Language Repository for Long Video Understanding
This addresses the challenge of declining LLM effectiveness with long input lengths for applications such as long-form video understanding, representing an incremental improvement through a novel repository-based approach.
The paper tackles the problem of long-form video understanding by introducing a Language Repository (LangRepo) for LLMs, which maintains concise, structured textual representations from multi-scale video chunks, achieving state-of-the-art performance on zero-shot visual question-answering benchmarks like EgoSchema, NExT-QA, IntentQA, and NExT-GQA.
Language has become a prominent modality in computer vision with the rise of LLMs. Despite supporting long context-lengths, their effectiveness in handling long-term information gradually declines with input length. This becomes critical, especially in applications such as long-form video understanding. In this paper, we introduce a Language Repository (LangRepo) for LLMs, that maintains concise and structured information as an interpretable (i.e., all-textual) representation. Our repository is updated iteratively based on multi-scale video chunks. We introduce write and read operations that focus on pruning redundancies in text, and extracting information at various temporal scales. The proposed framework is evaluated on zero-shot visual question-answering benchmarks including EgoSchema, NExT-QA, IntentQA and NExT-GQA, showing state-of-the-art performance at its scale. Our code is available at https://github.com/kkahatapitiya/LangRepo.