Multimodal large language models (MLLMs) have achieved remarkable performance across vision-language tasks, but their progress depends heavily on large-scale, high-quality multimodal data that are costly to annotate. Self-augmentation offers a promising alternative by enabling models to expand their own training data without external supervision.
arXiv:2608. 03733v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) have achieved remarkable performance across vision-language tasks, but their progress depends heavily on large-scale, high-quality multimodal data that are costly to annotate.
arXiv:2605.22208v3 Announce Type: replace Abstract: Multimodal Large Language Model (MLLM)-driven image restoration agent demonstrates effectiveness in degradation coupling scenarios by flexibly sele...
WeAgent-MMSearch introduces a multimodal search agent that preserves retrieved images as persistent references, enabling the model to inspect, process, and cite them throughout a search trajectory. The system includes a harness (WeAgent-Harness), a post‑training method (FA‑GSPO) that recovers salvageable rollouts, and a new benchmark (VisTarget‑Bench) to evaluate image‑retrieval versus visual‑perception failures. Evaluation shows that agentic post‑training boosts performance by 19.22 points, allowing the model to outperform similarly sized open‑source models and compete with much larger ones.
ClearText-Video (CTVid) is a large-scale, scene-text-aware benchmark that examines text-centric video understanding under varying quality conditions. It comprises 4,639 real-world egocentric videos, over 550,000 frames, 1.6 million human-verified scene-text annotations, and more than 220,000 spatial/temporal question–answer pairs in Chinese and English. For each high-quality video, CTVid provides matched degraded- and restored-quality variants, enabling studies of Text-Centric Video Restoration and Multi-Quality VideoQA, and revealing that visual enhancement does not always improve textual fidelity or downstream reasoning.
arXiv:2606. 27974v1 Announce Type: cross Abstract: Knowledge-based Visual Question Answering (KB-VQA) requires models to combine image understanding with external knowledge.
ShredBench is a new benchmark for evaluating multimodal large language models (MLLMs) on the task of reconstructing documents from shredded fragments. It uses an automated pipeline that generates fragmented documents from Markdown, covering English, Chinese, code, and table scenarios with varying fragmentation levels. Experiments show that while MLLMs perform well on intact documents, their performance drops sharply when documents are shredded, revealing a gap in fine‑grained cross‑modal reasoning.
NoteVQA is a new benchmark that collects 252 real‑life visual questions from the Chinese image‑sharing platform Xiaohongshu, covering 12 topics and 7 user intents. Each question is paired with a concise expert reference and a human‑audited interleaved answer that blends text and visual evidence. The study evaluates VLMs on short‑answer correctness and interleaved answer quality using a new AgenticInterleave framework and a 12‑dimensional IVR‑12 rubric, finding that even state‑of‑the‑art models achieve only about 53% accuracy and lag behind human references in content quality.
arXiv:2609.22942v1 Announce Type: new Abstract: Generative models are rapidly expanding image quality assessment (IQA) beyond traditional fidelity factors to emerging dimensions such as physical plau...
The paper introduces ImIR, a method that tunes a large pretrained image‑editing model for all‑in‑one image restoration by replacing text prompts with continuous image‑derived instructions. The approach uses a lightweight token mapper to shift the degraded image’s vision‑language embedding toward that of a clean image, enabling a single adapter to handle six restoration tasks in about three hours on one GPU. ImIR outperforms text conditioning in matched comparisons and supports task‑agnostic restoration without requiring a degradation label.
ATP‑Bench proposes a new benchmark for evaluating agentic tool planning in multimodal large language models (MLLMs) that generate interleaved text-and-image responses. The benchmark contains 7,702 QA pairs, including 1,592 visual‑question‑answer pairs, across eight categories and 25 visual‑critical intents, all verified by humans. A Multi‑Agent MLLM‑as‑a‑Judge (MAM) system is introduced to assess tool‑call precision, missed opportunities, and overall response quality without relying on ground‑truth references.
Degradations vary widely across images, so a practical restoration system has to handle many degradation types with one model. A recent and effective recipe adapts a large pretrained image-editing mod...