A project to align AMRs with CCG syntax.
ablodge / amr-ccg-alignment Goto Github PK
View Code? Open in Web Editor NEWA project to align AMRs with CCG syntax
A project to align AMRs with CCG syntax
The javascript interface should:
A minor annoyance is that currently, to adjust an alignment from preprocessing, I have to delete it and create a new one, which is placed last in the list. Then it takes time to go back and find the next prealignment to review. This could be improved by any of the following:
Consider adding an option to show all alignments for an AMR at the same time. Overlapping alignments could have multiple shading: multi-shades css.
Add dependencies (displayed over sentence) for each amr.
Add a way to remove an alignment.
When hovering on a word or part of the AMR which is part of one or more existing alignments, it would be helpful if the alignment buttons for those alignments would be darkened. That way when I'm browsing the alignments for a sentence, it is easy to select one for a particular word or AMR node/edge.
For example, the border of the input box could turn red.
I should make a tool that reads an AMR/UD alignment file and outputs html with the (lexical) alignments colored.
Users should be able to align AMR and CCG elements. For CCG visualization, see https://github.com/nschneid/ccg-cxns.
E.g. control verbs should coindex the subject NP of the complement with the subject of the matrix verb: something like (S\NP.1)/(S[inf]\NP.1) for "want" and (S[inf]\NP.1)/(S[b]\NP.1) for "to" in "I want to eat cake". Is it possible to obtain this information, perhaps using GraphParser?
If so, then we could tell annotators not to align anything coindexed with part of an outer argument: in the example, the semantic relation corresponding to expected subject of the embedded verb would account for the NP.1 argument, so "want" would not align to any of the semantic args of "eat".
Today I suggested there could be type conversion rules to turn S[b]\NP into S[imp] for imperatives and NP into S[voc]/S[dcl] for vocatives. But this would make alignment to the AMR difficult, and in retrospect it violates the spirit of CCG which is to bake as much as possible into the lexicon.
What about lexical rules as follows:
Thus:
This all does feel like a hack, though.
Users should be able to upload AMRs and other data.
Planned support for:
I should make a tool that reads an AMR/UD alignment file and outputs html with the (structural) alignments colored. Extends #1 for alignments between subgraphs.
Users should be able to download text file of alignments. Consider supporting uploading partial alignments, for additions/corrections.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.