The paper proposes COVRAG, a depth-based memory retrieval framework that maximizes the coverage of target-view regions to significantly improve long-term geometric consistency in autoregressive long v…

VideoMLA introduces a novel Multi-Head Latent Attention (MLA) mechanism that replaces per-head KV caches with a shared low-rank content latent, significantly reducing memory and improving throughput f…

MORPHOS is a novel autoregressive framework that generates dynamic 3D assets (like meshes and radiance fields) from videos by using a unified 4D representation to ensure temporal consistency and handl…

The paper proposes ST-DRC, a Spatial-Temporal Decoupled Reference Conditioning framework that effectively balances high-level semantic control and low-level identity fidelity for text-to-video generat…

The paper proposes a unified framework that decouples long-video reasoning into semantic and visual evidence, significantly improving performance on the HD-EPIC VQA Challenge.

TunerDiT introduces a training-free progressive steering method to enhance multi-event video generation using Diffusion Transformers, achieving state-of-the-art performance by explicitly managing even…

The paper introduces Knowledge-Intensive Video Generation (KIVI) as a challenging benchmark for evaluating video models on factuality and practical usefulness, showing that current state-of-the-art sy…

The paper proposes a training-free framework, Visual Representation-Guided Video-LLM Reasoning, to perform composed video retrieval by using visual examples and text instructions, achieving strong per…

This study systematically evaluates a wide range of chunking methods for Retrieval-Augmented Generation (RAG) to assess their effectiveness and highlight the overlooked challenges associated with chun…

FLARE is a systematic conversion framework that enables a single checkpoint to support both autoregressive (AR) and diffusion-style parallel decoding for hybrid-attention large language models, achiev…

Moment-KV introduces a novel momentum-based technique to compress the Key-Value (KV) cache during the decoding phase of LLM generation, significantly improving fidelity in long-generation tasks.

The paper proposes Dynamic Adapter Routing (DAR), a novel method that significantly improves continual multimodal retrieval by adaptively selecting and merging specialized adapters.

The paper introduces Moment-Video, a new benchmark that diagnoses the ability of video MLLMs to understand brief, critical visual events, revealing that current models struggle significantly with temp…

InfoMerge is a novel, training-free method that significantly compresses visual tokens for Video-LLMs by estimating temporal redundancy and allocating tokens based on content richness, achieving high…

Lumos-Nexus is a training-efficient framework that enhances video generation quality by progressively bridging generation from a lightweight model to a high-fidelity generator in a shared latent space…

The paper proposes VISTA, a multi-level event semantics mining framework, to accurately predict complex events in long videos, addressing the limitations of current LLMs in this domain.

This paper proposes Tail-Aware Adaptive-k (TAA-k), a training-free framework for adaptive context selection in retrieval-augmented generation systems using Extreme Value Theory.