We present MM-Narrator, a novel system leveraging GPT-4 with multimodal in-context learning for the generation of audio descriptions (AD). Unlike previous methods that primarily focused on downstream fine-tuning with short video clips, MM-Narrator excels in generating precise audio descriptions for videos of extensive lengths, even beyond hours, in an autoregressive manner. This capability is made possible by the proposed memory-augmented generation process, which effectively utilizes both the short-term textual context and long-term visual memory through an efficient register-and-recall mechanism. These contextual memories compile pertinent past information, including storylines and character identities, ensuring an accurate tracking and depicting of story-coherent and character-centric audio descriptions. Maintaining the training-free design of MM-Narrator, we further propose a complexity-based demonstration selection strategy to largely enhance its multi-step reasoning capability via few-shot multimodal in-context learning (MM-ICL). Experimental results on MAD-eval dataset demonstrate that MM-Narrator consistently outperforms both the existing fine-tuning-based approaches and LLM-based approaches in most scenarios, as measured by standard evaluation metrics. Additionally, we introduce the first segment-based evaluator for recurrent text generation. Empowered by GPT-4, this evaluator comprehensively reasons and marks AD generation performance in various extendable dimensions.
相關內容
We propose Diffusion Inference-Time T-Optimization (DITTO), a general-purpose frame-work for controlling pre-trained text-to-music diffusion models at inference-time via optimizing initial noise latents. Our method can be used to optimize through any differentiable feature matching loss to achieve a target (stylized) output and leverages gradient checkpointing for memory efficiency. We demonstrate a surprisingly wide-range of applications for music generation including inpainting, outpainting, and looping as well as intensity, melody, and musical structure control - all without ever fine-tuning the underlying model. When we compare our approach against related training, guidance, and optimization-based methods, we find DITTO achieves state-of-the-art performance on nearly all tasks, including outperforming comparable approaches on controllability, audio quality, and computational efficiency, thus opening the door for high-quality, flexible, training-free control of diffusion models. Sound examples can be found at //DITTO-Music.github.io/web/.
The video-language (VL) pretraining has achieved remarkable improvement in multiple downstream tasks. However, the current VL pretraining framework is hard to extend to multiple modalities (N modalities, N>=3) beyond vision and language. We thus propose LanguageBind, taking the language as the bind across different modalities because the language modality is well-explored and contains rich semantics. Specifically, we freeze the language encoder acquired by VL pretraining, then train encoders for other modalities with contrastive learning. As a result, all modalities are mapped to a shared feature space, implementing multi-modal semantic alignment. While LanguageBind ensures that we can extend VL modalities to N modalities, we also need a high-quality dataset with alignment data pairs centered on language. We thus propose VIDAL-10M with Video, Infrared, Depth, Audio and their corresponding Language, naming as VIDAL-10M. In our VIDAL-10M, all videos are from short video platforms with complete semantics rather than truncated segments from long videos, and all the video, depth, infrared, and audio modalities are aligned to their textual descriptions. LanguageBind has achieved superior performance on a wide range of 15 benchmarks covering video, audio, depth, and infrared. Moreover, multiple experiments have provided evidence for the effectiveness of LanguageBind in achieving indirect alignment and complementarity among diverse modalities. Code address: //github.com/PKU-YuanGroup/LanguageBind
Mixed-reality (MR) soundscapes blend real-world sound with virtual audio from hearing devices, presenting intricate auditory information that is hard to discern and differentiate. This is particularly challenging for blind or visually impaired individuals, who rely on sounds and descriptions in their everyday lives. To understand how complex audio information is consumed, we analyzed online forum posts within the blind community, identifying prevailing challenges, needs, and desired solutions. We synthesized the results and proposed Sound Unblending for increasing MR sound awareness, which includes six sound manipulations: Ambience Builder, Feature Shifter, Earcon Generator, Prioritizer, Spatializer, and Stylizer. To evaluate the effectiveness of sound unblending, we conducted a user study with 18 blind participants across three simulated MR scenarios, where participants identified specific sounds within intricate soundscapes. We found that sound unblending increased MR sound awareness and minimized cognitive load. Finally, we developed three real-world example applications to demonstrate the practicality of sound unblending.
Recent advancements in deep learning-based image compression are notable. However, prevalent schemes that employ a serial context-adaptive entropy model to enhance rate-distortion (R-D) performance are markedly slow. Furthermore, the complexities of the encoding and decoding networks are substantially high, rendering them unsuitable for some practical applications. In this paper, we propose two techniques to balance the trade-off between complexity and performance. First, we introduce two branching coding networks to independently learn a low-resolution latent representation and a high-resolution latent representation of the input image, discriminatively representing the global and local information therein. Second, we utilize the high-resolution latent representation as conditional information for the low-resolution latent representation, furnishing it with global information, thus aiding in the reduction of redundancy between low-resolution information. We do not utilize any serial entropy models. Instead, we employ a parallel channel-wise auto-regressive entropy model for encoding and decoding low-resolution and high-resolution latent representations. Experiments demonstrate that our method is approximately twice as fast in both encoding and decoding compared to the parallelizable checkerboard context model, and it also achieves a 1.2% improvement in R-D performance compared to state-of-the-art learned image compression schemes. Our method also outperforms classical image codecs including H.266/VVC-intra (4:4:4) and some recent learned methods in rate-distortion performance, as validated by both PSNR and MS-SSIM metrics on the Kodak dataset.
