Dank Learning: Generating Memes Using Deep Neural Networks

TL;DR

This paper presents a deep learning system that generates humorous and relevant memes from images, allowing user control over meme content through labels, and demonstrates its effectiveness through human and perplexity evaluations.

Contribution

The paper introduces a novel meme generation approach combining image embeddings and attention-based LSTM, with diversity-promoting decoding, achieving realistic meme creation.

Findings

Generated memes are often indistinguishable from real ones.

The model produces diverse and relevant meme captions.

Human assessments confirm meme quality and relevance.

Abstract

We introduce a novel meme generation system, which given any image can produce a humorous and relevant caption. Furthermore, the system can be conditioned on not only an image but also a user-defined label relating to the meme template, giving a handle to the user on meme content. The system uses a pretrained Inception-v3 network to return an image embedding which is passed to an attention-based deep-layer LSTM model producing the caption - inspired by the widely recognised Show and Tell Model. We implement a modified beam search to encourage diversity in the captions. We evaluate the quality of our model using perplexity and human assessment on both the quality of memes generated and whether they can be differentiated from real ones. Our model produces original memes that cannot on the whole be differentiated from real ones.