Mova · papers
Main list
The 105 works of Mova's own list, with their effect and the stage of the pipeline that uses them. measured marks a work whose effect Mova measured in its own experiments — 30 of them, including ideas that did not help; the others are rule sources, designs or background. Each stage has its page.
text → tokens
Splitting the text into words with their part of speech, lemma and features: the tagger and its lexicon.
3 papers · 1 measuredtokens → tree
Building the Universal Dependencies tree: the transition parser, the treebanks, the grammars of English as rule sources.
50 papers · 3 measuredtree → frames
Reading the tree into frames — who did what to whom — with speech acts, hedges and coreference.
20 papers · 8 measuredframes → world
Keeping the world of the text: entities, their states, quantities and compiled commonsense.
9 papers · 5 measuredworld → answer
Answering from the world, with a reason: school tests, reading questions, word problems.
11 papers · 6 measuredand back, byte for byte
Giving the text back exactly: round trips and converters between dialects of UD.
2 papers · 0 measuredhow it learns
How the parsers learn: perceptrons, instance-based learning, bootstrapping and the teacher loop.
10 papers · 7 measured