Request edit access
What is Recurrent Neural Net Grammars (RNNG), hint:
it's only discriminative model that gives a probability of a parse tree for the generated sentence
it's only generative model for jointly building tree and a sentence together which uses stack and rnn
it's a common name for both generative and discriminative models mentioned above
it's recursive neural networks with stack
Choose all the correct statements:
Discriminative models are good at building Parse tree
RNNGs are bad at capturing of hierarchical relationships among words and phrases
Sequential models are good for language modeling
Parse tree by RNNG is constructed top down
Recursive Neural Networks have better inductive bias than RNNs
What possible ways of word representation do you know from the lecture?
Consider the buffer that contains "A smart student learns reinforcement learning with reinforcement.". What will be the stack look like after evaluation the following sequence of operations: "NT(S) | NT(NP) | SHIFT | SHIFT | SHIFT"?
Write down the possible bracketed expression of constituency-based tree using the following abbreviations: S, VP, NP, DP, V, N, D for given sentence: "A smart student learns reinforcement learning with reinforcement.".
Never submit passwords through Google Forms.
This content is neither created nor endorsed by Google.
Terms of Service