Classifier-free guidance (CFG) is crucial for improving both generation quality and alignment between the input condition and final output in diffusion models. While a high guidance scale is generally required to enhance these aspects, it also causes oversaturation and unrealistic artifacts. In this paper, we revisit the CFG update rule and introduce modifications to address this issue. We first decompose the update term in CFG into parallel and orthogonal components with respect to the conditional model prediction and observe that the parallel component primarily causes oversaturation, while the orthogonal component enhances image quality. Accordingly, we propose down-weighting the parallel component to achieve high-quality generations without oversaturation. Additionally, we draw a connection between CFG and gradient ascent and introduce a new rescaling and momentum method for the CFG update rule based on this insight. Our approach, termed adaptive projected guidance (APG), retains the quality-boosting advantages of CFG while enabling the use of higher guidance scales without oversaturation. APG is easy to implement and introduces practically no additional computational overhead to the sampling process. Through extensive experiments, we demonstrate that APG is compatible with various conditional diffusion models and samplers, leading to improved FID, recall, and saturation scores while maintaining precision comparable to CFG, making our method a superior plug-and-play alternative to standard classifier-free guidance.

Classifier-free guidance (CFG) has become the standard method for enhancing the quality of conditional diffusion models. However, employing CFG requires either training an unconditional model alongside the main diffusion model or modifying the training procedure by periodically inserting a null condition. There is also no clear extension of CFG to unconditional models. In this paper, we revisit the core principles of CFG and introduce a new method, independent condition guidance (ICG), which provides the benefits of CFG without the need for any special training procedures. Our approach streamlines the training process of conditional diffusion models and can also be applied during inference on any pre-trained conditional model. Additionally, by leveraging the time-step information encoded in all diffusion networks, we propose an extension of CFG, called time-step guidance (TSG), which can be applied to *any* diffusion model, including unconditional ones. Our guidance techniques are easy to implement and have the same sampling cost as CFG. Through extensive experiments, we demonstrate that ICG matches the performance of standard CFG across various conditional diffusion models. Moreover, we show that TSG improves generation quality in a manner similar to CFG, without relying on any conditional information.

Diffusion models have emerged as the leading approach for image synthesis, demonstrating exceptional photorealism and diversity. However, training diffusion models at high resolutions remains computationally prohibitive, and existing zero-shot generation techniques for synthesizing images beyond training resolutions often produce artifacts, including object duplication and spatial incoherence. In this paper, we introduce HiWave, a training-free, zero-shot approach that substantially enhances visual fidelity and structural coherence in ultra-high-resolution image synthesis using pretrained diffusion models. Our method employs a two-stage pipeline: generating a base image from the pretrained model followed by a patch-wise DDIM inversion step and a novel wavelet-based detail enhancer module. Specifically, we first utilize inversion methods to derive initial noise vectors that preserve global coherence from the base image. Subsequently, during sampling, our wavelet-domain detail enhancer retains low-frequency components from the base image to ensure structural consistency, while selectively guiding high-frequency components to enrich fine details and textures. Extensive evaluations using Stable Diffusion XL demonstrate that HiWave effectively mitigates common visual artifacts seen in prior methods, achieving superior perceptual quality. A user study confirmed HiWave’s performance, where it was preferred over the state-of-the-art alternative in more than 80% of comparisons, highlighting its effectiveness for high-quality, ultra-high-resolution image synthesis without requiring retraining or architectural modifications.

While conditional diffusion models are known to have good coverage of the data distribution, they still face limitations in output diversity, particularly when sampled with a high classifier-free guidance scale for optimal image quality or when trained on small datasets. We attribute this problem to the role of the conditioning signal in inference and offer an improved sampling strategy for diffusion models that can increase generation diversity, especially at high guidance scales, with minimal loss of sample quality. Our sampling strategy anneals the conditioning signal by adding scheduled, monotonically decreasing Gaussian noise to the conditioning vector during inference to balance diversity and condition alignment. Our Condition-Annealed Diffusion Sampler (CADS) can be used with any pretrained model and sampling algorithm, and we show that it boosts the diversity of diffusion models in various conditional generation tasks. Further, using an existing pretrained diffusion model, CADS achieves a new state-of-the-art FID of 1.70 and 2.31 for class-conditional ImageNet generation at 256$\times$256 and 512$\times$512 respectively.

