An MLLM-guided architecture with a mixture of frequency experts and relational alignment loss achieves state-of-the-art all-in-one image restoration, outperforming prior methods by up to 1.35 dB on the CDD11 dataset.
hub
In: International conference on machine learning
17 Pith papers cite this work. Polarity classification is still indexing.
17
Pith papers citing it
hub tools
citation-role summary
background 1
method 1
citation-polarity summary
representative citing papers
DART routes zero-shot video temporal grounding queries by difficulty using DPP entropy, achieving up to 3.5 mIoU gains with 7x fewer frames on Charades-STA and ActivityNet Captions.
MemLearner introduces a learning-based adaptive context query method using query tokens in video world models to improve long-term scene consistency over rule-based retrieval.
A temperature-perturbed black-box attack infers video training membership in VideoLLMs with 0.68 AUC by exploiting sharper generation behavior on member samples.
DualTrack uses decoupled local spatiotemporal and global anatomical encoders with a fusion module to estimate probe trajectories from 2D ultrasound sequences, achieving sub-5 mm average reconstruction error on public benchmarks.
An agentic mRAG framework uses GRPO-trained visual reranking and active rejection to verify retrieved candidate entities, achieving state-of-the-art on three KB-VQA benchmarks.
Ego3DLM jointly predicts past and future 3D body pose and motion descriptions in a single autoregressive pass, conditioned on egocentric video, 3D scene features, and three-point tracking, achieving state-of-the-art on the Nymeria benchmark.
StochasT uses stochastic clustering of language tasks into varying turn depths for the same image to improve LVLMs on both single-turn and multi-turn scenarios without discarding data.
Audit of KB-VQA benchmarks reveals systematic violations of answer derivability, question clarity, and visual disambiguation assumptions, with new repair and multi-entity augmentation protocols producing different model performance trends.
MV-GEL uses a learned view selector and a fine-tuned vision-language segmentation model to localize text-described faces and edges on 3D meshes.
WaterGen decouples scene generation from medium degradation in a two-stage latent diffusion process to produce controllable realistic underwater images that improve downstream restoration and segmentation.
Hi-GaTA is a hierarchical gated temporal aggregation adapter that uses short-to-long temporal pyramids and gated fusion to enable surgical video report generation, backed by a new 214-video benchmark and a surgical ViViT pretrained on 40,000 minutes of video.
SemEnrich enriches radiology reports with positive/neutral findings via self-supervised semantic clustering, yielding average gains of 5-7% on COMET, BERT score, Sentence BLEU, CheXbert-F1 and RadGraph-F1 after fine-tuning, plus further gains when cluster info is added to GRPO rewards.
Motion-MLLM integrates IMU egomotion data into MLLMs using cascaded filtering and asymmetric fusion to ground visual content in physical trajectories for scale-aware 3D understanding, achieving competitive accuracy at higher speed.
A Medical Entity Tree organizes medical knowledge to engineer higher-quality training data that boosts general MLLMs on medical benchmarks.
Robust CLIP models amplify vulnerabilities to natural adversarial scenarios while standard CLIP shows large performance drops on natural language-induced adversarial examples in zero-shot classification, segmentation, and VQA.
U-CESE integrates three CESE modules into a unified clip-based pipeline with DAKE keyframe extraction and ReCap captioning to support consistent multimodal event retrieval across video sources.
citing papers explorer
-
Leveraging Multimodal Large Language Models for All-in-One Image Restoration via a Mixture of Frequency Experts
An MLLM-guided architecture with a mixture of frequency experts and relational alignment loss achieves state-of-the-art all-in-one image restoration, outperforming prior methods by up to 1.35 dB on the CDD11 dataset.
-
DART: Difficulty-Adaptive Routing for Zero-Shot Video Temporal Grounding
DART routes zero-shot video temporal grounding queries by difficulty using DPP entropy, achieving up to 3.5 mIoU gains with 7x fewer frames on Charades-STA and ActivityNet Captions.
-
MemLearner: Learning to Query Context memory for Video World Models
MemLearner introduces a learning-based adaptive context query method using query tokens in video world models to improve long-term scene consistency over rule-based retrieval.
