Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Rongrong JiSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works138 from public data
- MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework2,626
MetaGPT is introduced, an innovative framework that incorporates efficient human workflows as a meta programming approach into LLM-based multi-agent collaboration and leverages the assembly line paradigm to assign diverse roles to various agents, thereby establishing a framework that can effectively and cohesively deconstruct complex multi- agent collaborative problems.
- Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis1,687
Video-MME is introduced, the first-ever full-spectrum, Multi-Modal Evaluation benchmark of MLLMs in Video analysis, and it is revealed that Gemini 1.5 Pro is the best-performing commercial model, significantly outperforming the open-source models with an average accuracy of 75%, compared to 71.9% for GPT-4o.
- VITA: Towards Open-Source Interactive Omni Multimodal LLM204
VITA, the first-ever open-source Multimodal Large Language Model (MLLM) adept at simultaneous processing and analysis of Video, Image, Text, and Audio modalities, and meanwhile has an advanced multimodal interactive experience is introduced.
- 143
- Feast Your Eyes: Mixture-of-Resolution Adaptation for Multimodal Large Language Models120
This work proposes a novel and efficient method for MLLMs, termed Mixture-of-Resolution Adaptation (MRA), which adopts two visual pathways for images with different resolutions, where high-resolution visual information is embedded into the low-resolution pathway via the novel mixture-of-resolution adapters (MR-Adapters).
- Meta Architecture for Point Cloud Analysis115
This paper takes the initiative to explore and propose a unified framework called PointMeta, to which the popular 3D point cloud analysis approaches could fit, and shows very strong performance in efficiency and effectiveness through extensive experiments on challenging benchmarks, and verifies the necessity and benefits of high-level interpretation, contrast, and comparison like PointMeta.
- Multinomial Distribution Learning for Effective Neural Architecture Search115
A Multinomial Distribution Learning for extremely effective NAS is proposed, which considers the search space as a joint multinomial distribution, and the optimal network structure is obtained by the operations with the most likely probability in this distribution.
- FastVGGT: Training-Free Acceleration of Visual Geometry Transformer108
This work proposes FastVGGT, which, for the first time, leverages token merging in the 3D domain through a training-free mechanism for accelerating VGGT, and devise a unique token partitioning strategy tailored to 3D architectures and tasks.
- PAMS: Quantized Super-Resolution via Parameterized Max Scale104
A new quantization scheme termed PArameterized Max Scale (PAMS), which applies the trainable truncated parameter to explore the upper bound of the quantization range adaptively and a structured knowledge transfer (SKT) loss is introduced to fine-tune the quantized network.
- AutoDiffusion: Training-Free Optimization of Time Steps and Architectures for Automated Diffusion Model Acceleration94
This work designs a unified search space that consists of all possible time steps and various architectures and introduces a two stage evolutionary algorithm to find the optimal solution in the designed search space.
- 94
- Towards Optimal Fine Grained Retrieval via Decorrelated Centralized Loss with Normalize-Scale Layer77
A novel metric learning scheme, termed Normalize-Scale Layer and Decorrelated Global Centralized Ranking Loss, which achieves extremely efficient and discriminative learning, i.e., 5× speedup over triplet loss and 12% recall boost on CARS196.
- Cantor: Inspiring Multimodal Chain-of-Thought of MLLM74
It is argued that converging visual context acquisition and logical reasoning is pivotal for tackling visual reasoning tasks and an innovative multimodal CoT framework, termed Cantor, characterized by a perception-decision architecture is proposed, demonstrating the efficacy of the proposed framework.
- AffineQuant: Affine Transformation Quantization for Large Language Models72
This paper advocates for the direct optimization using equivalent Affine transformations in PTQ (AffineQuant) using the corresponding inverse matrix to ensure equivalence between the pre- and post-quantization outputs of PTQ, thereby maintaining its efficiency and generalization capabilities.
- OMPQ: Orthogonal Mixed Precision Quantization72
This work proposes to optimize a proxy metric of network orthogonality that can be efficiently solved with linear programming, which proves to be highly correlated with quantized model accuracy and bit-width and achieves high accuracy without compromise on quantization accuracy.
- Training-free Transformer Architecture Search68
Experimental results demonstrate that the TF- TAS achieves a competitive performance against the state-of-the-art manually or automatically design ViT architectures, and it promotes the searching efficiency in ViT search space greatly.
