# Evaluating the Representational Hub of Language and Vision Models

## TL;DR

This paper investigates how multimodal neural network encoders, inspired by cognitive science's 'Hub and Spoke' model, process and combine visual and linguistic information across various tasks.

## Contribution

It provides a comprehensive analysis of the representations learned by vision-language encoders and their effectiveness in multimodal understanding tasks.

## Key findings

- Encoder representations vary across tasks
- Pre-trained encoders show different levels of semantic understanding
- Analysis reveals how modalities are integrated in the encoder

## Abstract

The multimodal models used in the emerging field at the intersection of computational linguistics and computer vision implement the bottom-up processing of the `Hub and Spoke' architecture proposed in cognitive science to represent how the brain processes and combines multi-sensory inputs. In particular, the Hub is implemented as a neural network encoder. We investigate the effect on this encoder of various vision-and-language tasks proposed in the literature: visual question answering, visual reference resolution, and visually grounded dialogue. To measure the quality of the representations learned by the encoder, we use two kinds of analyses. First, we evaluate the encoder pre-trained on the different vision-and-language tasks on an existing diagnostic task designed to assess multimodal semantic understanding. Second, we carry out a battery of analyses aimed at studying how the encoder merges and exploits the two modalities.