Ensemble based discriminative models for Visual Dialog Challenge 2018

TL;DR

This paper presents an ensemble of discriminative models for the Visual Dialog Challenge 2018, achieving competitive results by combining different encoders and decoders to improve dialog understanding.

Contribution

Introduces an ensemble approach with diverse discriminative models for visual dialog, achieving top-tier challenge performance.

Findings

NDCG score of 55.46 on test-std split

MRR value of 63.77 on test-std split

Secured third position in the challenge

Abstract

This manuscript describes our approach for the Visual Dialog Challenge 2018. We use an ensemble of three discriminative models with different encoders and decoders for our final submission. Our best performing model on 'test-std' split achieves the NDCG score of 55.46 and the MRR value of 63.77, securing third position in the challenge.