SharpMoE leverages predicted clean latent features for saliency-aware routing, boosting diffusion MoE image generation by ~20% in FID.

Haoyou Deng, Keyu Yan, Chaojie Mao et al.

PortraitGen injects real exemplars into GRPO and combines OmniReward with AI-Portrait, reaching 96.35% artifact-detection accuracy.

Xiaomin Li, Qian Liang, Yinan Li et al.

PhysEditWorld, a UE5-based multimodal dataset, enables explicit control of physics (gravity) via matched replays, supporting physics-aware world modeling.

Bin Hu, Yanwen Ma, Jiehui Huang et al.

DiCoBench introduces a high-resolution multi-image benchmark focusing on differential and commonality visual cues, revealing models' significant gaps in fine-grained perception, with top accuracy at 58.1%.

Geng Li, Yuxin Peng

Introduces PhyEditBench, a real-world video-based benchmark for physics reasoning, revealing current models' limitations in physical understanding.

Shengbin Guo, Shaokang He, Chaoyue Meng et al.

SimPhysNet combines physics-informed neural networks with self-supervised contrastive learning to achieve 96.06% accuracy in laser welding penetration prediction using only 200 labeled images.

Sen Li, Xiaoying Liu, Xiaojian Xu et al.

UniTeD introduces a joint temporal diffusion model for perception and planning, achieving state-of-the-art results with 90.24 PDMS on NAVSIM v1.

Bo Zhao, Xinting Zhao, Naifan Li et al.

Proposed SSMNBench, a diagnostic benchmark with 3300 QA pairs, distinguishes single-view sufficiency and multi-view necessity to evaluate cross-view understanding of 17 models.

Tianchen Guo, Chen Liu, Ling Chen et al.

LCG uses Sparse Relational Attention for consistent long-context image generation, improving consistency and scalability.

Zihao Wang, Yijia Xu, Haoze Zheng et al.

ScaleHP optimizes coupled errors in camera-space hand pose estimation, achieving 35.8mm CS-MPJPE on FreiHand.

Ruitao Jing, Xingyu Chen, Hongyang Li et al.

VPA-Guard combines retrieval-augmented reasoning and self-evolution to reduce visual prompt attack success rates by 44.2%, enhancing video generation safety.

Yining Sun, Haoyu Kang, Jiajun Wu et al.

BLOCK replaces model bottleneck layers with trained transcoders for persistent concept removal in SD3.5, Flux, Infinity, outperforming state-of-the-art methods.

Aditya Kumar, Pierre Joly, Adam Dziedzic et al.

C3-Bench creates a multi-domain change captioning benchmark with 51 real-world scenarios, integrating LLM-based fine-grained evaluation for robustness.

Jae-Woo Kim, Hyeongbeom Kim, Ue-Hwan Kim

Proposes Teach-to-Reason (T2R), a framework using self-improving teachers and contrastive supervision to enhance reasoning in CXR VQA, outperforming baselines.

Xiao Han, Hao Liu, Zhimin Bao et al.

FLUX3D employs Diffusion-Aligned Structured Latents and sparse-structure-aware diffusion transformer to generate high-fidelity 3D Gaussian point clouds, outperforming SOTA methods.

Haorui Ji, Weizhe Liu, Hongdong Li et al.

TuringViT introduces Turing Linear Attention, VISTA-Curation, and dynamic-resolution pretraining, achieving high performance with only 10% of the data and improved efficiency.

Qiman Wu, Hanlin Chen, Lyujie Chen et al.

Proposes a semantic browsing framework for controllable diversity in image generation, leveraging rich textual representations to enable structured, interpretable exploration.

Sara Dorfman, Maya Vishnevsky, Omer Dahary et al.

Proposes an iterative self-filtering method using CLIP to improve vision-language data quality without pretraining, boosting downstream performance.

Andrei Liviu Nicolicioiu, Sarvjeet Singh Ghotra, Morgane M. Moss et al.

InteractiveAvatar uses LSVM and RRM to achieve long-term visual consistency and intent-aware real-time avatar generation.

Quanyue Song, Yishan He, Yanfei Zhang et al.

Proposed Sol Video Inference Engine uses five techniques—cache, sparse attention, token pruning, quantization, kernel fusion—for 2× acceleration of video diffusion models.

Yitong Li, Junsong Chen, Haopeng Li et al.