Paper 2602.13013
Towards Universal Video MLLMs with Attribute-Structured and Quality-Verified Instructions
- Published
- Feb 2026
- Research lab
- Independent
- Citations
- 5
- GitHub
- 68 stars
01 In brief
Summary
The paper introduces ASID-1M, an open-source dataset of one million attribute-structured audiovisual instruction annotations, along with ASID-Verify, a multi-stage data curation pipeline for automatic annotation, verification, and refinement, and ASID-Captioner, a video understanding model trained via supervised fine-tuning on this dataset.
The pipeline generates multi-source captions using models like AVoCaDO and Seed-1.6, ensembles them with ASR alignment and temporal consistency checks, and performs attribute-level evaluation and targeted refinement to reduce hallucinations and missing details.
ASID-Captioner is trained progressively from single-attribute to all-attribute supervision on short and long video clips.
Evaluations on seven benchmarks (including video-SALMONN-2, UGC-VideoCap, VDC, VidCapBench-AE, Daily-Omni, World-Sense, and Charades-STA) show that ASID-Captioner achieves state-of-the-art performance among open-source models, with the 7B variant competitive with Gemini-3-Pro.
Ablations confirm that multi-attribute supervision and the three-stage training recipe improve caption quality and instruction following.
The dataset, model, and code are publicly released.
02 From the paper
Abstract
Universal video understanding requires modeling fine-grained visual and audio information over time in diverse real-world scenarios. However, the performance of existing models is primarily constrained by video-instruction data that represents complex audiovisual content as single, incomplete descriptions, lacking fine-grained organization and reliable annotation. To address this, we introduce: (i) ASID-1M, an open-source collection of one million structured, fine-grained audiovisual instruction annotations with single- and multi-attribute supervision; (ii) ASID-Verify, a scalable data curation pipeline for annotation, with automatic verification and refinement that enforces semantic and temporal consistency between descriptions and the corresponding audiovisual content; and (iii) ASID-Captioner, a video understanding model trained via Supervised Fine-Tuning (SFT) on the ASID-1M. Experiments across seven benchmarks covering audiovisual captioning, attribute-wise captioning, caption-based QA, and caption-based temporal grounding show that ASID-Captioner improves fine-grained caption quality while reducing hallucinations and improving instruction following. It achieves state-of-the-art performance among open-source models and is competitive with Gemini-3-Pro.