- Centralized Ranking Loss with Weakly Supervised Localization for Fine-Grained Object Retrieval63
A novel centralized ranking loss (CRL) is introduced, which achieves a very efficient (1,000times training speedup comparing to the triplet loss) and discriminative feature learning by a ?
- Video-MME-v2: Towards the Next Stage in Benchmarks for Comprehensive Video Understanding53
This work introduces Video-MME-v2, a comprehensive benchmark designed to rigorously evaluate the robustness and faithfulness of video understanding, and designs a progressive tri-level hierarchy that incrementally increases the complexity of video comprehension.
- 48
- Knowledge Prompt-tuning for Sequential Recommendation46
This work introduces an external knowledge base and proposes Knowledge Prompt-tuning for Sequential Recommendation (KP4SR), which improves NDCG@5 and HR@5 by 40.65% and 36.65%, and proposes a knowledge tree mask, which restores the data structure in a mask matrix form, thus mitigating the noise problem.
- Solving the Catastrophic Forgetting Problem in Generalized Category Discovery43
This work designs a novel learning approach, LegoGCD, which is seamlessly integrated into previous methods to enhance the discrimination of novel classes while maintaining performance on previously encountered known classes and effectively addresses the known category forgetting issue across all datasets.
- Long-VITA: Scaling Large Multi-modal Models to 1 Million Tokens with Leading Short-Context Accuracy40
An effective multi-modal training schema that starts with large language models and proceeds through vision-language alignment, general knowledge learning, and two sequential stages of long-sequence fine-tuning to serve as a competitive baseline and offer valuable insights for the open-source community in advancing long-context multi-modal understanding.
- An Information Theory-Inspired Strategy for Automated Network Pruning36
This paper introduces a new theorem to illustrate that the hidden representation should compress information with each other to achieve a better generalization, and provides rigorous proof to reveal that optimizing the normalized HSIC simultaneously minimizes the mutual information between different layers.
- 36
- UI-AGILE: Advancing GUI Agents with Effective Reinforcement Learning and Precise Inference-Time Grounding34
UI-AGILE is introduced, a suite of improvements to the Reinforcement Fine-Tuning process to balance planning with speed and grounding accuracy, and a cropping-based resampling strategy to mitigate the sparse reward problem and improve learning on complex tasks.
- 33
- Rethinking Performance Estimation in Neural Architecture Search32
This paper provides a novel yet systematic rethinking of PE in a resource constrained regime, termed budgeted PE (BPE), which precisely and effectively estimates the performance of an architecture sampled from an architecture space.
- DDPNAS: Efficient Neural Architecture Search via Dynamic Distribution Pruning30
This paper proposes an efficient and unified NAS framework termed DDPNAS via dynamic distribution pruning, facilitating a theoretical bound on accuracy and efficiency and directly obtains the optimal neural architectures on given constraints, practical for on-device models across diverse search spaces and constraints.
- Multimodal Quantitative Language for Generative Recommendation29
This work proposes a novel approach called Multimodal Quantitative Language for Generative Recommendation (MQL4GRec), which aims to transform items from different domains and modalities into a unified language, which can serve as a bridge for transferring recommendation knowledge.
- Binarized Neural Architecture Search29
This work introduces channel sampling and operation space reduction into a differentiable NAS to significantly reduce the cost of searching and achieves a performance comparable to NAS on both CIFAR and ImageNet databases.
- Motion-aware Latent Diffusion Models for Video Frame Interpolation28
A novel diffusion framework, Motion-Aware latent Diffusion models (MADiff), which is specifically designed for the VFI task and achieves state-of-the-art performance significantly outperforming existing approaches, especially under challenging scenarios involving dynamic textures with complex motion.
- Solving Oscillation Problem in Post-Training Quantization Through a Theoretical Perspective28
This paper first formulate the oscillation in PTQ and prove the problem is caused by the difference in module capacity, and then tries to solve this problem by introducing a principled and generalized frame-work theoretically.
- A Unified Framework for Soft Threshold Pruning28
This work reformulates soft threshold pruning as an implicit optimization problem solved using the Iterative Shrinkage-Thresholding Algorithm (ISTA), a classic method from the fields of sparse recovery and compressed sensing, and derives a family of pruning methods, including sparsify-during-training, early pruning, and pruning at initialization.
- 27
- Flow caching for autoregressive video generation26
FlowCache is presented, the first caching framework specifically designed for autoregressive video generation, which introduces a chunkwise caching strategy that dynamically adapts to the unique denoising characteristics of each chunk, complemented by a joint importance-redundancy optimized KV cache compression mechanism that maintains fixed memory bounds while preserving generation quality.
