视听生成

The role of data in building AI systems has recently been emphasized by the emerging concept of data-centric AI. Unfortunately, in the real-world, datasets may contain dirty samples, such as poisoned samples from backdoor attack, noisy labels in crowdsourcing, and even hybrids of them. The presence of such dirty samples makes the DNNs vunerable and unreliable. Hence, it is critical to detect dirty samples to improve the quality and realiability of dataset. Existing detectors only focus on detecting poisoned samples or noisy labels, that are often prone to weak generalization when dealing with dirty samples from other fields. In this paper, we find a commonality of various dirty samples is visual-linguistic inconsistency between images and associated labels. To capture the semantic inconsistency between modalities, we propose versatile data cleanser (VDC) leveraging the surpassing capabilities of multimodal large language models (MLLM) in cross-modal alignment and reasoning. It consists of three consecutive modules: the visual question generation module to generate insightful questions about the image; the visual question answering module to acquire the semantics of the visual content by answering the questions with MLLM; followed by the visual answer evaluation module to evaluate the inconsistency. Extensive experiments demonstrate its superior performance and generalization to various categories and types of dirty samples. The code is available at [https://github.com/zihao-ai/vdc](https://github.com/zihao-ai/vdc).

In this work, we investigate the capability of generating images from pre-trained diffusion models at much higher resolutions than the training image sizes. In addition, the generated images should have arbitrary image aspect ratios. When generating images directly at a higher resolution, 1024 x 1024, with the pre-trained Stable Diffusion using training images of resolution 512 x 512, we observe persistent problems of object repetition and unreasonable object structures. Existing works for higher-resolution generation, such as attention-based and joint-diffusion approaches, cannot well address these issues. As a new perspective, we examine the structural components of the U-Net in diffusion models and identify the crucial cause as the limited perception field of convolutional kernels. Based on this key observation, we propose a simple yet effective re-dilation that can dynamically adjust the convolutional perception field during inference. We further propose the dispersed convolution and noise-damped classifier-free guidance, which can enable ultra-high-resolution image generation (e.g., 4096 x 4096). Notably, our approach does not require any training or optimization. Extensive experiments demonstrate that our approach can address the repetition issue well and achieve state-of-the-art performance on higher-resolution image synthesis, especially in texture details. Our work also suggests that a pre-trained diffusion model trained on low-resolution images can be directly used for high-resolution visual generation without further tuning, which may provide insights for future research on ultra-high-resolution image and video synthesis. More results are available at the anonymous website: https://scalecrafter.github.io/ScaleCrafter/

Recent deep music generation studies have put much emphasis on long-term generation with structures. However, we are yet to see high-quality, well-structured **whole-song** generation. In this paper, we make the first attempt to model a full music piece under the realization of *compositional hierarchy*. With a focus on symbolic representations of pop songs, we define a hierarchical language, in which each level of hierarchy focuses on the semantics and context dependency at a certain music scope. The high-level languages reveal whole-song form, phrase, and cadence, whereas the low-level languages focus on notes, chords, and their local patterns. A cascaded diffusion model is trained to model the hierarchical language, where each level is conditioned on its upper levels. Experiments and analysis show that our model is capable of generating full-piece music with recognizable global verse-chorus structure and cadences, and the music quality is higher than the baselines. Additionally, we show that the proposed model is *controllable* in a flexible way. By sampling from the interpretable hierarchical languages or adjusting pre-trained external representations, users can control the music flow via various features such as phrase harmonic structures, rhythmic patterns, and accompaniment texture.

In this paper, we aim to adapt a model at test-time using a few unlabeled data to address distribution shifts. To tackle the challenges of extracting domain knowledge from a limited amount of data, it is crucial to utilize correlated information from pre-trained backbones and source domains. Previous studies fail to utilize recent foundation models with strong out-of-distribution generalization. Additionally, domain-centric designs are not flavored in their works. Furthermore, they employ the process of modelling source domains and the process of learning to adapt independently into disjoint training stages. In this work, we propose an approach on top of the pre-computed features of the foundation model. Specifically, we build a knowledge bank to learn the transferable knowledge from source domains. Conditioned on few-shot target data, we introduce a domain prompt generator to condense the knowledge bank into a domain-specific prompt. The domain prompt then directs the visual features towards a particular domain via a guidance module. Moreover, we propose a domain-aware contrastive loss and employ meta-learning to facilitate domain knowledge extraction. Extensive experiments are conducted to validate the domain knowledge extraction. The proposed method outperforms previous work on 5 large-scale benchmarks including WILDS and DomainNet.

