Video generation is a rapidly advancing research area, garnering significant attention due to its broad range of applications. One critical aspect of this field is the generation of long-duration videos, which presents unique challenges and opportunities. This paper presents the first survey of recent advancements in long video generation and summarises them into two key paradigms: divide and conquer temporal autoregressive. We delve into the common models employed in each paradigm, including aspects of network design and conditioning techniques. Furthermore, we offer a comprehensive overview and classification of the datasets and evaluation metrics which are crucial for advancing long video generation research. Concluding with a summary of existing studies, we also discuss the emerging challenges and future directions in this dynamic field. We hope that this survey will serve as an essential reference for researchers and practitioners in the realm of long video generation.
Nature is infinitely resolution-free. In the context of this reality, existing diffusion models, such as Diffusion Transformers, often face challenges when processing image resolutions outside of their trained domain. To overcome this limitation, we present the Flexible Vision Transformer (FiT), a transformer architecture specifically designed for generating images with unrestricted resolutions and aspect ratios. Unlike traditional methods that perceive images as static-resolution grids, FiT conceptualizes images as sequences of dynamically-sized tokens. This perspective enables a flexible training strategy that effortlessly adapts to diverse aspect ratios during both training and inference phases, thus promoting resolution generalization and eliminating biases induced by image cropping. Enhanced by a meticulously adjusted network structure and the integration of training-free extrapolation techniques, FiT exhibits remarkable flexibility in resolution extrapolation generation. Comprehensive experiments demonstrate the exceptional performance of FiT across a broad range of resolutions, showcasing its effectiveness both within and beyond its training resolution distribution. Repository available at https://github.com/whlzy/FiT.
Humans generally acquire new skills without compromising the old; however, the opposite holds for Large Language Models (LLMs), e.g., from LLaMA to CodeLLaMA. To this end, we propose a new post-pretraining method for LLMs with an expansion of Transformer blocks. We tune the expanded blocks using only new corpus, efficiently and effectively improving the model's knowledge without catastrophic forgetting. In this paper, we experiment on the corpus of code and math, yielding LLaMA Pro-8.3B, a versatile foundation model initialized from LLaMA2-7B, excelling in general tasks, programming, and mathematics. LLaMA Pro and its instruction-following counterpart (LLaMA Pro-Instruct) achieve advanced performance among various benchmarks, demonstrating superiority over existing open models in the LLaMA family and the immense potential of reasoning and addressing diverse tasks as an intelligent agent. Our findings provide valuable insights into integrating natural and programming languages, laying a solid foundation for developing advanced language agents that operate effectively in various environments.
Foundation models have achieved great advances in multi-task learning with a unified interface of unimodal and multimodal tasks. However, the potential of such multi-task learners has not been exploited during transfer learning. In this work, we present a universal parameter-efficient transfer learning method, termed Predict-Interpolate Tuning ($\pi$-Tuning), for vision, language, and vision-language tasks. It aggregates the parameters of lightweight task-specific experts learned from similar tasks to aid the target downstream task. The task similarities are predicted in a unified modality-independent space, yielding a scalable graph to demonstrate task relationships. $\pi$-Tuning has several appealing benefits. First, it flexibly explores both intra- and inter-modal transferability between similar tasks to improve the accuracy and robustness of transfer learning, especially in data-scarce scenarios. Second, it offers a systematical solution for transfer learning with multi-task prediction-and-then-interpolation, compatible with diverse types of parameter-efficient experts, such as prompt and adapter. Third, an extensive study of task-level mutual benefits on 14 unimodal and 6 multimodal datasets shows that $\pi$-Tuning surpasses fine-tuning and other parameter-efficient transfer learning methods both in full-shot and low-shot regimes. The task graph also enables an in-depth interpretable analysis of task transferability across modalities.