Advances in latent diffusion models (LDMs) have revolutionized high-resolution image generation, but the design space of the autoencoder that is central to these systems remains underexplored. In this paper, we introduce LiteVAE, a new autoencoder design for LDMs, which leverages the 2D discrete wavelet transform to enhance scalability and computational efficiency over standard variational autoencoders (VAEs) with no sacrifice in output quality. We investigate the training methodologies and the decoder architecture of LiteVAE and propose several enhancements that improve the training dynamics and reconstruction quality. Our base LiteVAE model matches the quality of the established VAEs in current LDMs with a six-fold reduction in encoder parameters, leading to faster training and lower GPU memory requirements, while our larger model outperforms VAEs of comparable complexity across all evaluated metrics (rFID, LPIPS, PSNR, and SSIM).

In this paper, we present a system for augmenting virtual AI characters with long-term memory, enabling them to remember facts about themselves, their world, and past experiences.We propose a memory-creation pipeline that converts raw text into condensed memories and a memory-retrieval system that utilizes these memories to generate character responses.Using a fact-checking pipeline based on GPT-4 (OpenAI), our evaluation demonstrates that the character responses are grounded in the retrieved memories and maintain factual accuracy.We discuss the implications of our system for creating engaging and consistent virtual characters and highlight areas for future research, including large language model (LLM) guardrailing and virtual character personality development. { chat_history }Create a search query for the character ' s memory that helps answer the last user message .You cannot ask for clarification .Provide only the query .

Abstract Face swapping is the process of applying a source actor's appearance to a target actor's performance in a video. This is a challenging visual effect that has seen increasing demand in film and television production. Recent work has shown that data‐driven methods based on deep learning can produce compelling effects at production quality in a fraction of the time required for a traditional 3D pipeline. However, the dominant approach operates only on 2D imagery without reference to the underlying facial geometry or texture, resulting in poor generalization under novel viewpoints and little artistic control. Methods that do incorporate geometry rely on pre‐learned facial priors that do not adapt well to particular geometric features of the source and target faces. We approach the problem of face swapping from the perspective of learning simultaneous convolutional facial autoencoders for the source and target identities, using a shared encoder network with identity‐specific decoders. The key novelty in our approach is that each decoder first lifts the latent code into a 3D representation, comprising a dynamic face texture and a deformable 3D face shape, before projecting this 3D face back onto the input image using a differentiable renderer. The coupled autoencoders are trained only on videos of the source and target identities, without requiring 3D supervision. By leveraging the learned 3D geometry and texture, our method achieves face swapping with higher quality than when using off‐the‐shelf monocular 3D face reconstruction, and overall lower FID score than state‐of‐the‐art 2D methods. Furthermore, our 3D representation allows for efficient artistic control over the result, which can be hard to achieve with existing 2D approaches.

We propose an adaptive sampling and reconstruction method for offline Monte Carlo rendering. Our method produces sampling maps constrained by a user-defined budget that minimize the expected future denoising error. Compared to other state-of-the-art methods, which produce the necessary training data on the fly by composing pre-rendered images, our method samples from analytic noise distributions instead. These distributions are compact and closely approximate the pixel value distributions stemming from Monte Carlo rendering. Our method can efficiently sample training data by leveraging only a few per-pixel statistics of the target distribution, which provides several benefits over the current state of the art. Most notably, our analytic distributions' modeling accuracy and sampling efficiency increase with sample count, essential for high-quality offline rendering. Although our distributions are approximate, our method supports joint end-to-end training of the sampling and denoising networks. Finally, we propose the addition of a global summary module to our architecture that accumulates valuable information from image regions outside of the network's receptive field. This information discourages sub-optimal decisions based on local information. Our evaluation against other state-of-the-art neural sampling methods demonstrates denoising quality and data efficiency improvements.

Abstract In this paper, we propose an algorithm for fully automatic neural face swapping in images and videos. To the best of our knowledge, this is the first method capable of rendering photo‐realistic and temporally coherent results at megapixel resolution. To this end, we introduce a progressively trained multi‐way comb network and a light‐ and contrast‐preserving blending method. We also show that while progressive training enables generation of high‐resolution images, extending the architecture and training data beyond two people allows us to achieve higher fidelity in generated expressions. When compositing the generated expression onto the target face, we show how to adapt the blending strategy to preserve contrast and low‐frequency lighting. Finally, we incorporate a refinement strategy into the face landmark stabilization algorithm to achieve temporal stability, which is crucial for working with high‐resolution videos. We conduct an extensive ablation study to show the influence of our design choices on the quality of the swap and compare our work with popular state‐of‐the‐art methods.

We present a novel approach to modeling stories using recurrent neural networks. Different story features are extracted using natural language processing techniques and used to encode the stories as sequences. These sequences can be learned by deep neural networks, in order to predict the next story events. The predictions can be used as an inspiration for writers who experience a writer's block. We further assist writers in their creative process by generating visualizations of the character interactions in the story. We show that suggestions from our model are rated as highly as the real scenes from a set of films and that our visualizations can help people in gaining deeper story understanding.