- 25
- 24
- You only compress once: Towards effective and elastic BERT compression via exploit–explore stochastic nature gradient23
A novel approach, YOCO-BERT, to achieve compress once and deploy everywhere and proposes a novel stochastic nature gradient optimization method to guide the generation of optimal candidate architecture which could keep a balanced trade-off between explorations and exploitation.
- Event-Diffusion: Event-Based Image Reconstruction and Restoration with Diffusion Models23
It is argued that the key to event-based image reconstruction is to enhance the edge information of objects and restore the artifacts in the reconstructed images, and a new framework which incorporate it into the reconstruction pipeline to obtain high-quality results which effectively remove artifacts and blur in reconstructed images.
- Binarized Neural Architecture Search for Efficient Object Recognition23
Binarized neural architecture search (BNAS), with a search space of binarized convolutions, is introduced to produce extremely compressed models to reduce huge computational cost on embedded devices for edge computing.
- Uncovering the Over-Smoothing Challenge in Image Super-Resolution: Entropy-Based Quantification and Contrastive Optimization21
This paper proposes an explicit solution to the center-oriented optimization (COO) problem, called Detail Enhanced Contrastive Loss (DECLoss), which utilizes the clustering property of contrastive learning to directly reduce the variance of the potential high-resolution distribution and thereby decrease the entropy.
- 20
- Determining Layer-wise Sparsity for Large Language Models Through a Theoretical Perspective19
This paper identifies a critical issue of construction error explosion in existing LLMs sparsification methods and derives a simple yet effective approach to layer-wise sparsity allocation that mitigates this issue, and uses a monotonically increasing arithmetic progression to identify optimal layer-wise sparsity rates.
- GraCo: Granularity-Controllable Interactive Segmentation19
An any-granularity mask generator is designed that exploits the semantic property of the pre-trained IS model to automatically gen-erate abundant mask-granularity pairs without requiring additional manual annotation and a granularity-controllable learning strategy that efficiently imparts the granularity controllability to the IS model is proposed.
- Semi-Supervised Blind Image Quality Assessment through Knowledge Distillation and Incremental Learning18
A unified framework that combines semi-supervised and incremental learning to address the issue of catastrophic forgetting in Blind Image Quality Assessment and demonstrates its potential to tackle the challenges in real-world production processes.
- Dynamic Distribution Pruning for Efficient Network Architecture Search18
A dynamic distribution pruning method towards extremely efficient NAS, which samples architectures from a joint categorical distribution, and the optimal neural architecture is obtained when there is only one structure remained.
- 17
- Bilateral Event Mining and Complementary for Event Stream Super-Resolution16
This paper proposes a bilateral event mining and complementary network (BMCNet) to fully leverage the potential of each event and capture the shared information to complement each other simultaneously and proposes a bilateral information exchange (BIE) module to facilitate the exchange of information between two streams.
- 16
- 14
- 13
- Out-of-Distributed Semantic Pruning for Robust Semi-Supervised Learning13
A unified framework termed OOD Semantic Pruning (OSP), which aims at pruning OOD semantics out from in-distribution (ID) features by designing a soft orthogonality regularization, which first transforms each ID feature by suppressing its semantic component that is collinear with paired OOD sample.
- Wavelet-based Frame Selection by Detecting Semantic Boundary for Long Video Understanding12
Wavelet-based Frame Selection by Detecting Semantic Boundary by Detecting Semantic Boundary (WFS-SB) is introduced, a training-free framework that presents a new perspective on effective video understanding that significantly boosts LVLM performance.
- SocialOmni: Benchmarking Audio-Visual Social Interactivity in Omni Models12
SocialOmni is introduced, an offline diagnostic benchmark that separates three turn-level decisions: identifying who is speaking, deciding when a designated participant should enter at an annotated query time, and determining how that participant should continue the dialogue.
- Discovering Important Experts for Mixture-of-Experts Models Pruning Through a Theoretical Perspective12
Shapley-MoE, an efficient pruning method for MoE models inspired by cooperative game theory, is proposed by quantifying each expert’s contribution via Shapley value, and two novel methods to improve sampling accuracy and efficiency are proposed.
- 12
- 12
- 11
- 11
- 9
- 9
- 8
- Few-Shot Image Quality Assessment via Adaptation of Vision-Language Models8
The Gradient-Regulated MetaPrompt IQA Framework (GRMP-IQA), designed to efficiently adapt the visual-language pre-trained model, CLIP, to IQA tasks, achieving high accuracy even with limited data, is proposed.