Photos serve as a way for humans to record what they experience in their daily lives, and they are often regarded as trustworthy sources of information. However, there is a growing concern that the advancement of artificial intelligence (AI) technology may produce fake photos, which can create confusion and diminish trust in photographs. This study aims to answer the question of whether the current state-of-the-art AI-based visual content generation models can consistently deceive human eyes and convey false information. By conducting a high-quality quantitative study with fifty participants, we reveal, for the first time, that humans cannot distinguish between real photos and AI-created fake photos to a significant degree 38.7%. Our study also finds that an individual's background, such as their gender, age, and experience with AI-generated content (AIGC), does not significantly affect their ability to distinguish AI-generated images from real photographs. However, we do observe that there tend to be certain defects in AI-generated images that serve as cues for people to distinguish between real and fake photos. We hope that our study can raise awareness of the potential risks of AI-generated images and encourage further research to prevent the spread of false information. From a positive perspective, AI-generated images have the potential to revolutionize various industries and create a better future for humanity if they are used and regulated properly.
Diffusion models have attained impressive visual quality for image synthesis. However, how to interpret and manipulate the latent space of diffusion models has not been extensively explored. Prior work diffusion autoencoders encode the semantic representations into a semantic latent code, which fails to reflect the rich information of details and the intrinsic feature hierarchy. To mitigate those limitations, we propose Hierarchical Diffusion Autoencoders (HDAE) that exploit the fine-grained-to-abstract and lowlevel-to-high-level feature hierarchy for the latent space of diffusion models. The hierarchical latent space of HDAE inherently encodes different abstract levels of semantics and provides more comprehensive semantic representations. In addition, we propose a truncated-feature-based approach for disentangled image manipulation. We demonstrate the effectiveness of our proposed approach with extensive experiments and applications on image reconstruction, style mixing, controllable interpolation, detail-preserving and disentangled image manipulation, and multi-modal semantic image synthesis.
The purpose of image inpainting is to recover scratches and damaged areas using context information from remaining parts. In recent years, thanks to the resurgence of convolutional neural networks (CNNs), image inpainting task has made great breakthroughs. However, most of the work consider insufficient types of mask, and their performance will drop dramatically when encountering unseen masks. To combat these challenges, we propose a simple yet general method to solve this problem based on the LaMa image inpainting framework, dubbed GLaMa. Our proposed GLaMa can better capture different types of missing information by using more types of masks. By incorporating more degraded images in the training phase, we can expect to enhance the robustness of the model with respect to various masks. In order to yield more reasonable results, we further introduce a frequency-based loss in addition to the traditional spatial reconstruction loss and adversarial loss. In particular, we introduce an effective reconstruction loss both in the spatial and frequency domain to reduce the chessboard effect and ripples in the reconstructed image. Extensive experiments demonstrate that our method can boost the performance over the original LaMa method for each type of mask on FFHQ, ImageNet, Places2 and WikiArt dataset. The proposed GLaMa was ranked first in terms of PSNR, LPIPS and SSIM in the NTIRE 2022 Image Inpainting Challenge Track 1 Unsupervised.
The tactile sensing capabilities of human hands are essential in performing daily activities. Simultaneously perceiving normal and shear forces via the mechanoreceptors integrated into the hands enables humans to achieve daily tasks like grasping delicate objects. In this paper, we design and fabricate a novel biomimetic tactile sensor with skin-like heterogeneity that perceives normal and shear contact forces simultaneously. It mimics the multilayers of mechanoreceptors by combining an extrinsic layer (piezoresistive sensors) and an intrinsic layer (a Hall sensor) so that it can perform estimation of contact force directions, locations, and joint-level torque. By integrating our sensors, a robotic gripper can obtain contact force feedback at fingertips; accordingly, robots can perform challenging tasks, such as tweezers usage, and egg grasping. This insightful sensor design can be customized and applied in different areas of robots and provide them with heterogeneous force sensing, potentially supporting robotics in acquiring skin-like tactile feedback.