The HaliVer tool integrates deductive verification into the popular scheduling language Halide, used for image processing pipelines and array computations. HaliVer uses Vercors, a separation logic-based verifier, to verify the correctness of (1) the Halide algorithms and (2) the optimised parallel code produced by \halide when an optimisation schedule is applied to the algorithm. This allows proving complex, optimised code correct while reducing the effort to provide the required verification annotations. For both approaches, the same specification is used. We evaluated the tool on several optimised programs generated from characteristic Halide algorithms, using all but one of the essential scheduling directives available in Halide. Without annotation effort, Haliver proves memory safety in almost all programs. With annotations Haliver, additionally, proves functional correctness properties. We show that the approach is viable and reduces the manual annotation effort by an order of magnitude.
As a critical clue of video super-resolution (VSR), inter-frame alignment significantly impacts overall performance. However, accurate pixel-level alignment is a challenging task due to the intricate motion interweaving in the video. In response to this issue, we introduce a novel paradigm for VSR named Semantic Lens, predicated on semantic priors drawn from degraded videos. Specifically, video is modeled as instances, events, and scenes via a Semantic Extractor. Those semantics assist the Pixel Enhancer in understanding the recovered contents and generating more realistic visual results. The distilled global semantics embody the scene information of each frame, while the instance-specific semantics assemble the spatial-temporal contexts related to each instance. Furthermore, we devise a Semantics-Powered Attention Cross-Embedding (SPACE) block to bridge the pixel-level features with semantic knowledge, composed of a Global Perspective Shifter (GPS) and an Instance-Specific Semantic Embedding Encoder (ISEE). Concretely, the GPS module generates pairs of affine transformation parameters for pixel-level feature modulation conditioned on global semantics. After that, the ISEE module harnesses the attention mechanism to align the adjacent frames in the instance-centric semantic space. In addition, we incorporate a simple yet effective pre-alignment module to alleviate the difficulty of model training. Extensive experiments demonstrate the superiority of our model over existing state-of-the-art VSR methods.
Text-video retrieval is a critical multi-modal task to find the most relevant video for a text query. Although pretrained models like CLIP have demonstrated impressive potential in this area, the rising cost of fully finetuning these models due to increasing model size continues to pose a problem. To address this challenge, prompt tuning has emerged as an alternative. However, existing works still face two problems when adapting pretrained image-text models to downstream video-text tasks: (1) The visual encoder could only encode frame-level features and failed to extract global-level general video information. (2) Equipping the visual and text encoder with separated prompts failed to mitigate the visual-text modality gap. To this end, we propose DGL, a cross-modal Dynamic prompt tuning method with Global-Local video attention. In contrast to previous prompt tuning methods, we employ the shared latent space to generate local-level text and frame prompts that encourage inter-modal interaction. Furthermore, we propose modeling video in a global-local attention mechanism to capture global video information from the perspective of prompt tuning. Extensive experiments reveal that when only 0.67% parameters are tuned, our cross-modal prompt tuning strategy DGL outperforms or is comparable to fully finetuning methods on MSR-VTT, VATEX, LSMDC, and ActivityNet datasets. Code will be available at //github.com/knightyxp/DGL
To enhance the performance and effect of AR/VR applications and visual assistance and inspection systems, visual simultaneous localization and mapping (vSLAM) is a fundamental task in computer vision and robotics. However, traditional vSLAM systems are limited by the camera's narrow field-of-view, resulting in challenges such as sparse feature distribution and lack of dense depth information. To overcome these limitations, this paper proposes a 360ORB-SLAM system for panoramic images that combines with a depth completion network. The system extracts feature points from the panoramic image, utilizes a panoramic triangulation module to generate sparse depth information, and employs a depth completion network to obtain a dense panoramic depth map. Experimental results on our novel panoramic dataset constructed based on Carla demonstrate that the proposed method achieves superior scale accuracy compared to existing monocular SLAM methods and effectively addresses the challenges of feature association and scale ambiguity. The integration of the depth completion network enhances system stability and mitigates the impact of dynamic elements on SLAM performance.
We advance the field of Parameter-Efficient Fine-Tuning (PEFT) with our novel multi-adapter method, OrchMoE, which capitalizes on modular skill architecture for enhanced forward transfer in neural networks. Unlike prior models that depend on explicit task identification inputs, OrchMoE automatically discerns task categories, streamlining the learning process. This is achieved through an integrated mechanism comprising an Automatic Task Classification module and a Task-Skill Allocation module, which collectively deduce task-specific classifications and tailor skill allocation matrices. Our extensive evaluations on the 'Super Natural Instructions' dataset, featuring 1,600 diverse instructional tasks, indicate that OrchMoE substantially outperforms comparable multi-adapter baselines in terms of both performance and sample utilization efficiency, all while operating within the same parameter constraints. These findings suggest that OrchMoE offers a significant leap forward in multi-task learning efficiency.
Vision-language foundation models, represented by Contrastive language-image pre-training (CLIP), have gained increasing attention for jointly understanding both vision and textual tasks. However, existing approaches primarily focus on training models to match global image representations with textual descriptions, thereby overlooking the critical alignment between local regions and corresponding text tokens. This paper extends CLIP with multi-granularity alignment. Notably, we deliberately construct a new dataset comprising pseudo annotations at various levels of granularities, encompassing image-level, region-level, and pixel-level captions/tags. Accordingly, we develop a unified multi-granularity learning framework, named UMG-CLIP, that simultaneously empowers the model with versatile perception abilities across different levels of detail. Equipped with parameter efficient tuning, UMG-CLIP surpasses current widely used CLIP models and achieves state-of-the-art performance on diverse image understanding benchmarks, including open-world recognition, retrieval, semantic segmentation, and panoptic segmentation tasks. We hope UMG-CLIP can serve as a valuable option for advancing vision-language foundation models.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191