value” by contributing to the ultimate goal of generating texts that easy for people to read and understand.
I think surface realisation becomes especially challenging when syntax depends on semantics or pragmatics. From engineering perspective, handling phenomena that only occur in a few languages can be painful.
Many students get stressed about their PhD viva (oral exam) even though they are very unlikely to fail. I present some rules and a flowchart to suggest when there is real cause for concern, and when there is not.
A few comments on how I review papers (what I actually do, not what I am supposed to do), and associated advice for authors.
Lexical choice is an area of NLG which really needs machine-learning and data-based techniques.
Perhaps being somewhat idealistic, I think academic researchers should act as “scouts” exploring unknown research terrain
I went to my first developers conference last week and was impressed, not least by the sensible attitude towards deep learning and other trendy AI technology.
Good software engineering is criticial when building NLG systems, including requirements analysis, design, testing, and support.
If you are writing a scientific paper which presents statistics, please use two-tailed p values unless you **really** know what you are doing.
Ehud Reiter's Blog 