Recent progress in text-to-3D generation has been achieved through the utilization of score distillation methods: they make use of the pre-trained text-to-image (T2I) diffusion models by distilling via the diffusion model training objective. However, such an approach inevitably results in the use of random timesteps at each update, which increases the variance of the gradient and ultimately prolongs the optimization process. In this paper, we propose to enhance the text-to-3D optimization by leveraging the T2I diffusion prior in the generative sampling process with a predetermined timestep schedule. To this end, we interpret text-to3D optimization as a multi-view image-to-image translation problem, and propose a solution by approximating the probability flow. By leveraging the proposed novel optimization algorithm, we design DreamFlow, a practical three-stage coarseto-fine text-to-3D optimization framework that enables fast generation of highquality and high-resolution (i.e., 1024x1024) 3D contents. For example, we demonstrate that DreamFlow is 5 times faster than the existing state-of-the-art text-to-3D method, while producing more photorealistic 3D contents. Visit our project page (https://kyungmnlee.github.io/dreamflow.github.io/) for visualizations.
This paper considers image-based virtual try-on, which renders an image of a person wearing a curated garment, given a pair of images depicting the person and the garment, respectively. Previous works adapt existing exemplar-based inpainting diffusion models for virtual try-on to improve the naturalness of the generated visuals compared to other methods (e.g., GAN-based), but they fail to preserve the identity of the garments. To overcome this limitation, we propose a novel diffusion model that improves garment fidelity and generates authentic virtual try-on images. Our method, coined IDM-VTON, uses two different modules to encode the semantics of garment image; given the base UNet of the diffusion model, 1) the high-level semantics extracted from a visual encoder are fused to the cross-attention layer, and then 2) the low-level features extracted from parallel UNet are fused to the self-attention layer. In addition, we provide detailed textual prompts for both garment and person images to enhance the authenticity of the generated visuals. Finally, we present a customization method using a pair of person-garment images, which significantly improves fidelity and authenticity. Our experimental results show that our method outperforms previous approaches (both diffusion-based and GAN-based) in preserving garment details and generating authentic virtual try-on images, both qualitatively and quantitatively. Furthermore, the proposed customization method demonstrates its effectiveness in a real-world scenario.
Text-to-image (T2I) diffusion models, when fine-tuned on a few personal images, are able to generate visuals with a high degree of consistency. However, they still lack in synthesizing images of different scenarios or styles that are possible in the original pretrained models. To address this, we propose to fine-tune the T2I model by maximizing consistency to reference images, while penalizing the deviation from the pretrained model. We devise a novel training objective for T2I diffusion models that minimally fine-tunes the pretrained model to achieve consistency. Our method, dubbed \emph{Direct Consistency Optimization}, is as simple as regular diffusion loss, while significantly enhancing the compositionality of personalized T2I models. Also, our approach induces a new sampling method that controls the tradeoff between image fidelity and prompt fidelity. Lastly, we emphasize the necessity of using a comprehensive caption for reference images to further enhance the image-text alignment. We show the efficacy of the proposed method on the T2I personalization for subject, style, or both. In particular, our method results in a superior Pareto frontier to the baselines. Generated examples and codes are in our project page( https://dco-t2i.github.io/).
Recently, the cascaded two-pass architecture has emerged as a strong contender for on-device automatic speech recognition (ASR). A cascade of causal and shallow non-causal encoders coupled with a shared decoder enables operation in both streaming and look-ahead modes. In this paper, we propose shallow cascaded model by combining various model compression techniques such as knowledge distillation, shared decoder, and tied-and-reduced transducer network in order to reduce the model footprint. The shared decoder is changed into a tied-and-reduced network. The cascaded two-pass model is further compressed using knowledge distillation using a Kullback-Leibler divergence loss on the model posteriors. We demonstrate a 50% reduction in the size of a 41 M parameter cascaded teacher model with no noticeable degradation in ASR accuracy and a 30% reduction in latency
Generative priors of large-scale text-to-image diffusion models enable a wide range of new generation and editing applications on diverse visual modalities. However, when adapting these priors to complex visual modalities, often represented as multiple images (e.g., video), achieving consistency across a set of images is challenging. In this paper, we address this challenge with a novel method, Collaborative Score Distillation (CSD). CSD is based on the Stein Variational Gradient Descent (SVGD). Specifically, we propose to consider multiple samples as "particles" in the SVGD update and combine their score functions to distill generative priors over a set of images synchronously. Thus, CSD facilitates seamless integration of information across 2D images, leading to a consistent visual synthesis across multiple samples. We show the effectiveness of CSD in a variety of tasks, encompassing the visual editing of panorama images, videos, and 3D scenes. Our results underline the competency of CSD as a versatile method for enhancing inter-sample consistency, thereby broadening the applicability of text-to-image diffusion models.
