I'm no expert, but my (very crude) understanding is that much of modern natural language parsing frameworks (e.g., Google's SyntaxNet) can be viewed as variatio

I'm no expert, but my (very crude) understanding is that much of modern natural language parsing frameworks (e.g., Google's SyntaxNet) can be viewed as variations of GLR where we slap a probability onto each possible transition.