# SuperChat: Dialogue Generation by Transfer Learning from Vision to Language using Two-dimensional Word Embedding and Pretrained ImageNet CNN Models

## TL;DR

This paper introduces SuperChat, a novel dialogue generation method that leverages two-dimensional word embeddings and pretrained CNN models from vision tasks, achieving high-quality responses in open domain conversations.

## Contribution

It adapts the Super Characters approach and 2D embeddings from text classification to dialogue generation, combining vision-based models with language processing.

## Key findings

- SuperChat generates high-quality conversational responses.
- The method outperforms traditional text-based models on a public dataset.
- An interactive demo demonstrates practical effectiveness.

## Abstract

The recent work of Super Characters method using two-dimensional word embedding achieved state-of-the-art results in text classification tasks, showcasing the promise of this new approach. This paper borrows the idea of Super Characters method and two-dimensional embedding, and proposes a method of generating conversational response for open domain dialogues. The experimental results on a public dataset shows that the proposed SuperChat method generates high quality responses. An interactive demo is ready to show at the workshop.