- VEGA: Learning Interleaved Image-Text Comprehension in Vision-Language Large Models8
By employing a multi-task, multi-scale post-training strategy, this work sets a robust baseline for MLLMs on the IITC task, attaining an $85.8% accuracy rate in image association and a $0.508$ Rouge score, which validate the effectiveness of the dataset in improving MLLMs capabilities for nuanced image-text comprehension.
- A Unified Framework for 3D Point Cloud Visual Grounding8
This paper proposes Superpoint Mask Branch, which serves a dual purpose and eliminates the heavy computational overhead on the high-resolution visual features for upsampling, and enables 3DRefTR to achieve both well-performing 3DRES and3DREC capacities with only a 6% additional latency compared to the original 3DREC model.
- Event-Anchored Frame Selection for Effective Long-Video Understanding7
Event-Anchored Frame Selection (EFS), a hierarchical, event-aware pipeline that partitions the video stream into visually homogeneous temporal segments, which serve as proxies for semantic events and ensures the final keyframe set jointly optimizes for event coverage, query relevance, and visual diversity.
- EBFT: Effective and Block-Wise Fine-Tuning for Sparse LLMs7
This work proposes an efficient and fast framework for fine-tuning sparse LLMs based on minimizing reconstruction error that achieves superior results on various benchmarks, and can be executed on a single 16GB GPU.
- DLIP: Distilling Language-Image Pre-training7
DLIP, a simple yet efficient Distilling Language-Image Pre-training framework, is presented, through which it is investigated how to distill a light VLP model, and experimental results reveal that DLIP can achieve a state-of-the-art accuracy/efficiency trade-off across diverse cross-modal tasks.
- Digest: Deeply Supervised Knowledge Transfer Network Learning for Brain Tumor Segmentation with Incomplete Multi-Modal MRI Scans7
A knowledge transfer learning frame is constructed, enabling a student model to learn modality-shared semantic information from a teacher model pretrained with the complete multi-modal MRI data, and a deeply supervised knowledge transfer loss is designed to ensure the consistency of the teacher-student structure at different decoding stages.
- Motion-Aware Caching for Efficient Autoregressive Video Generation6
MotionCache is proposed, a motion-aware cache framework that exploits inter-frame differences as a lightweight proxy for pixel-level motion characteristics and employs a coarse-to-fine strategy: an initial warm-up phase establishes semantic coherence, followed by motion-weighted cache reuse that dynamically adjusts update frequencies per token.
- Multimodal Inplace Prompt Tuning for Open-set Object Detection6
This study introduces a redundancy assessment metric to identify uniform attention patterns and incorporates multimodal inplace prompt tuning (MIPT) to enrich the text prompt with visual clues in areas with high redundancy.
- Depth-Guided Semi-Supervised Instance Segmentation6
This work proposes Depth Feature Fusion, which integrates features extracted from depth estimation into the SSIS process, and establishes a new benchmark for SSIS, outperforming previous methods.
- Probabilistic Modeling: Proving the Lottery Ticket Hypothesis in Spiking Neural Network6
This work proposes a novel probabilistic modeling approach for spiking neurons with complicated spatio-temporal dynamics and theoretically and experimentally proves that LTH holds in Spiking Neural Networks (SNNs), and concludes that pruning directly in accordance with the weight size in existing SNNs is clearly not optimal.
- Adaptive Feature Selection for No-Reference Image Quality Assessment by Mitigating Semantic Noise Sensitivity6
A Quality-Aware Feature Matching IQA metric (QFM-IQM) is proposed that employs contrastive learning to remove harmful features from the upstream task and achieves superior performance to the state-of-the-art NR-IQA.
- AlphaQ: Calibration-Free Bit Allocation for Mixture-of-Experts Quantization5
AlphaQ is proposed, a calibration-free bit-allocation method for MoE quantization that draws on HT-SR theory and operationalizes this principle by measuring expert-wise spectral heavy-tailedness and solving a budget-constrained optimization problem that minimizes total quantization error under a global bit-budget constraint.
- From Objects to Events: Unlocking Complex Visual Understanding in Object Detectors Via LLM-guided Symbolic Reasoning5
A novel framework that expands the capability of standard object detectors beyond mere object recognition to complex event understanding through LLM-guided symbolic reasoning, bridging the semantic gap between object detection and event understanding without requiring expensive taskspecific training.
