7 papers
Visual Latents Know More Than They Say: Unsilencing Latent Reasoning in MLLMs
Xin Zhang, Qiqi Tao, Jiawei Du +2
Continuous latent-space reasoning offers a compact alternative to textual chain-of-thought for multimodal models, enabling high-dimensional visual evidence to be integrated without…
Mitigating Entangled Steering in Large Vision-Language Models for Hallucination Reduction
Yuanhong Zhang, Zhaoyang Wang, Xin Zhang +2
Large Vision-Language Models (LVLMs) have achieved remarkable success across cross-modal tasks but remain hindered by hallucinations, producing textual outputs inconsistent with vi…
AEGIS: Authenticity Evaluation Benchmark for AI-Generated Video Sequences
Jieyu Li, Xin Zhang, Joey Tianyi Zhou
Recent advances in AI-generated content have fueled the rise of highly realistic synthetic videos, posing severe risks to societal trust and digital integrity. Existing benchmarks…
Modelship Attribution: Tracing Multi-Stage Manipulations Across Generative Models
Zhiya Tan, Xin Zhang, Joey Tianyi Zhou
As generative techniques become increasingly accessible, authentic visuals are frequently subjected to iterative alterations by various individuals employing a variety of tools. Cu…
Beyond Modality Collapse: Representations Blending for Multimodal Dataset Distillation
Xin Zhang, Ziruo Zhang, Jiawei Du +2
Multimodal Dataset Distillation (MDD) seeks to condense large-scale image-text datasets into compact surrogates while retaining their effectiveness for cross-modal learning. Despit…
Breaking Class Barriers: Efficient Dataset Distillation via Inter-Class Feature Compensator
Xin Zhang, Jiawei Du, Ping Liu +1
Dataset distillation has emerged as a technique aiming to condense informative features from large, natural datasets into a compact and synthetic form. While recent advancements ha…