Next Token Prediction Towards Multimodal Intelligence: A Comprehensive Survey

TL;DR

This survey reviews how Next Token Prediction (NTP) is expanding from language models to multimodal tasks, proposing a taxonomy to unify understanding and generation across modalities.

Contribution

It introduces a comprehensive taxonomy for multimodal NTP, covering tokenization, model architectures, task representation, datasets, and challenges, aiding future research.

Findings

NTP is effective across various modalities.

A unified taxonomy for multimodal NTP is proposed.

Open challenges in multimodal NTP are identified.

Abstract

Building on the foundations of language modeling in natural language processing, Next Token Prediction (NTP) has evolved into a versatile training objective for machine learning tasks across various modalities, achieving considerable success. As Large Language Models (LLMs) have advanced to unify understanding and generation tasks within the textual modality, recent research has shown that tasks from different modalities can also be effectively encapsulated within the NTP framework, transforming the multimodal information into tokens and predict the next one given the context. This survey introduces a comprehensive taxonomy that unifies both understanding and generation within multimodal learning through the lens of NTP. The proposed taxonomy covers five key aspects: Multimodal tokenization, MMNTP model architectures, unified task representation, datasets \& evaluation, and open challenges. This new taxonomy aims to aid researchers in their exploration of multimodal intelligence. An associated GitHub repository collecting the latest papers and repos is available at https://github.com/LMM101/Awesome-Multimodal-Next-Token-Prediction