Standard language models generate text by selecting tokens from a fixed, finite, and standalone vocabulary. We introduce a novel method that selects context-aware phrases from a collection of supporting documents. One of the most significant challenges for this paradigm shift is determining the training oracles, because a string of text can be segmented in various ways and each segment can be retrieved from numerous possible documents. To address this, we propose to initialize the training oracles using linguistic heuristics and, more importantly, bootstrap the oracles through iterative self-reinforcement. Extensive experiments show that our model not only outperforms standard language models on a variety of knowledge-intensive tasks but also demonstrates improved generation quality in open-ended text generation. For instance, compared to the standard language model counterpart, our model raises the accuracy from 23.47% to 36.27% on OpenbookQA, and improves the MAUVE score from 42.61% to 81.58% in open-ended text generation. Remarkably, our model also achieves the best performance and the lowest latency among several retrieval-augmented baselines. In conclusion, we assert that retrieval is more accurate generation and hope that our work will encourage further research on this new paradigm shift.

Controllable layout generation refers to the process of creating a plausible visual arrangement of elements within a graphic design (*e.g.*, document and web designs) with constraints representing design intentions. Although recent diffusion-based models have achieved state-of-the-art FID scores, they tend to exhibit more pronounced misalignment compared to earlier transformer-based models. In this work, we propose the **LA**yout **C**onstraint diffusion mod**E**l (LACE), a unified model to handle a broad range of layout generation tasks, such as arranging elements with specified attributes and refining or completing a coarse layout design. The model is based on continuous diffusion models. Compared with existing methods that use discrete diffusion models, continuous state-space design can enable the incorporation of continuous aesthetic constraint functions in training more naturally. For conditional generation, we propose injecting layout conditions in the form of masks or gradient guidance during inference. Empirical results show that LACE produces high-quality layouts and outperforms existing state-of-the-art baselines. We will release our source code and model checkpoints.

Our understanding of the visual world is centered around various concept axes, characterizing different aspects of visual entities. While different concept axes can be easily specified by language, e.g., color, the exact visual nuances along each axis often exceed the limitations of linguistic articulations, e.g., a particular style of painting. In this work, our goal is to learn a language-informed visual concept representation, by simply distilling large pre-trained vision-language models. Specifically, we train a set of concept encoders to encode the information pertinent to a set of language-informed concept axes, with an objective of reproducing the input image through a pre-trained Text-to-Image (T2I) model. To encourage better disentanglement of different concept encoders, we anchor the concept embeddings to a set of text embeddings obtained from a pre-trained Visual Question Answering (VQA) model. At inference time, the model extracts concept embeddings along various axes from new test images, which can be remixed to generate images with novel compositions of visual concepts. With a lightweight test-time finetuning procedure, it can also generalize to novel concepts unseen at training. Project page at https://cs.stanford.edu/~yzzhang/projects/concept-axes.

While Large Language Models (LLMs) are the dominant models for generative tasks in language, they do not perform as well as diffusion models on image and video generation. To effectively use LLMs for visual generation, one crucial component is the visual tokenizer that maps pixel-space inputs to discrete tokens appropriate for LLM learning. In this paper, we introduce \modelname{}, a video tokenizer designed to generate concise and expressive tokens for both videos and images using a common token vocabulary. Equipped with this new tokenizer, we show that LLMs outperform diffusion models on standard image and video generation benchmarks including ImageNet and Kinetics. In addition, we demonstrate that our tokenizer surpasses the previously top-performing video tokenizer on two more tasks: (1) video compression comparable to the next-generation video codec (VCC) according to human evaluations, and (2) learning effective representations for action recognition tasks.

