A Simple LLM Framework for Long-Range Video Question-Answering

TL;DR

LLoVi is a simple, effective framework for long-range video question-answering that leverages visual captioners and large language models, achieving state-of-the-art results without complex modeling techniques.

Contribution

The paper introduces LLoVi, a straightforward LVQA approach using captioning and LLMs, outperforming prior complex methods on multiple benchmarks.

Findings

Achieves 50.3% accuracy on EgoSchema, surpassing previous best by 18.1%.

Outperforms state-of-the-art on NeXT-QA and IntentQA datasets.

Extends to grounded LVQA, outperforming all prior methods.

Abstract

We present LLoVi, a language-based framework for long-range video question-answering (LVQA). Unlike prior long-range video understanding methods, which are often costly and require specialized long-range video modeling design (e.g., memory queues, state-space layers, etc.), our approach uses a frame/clip-level visual captioner (e.g., BLIP2, LaViLa, LLaVA) coupled with a Large Language Model (GPT-3.5, GPT-4) leading to a simple yet surprisingly effective LVQA framework. Specifically, we decompose short and long-range modeling aspects of LVQA into two stages. First, we use a short-term visual captioner to generate textual descriptions of short video clips (0.5-8s in length) densely sampled from a long input video. Afterward, an LLM aggregates the densely extracted short-term captions to perform long-range temporal reasoning needed to understand the whole video and answer a question. To analyze what makes our simple framework so effective, we thoroughly evaluate various components of our system. Our empirical analysis reveals that the choice of the visual captioner and LLM is critical for good LVQA performance. Furthermore, we show that a specialized prompt that asks the LLM first to summarize the noisy short-term visual captions and then answer a given input question leads to a significant LVQA performance boost. On EgoSchema, which is best known as a very long-form video question-answering benchmark, our method achieves 50.3% accuracy, outperforming the previous best-performing approach by 18.1% (absolute gain). In addition, our approach outperforms the previous state-of-the-art by 4.1% and 3.1% on NeXT-QA and IntentQA. We also extend LLoVi to grounded LVQA and show that it outperforms all prior methods on the NeXT-GQA dataset. We will release our code at https://github.com/CeeZh/LLoVi.