-
Membership Inference Attacks Against Video Large Language Models
A temperature-perturbed black-box attack infers video training membership in VideoLLMs with 0.68 AUC by exploiting sharper generation behavior on member samples.
-
DualTrack: Sensorless 3D Ultrasound needs Local and Global Context
DualTrack uses decoupled local spatiotemporal and global anatomical encoders with a fusion module to estimate probe trajectories from 2D ultrasound sequences, achieving sub-5 mm average reconstruction error on public benchmarks.
-
MMAgent-R$^2$: Learning to Rerank and Reject for Agentic mRAG
An agentic mRAG framework uses GRPO-trained visual reranking and active rejection to verify retrieved candidate entities, achieving state-of-the-art on three KB-VQA benchmarks.
-
Ego-Human Motion Prediction with 3D-Aware LLM
Ego3DLM jointly predicts past and future 3D body pose and motion descriptions in a single autoregressive pass, conditioned on egocentric video, 3D scene features, and three-point tracking, achieving state-of-the-art on the Nymeria benchmark.
-
StochasT: Learning with Stochastic Turn Depth for Visual Instruction Tuning
StochasT uses stochastic clustering of language tasks into varying turn depths for the same image to improve LVLMs on both single-turn and multi-turn scenarios without discarding data.
-
Identifying and Resolving Pitfalls of Knowledge-Based VQA Benchmarks: Auditing, Repairing, and Augmenting
Audit of KB-VQA benchmarks reveals systematic violations of answer derivability, question clarity, and visual disambiguation assumptions, with new repair and multi-entity augmentation protocols producing different model performance trends.
-
MV-GEL: Language-Driven Multi-View Geometric Entity Localization on Meshes
MV-GEL uses a learned view selector and a fine-tuned vision-language segmentation model to localize text-described faces and edges on 3D meshes.
-
WaterGen: Decoupling Scene and Medium in Underwater Image Generation
WaterGen decouples scene generation from medium degradation in a two-stage latent diffusion process to produce controllable realistic underwater images that improve downstream restoration and segmentation.
-
Hi-GaTA: Hierarchical Gated Temporal Aggregation Adapter for Surgical Video Report Generation
Hi-GaTA is a hierarchical gated temporal aggregation adapter that uses short-to-long temporal pyramids and gated fusion to enable surgical video report generation, backed by a new 214-video benchmark and a surgical ViViT pretrained on 40,000 minutes of video.
-
SemEnrich: Self-Supervised Semantic Enrichment of Radiology Reports for Vision-Language Learning
SemEnrich enriches radiology reports with positive/neutral findings via self-supervised semantic clustering, yielding average gains of 5-7% on COMET, BERT score, Sentence BLEU, CheXbert-F1 and RadGraph-F1 after fine-tuning, plus further gains when cluster info is added to GRPO rewards.
-
Feeling the Space: Egomotion-Aware Video Representation for Efficient and Accurate 3D Scene Understanding
Motion-MLLM integrates IMU egomotion data into MLLMs using cascaded filtering and asymmetric fusion to ground visual content in physical trajectories for scale-aware 3D understanding, achieving competitive accuracy at higher speed.
-
Learning from Medical Entity Trees: An Entity-Centric Medical Data Engineering Framework for MLLMs
A Medical Entity Tree organizes medical knowledge to engineer higher-quality training data that boosts general MLLMs on medical benchmarks.
-
Beyond Standard Benchmarks: A Systematic Audit of Vision-Language Model's Robustness to Natural Semantic Variation Across Diverse Tasks
Robust CLIP models amplify vulnerabilities to natural adversarial scenarios while standard CLIP shows large performance drops on natural language-induced adversarial examples in zero-shot classification, segmentation, and VQA.
-
U-CESE: Unified Clip-based Event Search Engine for AI Challenge HCMC 2025
U-CESE integrates three CESE modules into a unified clip-based pipeline with DAKE keyframe extraction and ReCap captioning to support consistent multimodal event retrieval across video sources.