- Aligning Instance Brownian Bridge with Texts for Open-Vocabulary Video Instance Segmentation5
BridgeText Alignment (BTA) is proposed to link frame-level instance representations as a Brownian Bridge to link frame-level instance representations as a Brownian Bridge and surpasses the SOTA (OV2Seg) by a clear margin.
- Benchmarking Abstract and Reasoning Abilities Through A Theoretical Perspective5
This paper designs a benchmark: systematic symbol remapping in rule-based tasks, which forces models to demonstrate genuine pattern recognition beyond superficial token matching, and introduces two novel complementary metrics: \(\scoreGamma\) measures basic reasoning accuracy and \(\scoreDelta\) quantifies a model's reliance on specific symbols rather than underlying patterns.
- Multi-Modal Prompt Learning on Blind Image Quality Assessment5
This paper introduces an innovative multi-modal prompt-based methodology for IQA that employs carefully crafted prompts that synergistically mine incremental semantic information from both visual and linguistic data.
- 5
- 5
- Mix-Ecom: Towards Mixed-Type E-Commerce Dialogues with Complex Domain Rules4
This work introduces a novel corpus, termed Mix-ECom, which is constructed based on real-world customer-service dialogues with post-processing to remove user privacy and add CoT process and proposes a dynamic framework to further improve the performance.
- Instance Brownian Bridge as Texts for Open-vocabulary Video Instance Segmentation4
This work builds a frozen video segmentor to generate frame-level instance queries, and design Temporal Instance Resampler (TIR) to generate queries with temporal context from frame queries to link frame-level instance representations as a Brownian Bridge to model instance dynamics and align bridge-level instance representation to class texts.
- 4
- 3
- 3
- Adaptive Fuzzy Positive Learning for Annotation-Scarce Semantic Segmentation3
Adaptive Fuzzy Positive Learning (A-FPL) is proposed for correctly learning unlabeled data in a plug-and-play fashion, targeting adaptively encouraging fuzzy positive predictions and suppressing highly probable negatives.
- 3
- LocLoc: Low-level Cues and Local-area Guides for Weakly Supervised Object Localization3
This paper introduces a GrabCut-Enhanced Generator to learn global semantic representations for localization based on graph cuts to enhance low-level information based on long-range dependencies captured by the transformer and designs a Local Feature Digging Module (LFDM) that utilizes low- level cues to guide the learning route of local feature representations for accurate classification.
- 3
- SpecEyes: Accelerating Agentic Multimodal LLMs via Speculative Perception and Planning2
This work proposes SpecEyes, an agentic-level speculative acceleration framework that breaks this sequential bottleneck in agentic multimodal large language models and introduces a cognitive gating mechanism based on answer separability, which quantifies the model's confidence for self-verification without requiring oracle labels.
- OmniView-Space: Reinforcing Spatial Reasoning via Multi-Perspective Spatial Mapping2
OmniView-Space is a framework designed to maintain spatial consistency through multimodal egocentric evidence and achieves state-of-the-art performance on single- and multi-image spatial reasoning benchmarks, while reducing reliance on external geometry pipelines.
- WaveZip: Wavelet-Driven Space-Time Decoupling for Video Token Condensation2
WaveZip is proposed, a joint signal-frequency-domain framework for efficient video inference that requires no task-specific training and can be seamlessly integrated into off-the-shelf LVLMs to boost inference efficiency.
- QuoTA: Query-oriented Token Assignment via CoT Query Decouple for Long Video Comprehension2
QuoTA strategically allocates frame-level importance scores based on query relevance, enabling one-time visual token assignment before cross-modal interactions in decoder layers, and decouple the query through Chain-of-Thoughts reasoning to facilitate more precise LVLM-based frame importance scoring.
- Towards Efficient Automatic Self-Pruning of Large Language Models2
This work introduces an end-to-end automatic self-pruning framework for LLMs, which efficiently search layer-wise pruning rates and leverages LLMs to autonomously execute the entire evolutionary search process to search for pruning rate configurations.
- Dynamic Clustering Convolutional Neural Network2
The extensive experiments demonstrate that the proposed DCCNeXt outperforms the mainstream Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), Vision Multi-layer Perceptrons (MLPs), Vision Graph Neural Networks (GNNs), and Vision Mambas.