We present a novel approach named OmniControl for incorporating flexible spatial control signals into a text-conditioned human motion generation model based on the diffusion process. Unlike previous methods that can only control the pelvis trajectory, OmniControl can incorporate flexible spatial control signals over different joints at different times with only one model. Specifically, we propose analytic spatial guidance that ensures the generated motion can tightly conform to the input control signals. At the same time, realism guidance is introduced to refine all the joints to generate more coherent motion. Both the spatial and realism guidance are essential and they are highly complementary for balancing control accuracy and motion realism. By combining them, OmniControl generates motions that are realistic, coherent, and consistent with the spatial constraints. Experiments on HumanML3D and KIT-ML datasets show that OmniControl not only achieves significant improvement over state-of-the-art methods on pelvis control but also shows promising results when incorporating the constraints over other joints. Project page: https://neu-vi.github.io/omnicontrol/.

Due to privacy or patent concerns, a growing number of large models are released without granting access to their training data, making transferring their knowledge inefficient and problematic. In response, Data-Free Knowledge Distillation (DFKD) methods have emerged as direct solutions. However, simply adopting models derived from DFKD for real-world applications suffers significant performance degradation, due to the discrepancy between teachers' training data and real-world scenarios (student domain). The degradation stems from the portions of teachers' knowledge that are not applicable to the student domain. They are specific to the teacher domain and would undermine students' performance. Hence, selectively transferring teachers' appropriate knowledge becomes the primary challenge in DFKD. In this work, we propose a simple but effective method AuG-KD. It utilizes an uncertainty-guided and sample-specific anchor to align student-domain data with the teacher domain and leverages a generative method to progressively trade off the learning process between OOD knowledge distillation and domain-specific information learning via mixup learning. Extensive experiments in 3 datasets and 8 settings demonstrate the stability and superiority of our approach.

Diffusion model based Text-to-Image has achieved impressive achievements recently. Although current technology for synthesizing images is highly advanced and capable of generating images with high fidelity, it is still possible to give the show away when focusing on the text area in the generated image, as synthesized text often contains blurred, unreadable, or incorrect characters, making visual text generation one of the most challenging issues in this field. To address this issue, we introduce AnyText, a diffusion-based multilingual visual text generation and editing model, that focuses on rendering accurate and coherent text in the image. AnyText comprises a diffusion pipeline with two primary elements: an auxiliary latent module and a text embedding module. The former uses inputs like text glyph, position, and masked image to generate latent features for text generation or editing. The latter employs an OCR model for encoding stroke data as embeddings, which blend with image caption embeddings from the tokenizer to generate texts that seamlessly integrate with the background. We employed text-control diffusion loss and text perceptual loss for training to further enhance writing accuracy. AnyText can write characters in multiple languages, to the best of our knowledge, this is the first work to address multilingual visual text generation. It is worth mentioning that AnyText can be plugged into existing diffusion models from the community for rendering or editing text accurately. After conducting extensive evaluation experiments, our method has outperformed all other approaches by a significant margin. Additionally, we contribute the first large-scale multilingual text images dataset, AnyWord-3M, containing 3 million image-text pairs with OCR annotations in multiple languages. Based on AnyWord-3M dataset, we propose AnyText-benchmark for the evaluation of visual text generation accuracy and quality. Our project will be open-sourced soon to improve and promote the development of text generation technology.

Despite significant advances in large-scale text-to-image models, achieving hyper-realistic human image generation remains a desirable yet unsolved task. Existing models like Stable Diffusion and DALL·E 2 tend to generate human images with incoherent parts or unnatural poses. To tackle these challenges, our key insight is that human image is inherently structural over multiple granularities, from the coarse-level body skeleton to fine-grained spatial geometry. Therefore, capturing such correlations between the explicit appearance and latent structure in one model is essential to generate coherent and natural human images. To this end, we propose a unified framework, HyperHuman, that generates in-the-wild human images of high realism and diverse layouts. Specifically, 1) we first build a large-scale human-centric dataset, named HumanVerse, which consists of 340M images with comprehensive annotations like human pose, depth, and surface normal. 2) Next, we propose a Latent Structural Diffusion Model that simultaneously denoises the depth and surface normal along with the synthesized RGB image. Our model enforces the joint learning of image appearance, spatial relationship, and geometry in a unified network, where each branch in the model complements to each other with both structural awareness and textural richness. 3) Finally, to further boost the visual quality, we propose a Structure-Guided Refiner to compose the predicted conditions for more detailed generation of higher resolution. Extensive experiments demonstrate that our framework yields the state-of-the-art performance, generating hyper-realistic human images under diverse scenarios.

