Combining Language and Vision with a Multimodal Skip-gram Model Angeliki Lazaridou* (University of Trento) Nghia The Pham (University of Trento) Marco Baroni (University of Trento ) Abstract ”We present MMSkip-gram, a method for inducing word representations, that extends the effective Skip-gram approach of Mikolov et al.[7]. MMSkip-gram, by exploiting visual information naturally occurying in images, is able to induce word representations that outperform Skip-gram both on general semantic tasks such as predicting word similarity and on multimodal tasks such as as zero-shot learning for image labeling.”
The paper is not available online. Please reach to the authors at
[email protected]*
[email protected] [email protected] for more information.
1
