Language Is Not All You Need: Aligning Perception with Language Models

TL;DR

Kosmos-1 is a multimodal large language model trained on diverse web-scale data, capable of understanding and generating across language, vision, and perception tasks without fine-tuning, advancing towards artificial general intelligence.

Contribution

This work introduces Kosmos-1, a novel multimodal large language model trained from scratch on interleaved text and images, demonstrating strong zero-shot and few-shot performance across multiple modalities.

Findings

Achieves high performance on language understanding and generation tasks.

Excels in perception-language tasks like VQA and image captioning.

Shows effective cross-modal transfer of knowledge.

Abstract

A big convergence of language, multimodal perception, action, and world modeling is a key step toward artificial general intelligence. In this work, we introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zero-shot). Specifically, we train Kosmos-1 from scratch on web-scale multimodal corpora, including arbitrarily interleaved text and images, image-caption pairs, and text data. We evaluate various settings, including zero-shot, few-shot, and multimodal chain-of-thought prompting, on a wide range of tasks without any gradient updates or finetuning. Experimental results show that Kosmos-1 achieves impressive performance on (i) language understanding, generation, and even OCR-free NLP (directly fed with document images), (ii) perception-language tasks, including multimodal dialogue, image captioning, visual question answering, and (iii) vision tasks, such as image recognition with descriptions (specifying classification via text instructions). We also show that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge from language to multimodal, and from multimodal to language. In addition, we introduce a dataset of Raven IQ test, which diagnoses the nonverbal reasoning capability of MLLMs.