Most neural networks for classification primarily learn features differentiated by input-domain related information such as visual similarity of objects in an image. While this focus is natural behavior, it can inadvertently introduce an inductive bias that conflicts with unseen relations in an implicit output-domain determined by human labeling based on their own world knowledge. Such conflicts can limit generalization of models by potential dominance of the input-domain focused bias in inference. To overcome this limitation without external resources, we introduce Output-Domain focused Biasing (ODB) training strategy that constructs inductive biases on features differentiated by only output labels. It has four steps: 1) it learns intermediate latent object features in an unsupervised manner; 2) it decouples their visual dependencies by assigning new independent embedding parameters; 3) it captures structured features optimized for the original classification task; and 4) it integrates the structured features with the original visual features for the final prediction. We implement the ODB on a vision transformer architecture, and achieved significant improvements on image classification benchmarks. This paper offers a straightforward and effective method to obtain and utilize output-domain focused inductive bias for classification mapping two different domains.

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.

As the scale of vision models continues to grow, the emergence of Visual Prompt Tuning (VPT) as a parameter-efficient transfer learning technique has gained attention due to its superior performance compared to traditional full-finetuning. However, the conditions favoring VPT (the "when") and the underlying rationale (the "why") remain unclear. In this paper, we conduct a comprehensive analysis across 19 distinct datasets and tasks. To understand the "when" aspect, we identify the scenarios where VPT proves favorable by two dimensions: task objectives and data distributions. We find that VPT is preferrable when there is 1) a substantial disparity between the original and the downstream task objectives ($e.g.$, transitioning from classification to counting), or 2) a notable similarity in data distributions between the two tasks ($e.g.$, both involve natural images). In exploring the "why" dimension, our results indicate VPT's success cannot be attributed solely to overfitting and optimization considerations. The unique way VPT preserves original features and adds parameters appears to be a pivotal factor. Our study provides insights into VPT's mechanisms, and offers guidance for its optimal utilization.

Scene graph generation (SGG) models have suffered from inherent problems regarding the benchmark datasets such as the long-tailed predicate distribution and missing annotation problems. In this work, we aim to alleviate the long-tailed problem of SGG by utilizing unannotated triplets. To this end, we introduce a **S**elf-**T**raining framework for **SGG** **(ST-SGG)** that assigns pseudo-labels for unannotated triplets based on which the SGG models are trained. While there has been significant progress in self-training for image recognition, designing a self-training framework for the SGG task is more challenging due to its inherent nature such as the semantic ambiguity and the long-tailed distribution of predicate classes. Hence, we propose a novel pseudo-labeling technique for SGG, called **C**lass-specific **A**daptive **T**hresholding with **M**omentum **(CATM)**, which is a model-agnostic framework that can be applied to any existing SGG models. Furthermore, we devise a graph structure learner (GSL) that is beneficial when adopting our proposed self-training framework to the state-of-the-art message-passing neural network (MPNN)-based SGG models. Our extensive experiments verify the effectiveness of ST-SGG on various SGG models, particularly in enhancing the performance on fine-grained predicate classes.

A remarkable ability of human beings resides in compositional reasoning, i.e., the capacity to make "infinite use of finite means". However, current large vision-language foundation models (VLMs) fall short of such compositional abilities due to their ``bag-of-words" behaviors and inability to construct words that correctly represent visual entities and the relations among the entities. To this end, we propose CoVLM, which can guide the LLM to explicitly compose visual entities and relationships among the text and dynamically communicate with the vision encoder and detection network to achieve vision-language communicative decoding. Specifically, we first devise a set of novel communication tokens for the LLM, for dynamic communication between the visual detection system and the language system. A communication token is generated by the LLM following a visual entity or a relation, to inform the detection network to propose regions that are relevant to the sentence generated so far. The proposed regions-of-interests (ROIs) are then fed back into the LLM for better language generation contingent on the relevant regions. The LLM is thus able to compose the visual entities and relationships through the communication tokens. The vision-to-language and language-to-vision communication are iteratively performed until the entire sentence is generated. Our framework seamlessly bridges the gap between visual perception and LLMs and outperforms previous VLMs by a large margin on compositional reasoning benchmarks (e.g., ~20% in HICO-DET mAP, ~14% in Cola top-1 accuracy, and ~3% on ARO top-1 accuracy). We also achieve state-of-the-art performances on traditional vision-language tasks such as referring expression comprehension and visual question answering.

