Recent advancement in multimodal LLMs (MLLMs) has demonstrated their remarkable capability to generate descriptive captions for input videos. However, these models suffer from factual inaccuracies in the generated descriptions, causing severe hallucination issues. While prior works have explored alleviating hallucinations for static images, jointly mitigating visual object and temporal action hallucinations for dynamic videos remains a challenging and unsolved task. To tackle this challenge, we propose a Self-Augmented Contrastive Alignment (SANTA) framework for enabling object and action faithfulness by exempting the spurious correlations and enforcing the emphasis on visual facts. SANTA employs a hallucinative self-augmentation scheme to identify the potential hallucinations that lie in the MLLM and transform the original captions to the contrasted negatives. Furthermore, we develop a tracklet-phrase contrastive alignment to match the regional objects and relation-guided actions with their corresponding visual and temporal phrases. Extensive experiments demonstrate that SANTA outperforms existing methods in alleviating object and action hallucinations, yielding superior performance on the hallucination examination benchmarks.

Dense video captioning aims to interpret and describe all temporally localized events throughout an input video. Recent state-of-the-art methods leverage large language models (LLMs) to provide detailed moment descriptions for video data. However, existing VideoLLMs remain challenging in identifying precise event boundaries in untrimmed videos, causing the generated captions to be not properly grounded. In this paper, we propose TA-Prompting, which enhances VideoLLMs via Temporal Anchors that learn to precisely localize events and prompt the VideoLLMs to perform temporal-aware video event understanding. During inference, in order to properly determine the output caption sequence from an arbitrary number of events presented within a video, we introduce an event coherent sampling strategy to select event captions with sufficient coherence across temporal events and cross-modal similarity with the given video. Through extensive experiments on benchmark datasets, we show that our TA-Prompting is favorable against state-of-the-art VideoLLMs, yielding superior performance on dense video captioning and temporal understanding tasks including moment retrieval and temporalQA.

It is challenging to update Large language models (LLMs) since real-world knowledge evolves.While existing Lifelong Knowledge Editing (LKE) methods efficiently update sequentially incoming edits, they often struggle to precisely overwrite the outdated knowledge with the latest one, resulting in conflicts that hinder LLMs from determining the correct answer.To address this Serial Lifelong Knowledge Editing (sLKE) problem, we propose a novel Mixture-of-Knowledge-Experts scheme with an Activation-guided Routing Mechanism (ARM), which assigns specialized experts to store domain-specific knowledge and ensures that each update completely overwrites old information with the latest data.Furthermore, we introduce a novel sLKE benchmark where answers to the same concept are updated repeatedly, to assess the ability of editing methods to refresh knowledge accurately.Experimental results on both LKE and sLKE benchmarks show that our ARM performs favorably against SOTA knowledge editing methods.

Customized text-to-video generation aims to produce high-quality videos that incorporate user-specified subject identities or motion patterns. However, existing methods mainly focus on personalizing a single concept, either subject identity or motion pattern, limiting their effectiveness for multiple subjects with the desired motion patterns. To tackle this challenge, we propose a unified framework Video-Mage for video customization over both multiple subjects and their interactive motions. VideoMage employs subject and motion LoRAs to capture personalized content from user-provided images and videos, along with an appearance-agnostic motion learning approach to disentangle motion patterns from visual appearance. Furthermore, we develop a spatial-temporal composition scheme to guide interactions among subjects within the desired motion patterns. Extensive experiments demonstrate that VideoMage outperforms existing methods, generating coherent, user-controlled videos with consistent subject identities and interactions. Project Page: https://jasper0314-huang.github.io/videomage-customization/

Open-source foundation models have seen rapid adoption and development, enabling powerful general-purpose capabilities across diverse domains. However, fine-tuning large foundation models for domain-specific or personalized tasks remains prohibitively expensive for most users due to the significant memory overhead beyond that of inference. We introduce EMLoC, an Emulator-based Memory-efficient fine-tuning framework with LoRA Correction, which enables model fine-tuning within the same memory budget required for inference. EMLoC constructs a task-specific light-weight emulator using activation-aware singular value decomposition (SVD) on a small downstream calibration set. Fine-tuning then is performed on this lightweight emulator via LoRA. To tackle the misalignment between the original model and the compressed emulator, we propose a novel compensation algorithm to correct the fine-tuned LoRA module, which thus can be merged into the original model for inference. EMLoC supports flexible compression ratios and standard training pipelines, making it adaptable to a wide range of applications. Extensive experiments demonstrate that EMLoC outperforms other baselines across multiple datasets and modalities. Moreover, without quantization, EMLoC enables fine-tuning of a 38B model, which originally required 95GB of memory, on a single 24GB consumer GPU—bringing efficient and practical model adaptation to individual users.

Generating 3D shapes according to specific textual input is a crucial topic in the multimedia application, with its potential enhancement to the VR/AR/XR usage that enables more diverse virtual scenes. Due to the recent success of diffusion models, text-guided 3D object generation has drawn a lot of attention recently. However, current latent diffusion-based methods are restricted to shape-only generation, requiring time-consuming and computationally expensive post-processing to obtain colored objects. In this paper, we propose an end-to-endShape-Color Diffusion Prior framework (SCDiff)to achieve colored text-to-3D object generation. Given a general text description as input, our SCDiff is able to distinguish shape and color-related priors in the text and generate a shape latent and a color latent for a pre-trained 3D object auto-encoder to derive colored 3D objects. Our SCDiff contains two 3D latent diffusion models (LDM), where one generates the shape latent from the input text and the other generates the color latent. To help the two LDMs focus on shape/color-related information, we further adopt a Large Language Model (LLM) to separate the input text into a shape phrase and a color phrase via an in-context learning technique so that our shape/color LDM would not be influenced by irrelevant information. Due to the separation of shape and color latent, we are able to manipulate the color of an object by giving different color phrases while maintaining the original shape. Experiments on a benchmark dataset would quantitatively and qualitatively verify the effectiveness and practicality of our proposed model. As an extension, we show the capability of our SCDiff on 3D object generation and manipulation based on various modality conditions, which further confirms the scalability and applications in multimedia of our proposed framework.

Natural Language Explanation (NLE) in vision and language tasks aims to provide human-understandable explanations for the associated decision-making process. In practice, one might encounter explanations which lack informativeness or contradict visual-grounded facts, known as implausibility and hallucination problems, respectively. To tackle these challenging issues, we consider the task of visual question answering (VQA) and introduce Rapper, a two-stage Reinforced Rationale-Prompted Paradigm. By knowledge distillation, the former stage of Rapper infuses rationale-prompting via large language models (LLMs), encouraging the rationales supported by language-based facts. As for the latter stage, a unique Reinforcement Learning from NLE Feedback (RLNF) is introduced for injecting visual facts into NLE generation. Finally, quantitative and qualitative experiments on two VL-NLE benchmarks show that Rapper surpasses state-of-the-art VQA-NLE methods while providing plausible and faithful NLE.