Vision-language models, such as contrastive language-image pre-training (CLIP), have demonstrated impressive results in natural image domains. However, these models often struggle when applied to specialized domains like remote sensing, and adapting to such domains is challenging due to the limited number of image-text pairs available for training. To address this, we propose S-CLIP, a semi-supervised learning method for training CLIP that utilizes additional unpaired images. S-CLIP employs two pseudo-labeling strategies specifically designed for contrastive learning and the language modality. The caption-level pseudo-label is given by a combination of captions of paired images, obtained by solving an optimal transport problem between unpaired and paired images. The keyword-level pseudo-label is given by a keyword in the caption of the nearest paired image, trained through partial label learning that assumes a candidate set of labels for supervision instead of the exact one. By combining these objectives, S-CLIP significantly enhances the training of CLIP using only a few image-text pairs, as demonstrated in various specialist domains, including remote sensing, fashion, scientific figures, and comics. For instance, S-CLIP improves CLIP by 10% for zero-shot classification and 4% for image-text retrieval on the remote sensing benchmark, matching the performance of supervised CLIP while using three times fewer image-text pairs.
Learning with few labeled tabular samples is often an essential requirement for industrial machine learning applications as varieties of tabular data suffer from high annotation costs or have difficulties in collecting new samples for novel tasks. Despite the utter importance, such a problem is quite under-explored in the field of tabular learning, and existing few-shot learning schemes from other domains are not straightforward to apply, mainly due to the heterogeneous characteristics of tabular data. In this paper, we propose a simple yet effective framework for few-shot semi-supervised tabular learning, coined Self-generated Tasks from UNlabeled Tables (STUNT). Our key idea is to self-generate diverse few-shot tasks by treating randomly chosen columns as a target label. We then employ a meta-learning scheme to learn generalizable knowledge with the constructed tasks. Moreover, we introduce an unsupervised validation scheme for hyperparameter search (and early stopping) by generating a pseudo-validation set using STUNT from unlabeled data. Our experimental results demonstrate that our simple framework brings significant performance gain under various tabular few-shot learning benchmarks, compared to prior semi- and self-supervised baselines. Code is available at https://github.com/jaehyun513/STUNT.
We aim to diagnose the potential biases in image classifiers. To this end, prior works manually labeled biased attributes or visualized biased features, which need high annotation costs or are often ambiguous to interpret. Instead, we leverage two types (generative and discriminative) of pre-trained vision-language models to describe the visual bias as a word. Specifically, we propose bias-to-text (B2T), which generates captions of the mispredicted images using a pre-trained captioning model to extract the common keywords that may describe visual biases. Then, we categorize the bias type as spurious correlation or majority bias by checking if it is specific or agnostic to the class, based on the similarity of class-wise mispredicted images and the keyword upon a pre-trained vision-language joint embedding space, e.g., CLIP. We demonstrate that the proposed simple and intuitive scheme can recover well-known gender and background biases, and discover novel ones in real-world datasets. Moreover, we utilize B2T to compare the classifiers using different architectures or training methods. Finally, we show that one can obtain debiased classifiers using the B2T bias keywords and CLIP, in both zero-shot and full-shot manners, without using any human annotation on the bias.
Training a deep learning model with artificially generated data can be an alternative when training data are scarce, yet it suffers from poor generalization performance due to a large domain gap. In this paper, we characterize the domain gap by using a causal framework for data generation. We assume that the real and synthetic data have common content variables but different style variables. Thus, a model trained on synthetic dataset might have poor generalization as the model learns the nuisance style variables. To that end, we propose causal invariance learning which encourages the model to learn a style-invariant representation that enhances the syn-to-real generalization. Furthermore, we propose a simple yet effective feature distillation method that prevents catastrophic forgetting of semantic knowledge of the real domain. In sum, we refer to our method as Guided Causal Invariant Syn-to-real Generalization that effectively improves the performance of syn-to-real generalization. We empirically verify the validity of proposed methods, and especially, our method achieves state-of-the-art on visual syn-to-real domain generalization tasks such as image classification and semantic segmentation.
Contrastive representation learning seeks to acquire useful representations by estimating the shared information between multiple views of data. Here, the choice of data augmentation is sensitive to the quality of learned representations: as harder the data augmentations are applied, the views share more task-relevant information, but also task-irrelevant one that can hinder the generalization capability of representation. Motivated by this, we present a new robust contrastive learning scheme, coined R\'enyiCL, which can effectively manage harder augmentations by utilizing R\'enyi divergence. Our method is built upon the variational lower bound of R\'enyi divergence, but a na\"ive usage of a variational method is impractical due to the large variance. To tackle this challenge, we propose a novel contrastive objective that conducts variational estimation of a skew R\'enyi divergence and provide a theoretical guarantee on how variational estimation of skew divergence leads to stable training. We show that R\'enyi contrastive learning objectives perform innate hard negative sampling and easy positive sampling simultaneously so that it can selectively learn useful features and ignore nuisance features. Through experiments on ImageNet, we show that R\'enyi contrastive learning with stronger augmentations outperforms other self-supervised methods without extra regularization or computational overhead. Moreover, we also validate our method on other domains such as graph and tabular, showing empirical gain over other contrastive methods.