Skip to content

Use a words-as-classifiers model to resolve references to objects in photos, including the ability to correctly interpret indefinite NPs and quantification

Notifications You must be signed in to change notification settings

betoma/words-as-classifiers

Repository files navigation

Quantification for words-as-classifiers

Both the project itself and this readme are still in their very early work-in-progress stages. That said, feel free to contact me with any suggestions.

An extension of the words-as-classifiers reference resolution performed in Schlangen et al. (2016). "Resolving References to Objects in Photographs using the Words-As-Classifiers Model" and Kennington et al. (2015). "Simple Learning and Compositional Application of Perceptually Grounded Word Meanings for Incremental Reference Resolution". The goal is to add the ability to parse quantificational noun phrases rather than merely definite singular noun phrases. At the moment, the plan is to implement a simulation using a threshold to turn the probability distributions produced by classifiers into traditional boolean sets and then to compare this with an implementation that uses fuzzy sets.

Note: this repo is out of date as I work on further related work for my thesis. Will update the repos in question eventually.

About

Use a words-as-classifiers model to resolve references to objects in photos, including the ability to correctly interpret indefinite NPs and quantification

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published