- Watch Where You Move: Region-Aware Dynamic Aggregation and Excitation for Gait Recognition2
A Region-aware Dynamic Aggregation and Excitation framework (GaitRDAE) that automatically searches for motion regions, assigns adaptive temporal scales and applies corresponding attention, and achieves state-of-the-art performance on several benchmark datasets.
- 2
- Fast Hybrid Search for Automatic Model Compression2
A hybrid search method, searching for the pruned model architecture and the pruning algorithm at the same time, which automatically finds the pruning ratio and pruning algorithm for each convolution layer, which achieves a clear improvement in performance compared to current state-of-the-art methods.
- 2
- 2
- Searching Lightweight Neural Network for Image Signal Processing2
An automated search framework is proposed that derives ISP models with high image quality while satisfying the low-computation requirement and adopts the weight-sharing strategy by introducing a supernet and decouple the architecture search into two stages, supernet training and hard-aware evolutionary search.
- Iterative Data Refinement for Self-Supervised MR Image Reconstruction2
Experimental results show that without utilizing any fully sampled MRI data, the proposed self-supervised framework possesses strong capabilities in capturing image details and structures at high acceleration factors.
- 1
- A2RBench: An Automatic Paradigm for Formally Verifiable Abstract Reasoning Benchmark Generation1
An automated pipeline named A2RBench is introduced, encompassing generation, expansion, evaluation, and analysis, and it is proved that programmatic verification--testing whether the inverse operation perfectly reverses the forward operation (cycle consistency)--guarantees a unique solution.
- OmniScope: Modality-Decoupled Token Compression for Omnimodal Large Language Models1
This work proposes OmniScope, a training-free token compression framework that uses the query as a shared semantic anchor while estimating relevance separately for audio and video, and suggests a simple design principle for OmniLLM inference: share the query across modalities, but not the salience estimates.
- 1
- Distilling Structural Knowledge for Platform-Aware Semantic Segmentation1
It is found that the student network with a more similar architecture against the teacher network obtains more performance gain from distillation, and a novel Structural Distillation (SD) method is proposed which introduces the structural similarity constraints into vanilla knowledge distillation.
- Probabilistic Automated Model Compression via Representation Mutual Information Optimization1
This work introduces a probabilistic framework for automated model compression (Prob-AMC) that optimizes pruning, quantization, and knowledge distillation simultaneously using information theory, and achieves a superior compression ratio on CIFAR-10 and ImageNet.
- Adaptive PromptNet for Auxiliary Glioma Diagnosis Without Contrast-Enhanced MRI1
An adaptive PromptNet for glioma grading utilizing only non-enhanced MRI data has been constructed, and competitive gliomas grading performance on NE-MRI data is achieved.
- –
- –
- –
- –
- Q-DeepSight: Incentivizing Thinking with Images for Image Quality Assessment and Refinement–
Q-DeepSight is proposed, a think-with-image framework that achieves state-of-the-art performance across diverse benchmarks, including natural, restored, and AI-generated content and demonstrates its practical value with Perceptual-in-Generation (PiG), a training-free framework where Q-DeepSight's diagnoses guide iterative image enhancement, effectively closing the loop between assessment and refinement.
- One Ranking, Any Budget: Matryoshka Evidence-to-Context Frame Selection for Long-Video Understanding–
- –
- –
- TimePLE: Rethinking Temporal Representation for Video Temporal Grounding–
TimePLE is proposed, which reformulates VTG from endpoint prediction to interval-native grounding by predicting a single joint distribution over valid temporal intervals, and curate 90K-scale grounded samples and human-verify 3K-scale benchmark annotations.
- Symbolic-driven agentic reasoning for environmental and behavioral event detection–
SDAR employs a symbolic reasoning engine to guide agentic decision making, connecting low-level visual cues with structured symbolic representations of events, and enables interpretable reasoning chains that capture causal relationships, contextual dependencies, and event categories.
- –
- –
- –
- –
- –
- ALGOGEN: Tool-Generated Verifiable Traces for Reliable Algorithm Visualization–
The results demonstrate that the decoupling paradigm effectively mitigates LLM hallucinations in complex AV tasks, providing a more reliable solution for automated generation of high-quality algorithm visualizations.
- –
- –
- –
- –
- –
- –
- Rethinking 3D Dense Caption and Visual Grounding in A Unified Framework through Prompt-based Localization–
Inspired by DETR, this work proposes a unified framework, 3DGCTR, to jointly solve these two distinct but closely related tasks in an end-to-end fashion, and facilitates simultaneous multi-task training on both tasks, mutually enhancing their performance.
- –
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.