Audio-visual speech separation methods aim to integrate different modalities to generate high-quality separated speech, thereby enhancing the performance of downstream tasks such as speech recognition. Most existing state-of-the-art (SOTA) models operate in the time domain. However, their overly simplistic approach to modeling acoustic features often necessitates larger and more computationally intensive models in order to achieve SOTA performance. In this paper, we present a novel time-frequency domain audio-visual speech separation method: Recurrent Time-Frequency Separation Network (RTFS-Net), which applies its algorithms on the complex time-frequency bins yielded by the Short-Time Fourier Transform. We model and capture the time and frequency dimensions of the audio independently using a multi-layered RNN along each dimension. Furthermore, we introduce a unique attention-based fusion technique for the efficient integration of audio and visual information, and a new mask separation approach that takes advantage of the intrinsic spectral nature of the acoustic features for a clearer separation. RTFS-Net outperforms the prior SOTA method in both inference speed and separation quality while reducing the number of parameters by 90% and MACs by 83%. This is the first time-frequency domain audio-visual speech separation method to outperform all contemporary time-domain counterparts.

We introduce MAGNeT, a masked generative sequence modeling method that operates directly over several streams of audio tokens. Unlike prior work, MAGNeT is comprised of a single-stage, non-autoregressive transformer. During training, we predict spans of masked tokens obtained from a masking scheduler, while during inference we gradually construct the output sequence using several decoding steps. To further enhance the quality of the generated audio, we introduce a novel rescoring method in which, we leverage an external pre-trained model to rescore and rank predictions from MAGNeT, which will be then used for later decoding steps. Lastly, we explore a hybrid version of MAGNeT, in which we fuse between autoregressive and non-autoregressive models to generate the first few seconds in an autoregressive manner while the rest of the sequence is being decoded in parallel. We demonstrate the efficiency of MAGNeT for the task of text-to-music and text-to-audio generation and conduct an extensive empirical evaluation, considering both objective metrics and human studies. The proposed approach is comparable to the evaluated baselines, while being significantly faster (x$7$ faster than the autoregressive baseline). Through ablation studies and analysis, we shed light on the importance of each of the components comprising MAGNeT, together with pointing to the trade-offs between autoregressive and non-autoregressive modeling, considering latency, throughput, and generation quality. Samples are available on our demo page https://pages.cs.huji.ac.il/adiyoss-lab/MAGNeT.

Evaluating text-to-image models is notoriously difficult. A strong recent approach for assessing text-image faithfulness is based on QG/A (question generation and answering), which uses pre-trained foundational models to automatically generate a set of questions and answers from the prompt, and output images are scored based on whether these answers extracted with a visual question answering model are consistent with the prompt-based answers. This kind of evaluation is naturally dependent on the quality of the underlying QG and VQA models. We identify and address several reliability challenges in existing QG/A work: (a) QG questions should respect the prompt (avoiding hallucinations, duplications, and omissions) and (b) VQA answers should be consistent (not asserting that there is no motorcycle in an image while also claiming the motorcycle is blue). We address these issues with Davidsonian Scene Graph (DSG), an empirically grounded evaluation framework inspired by formal semantics, which is adaptable to any QG/A frameworks. DSG produces atomic and unique questions organized in dependency graphs, which (i) ensure appropriate semantic coverage and (ii) sidestep inconsistent answers. With extensive experimentation and human evaluation on a range of model configurations (LLM, VQA, and T2I), we empirically demonstrate that DSG addresses the challenges noted above. Finally, we present DSG-1k, an open-sourced evaluation benchmark that includes 1,060 prompts, covering a wide range of fine-grained semantic categories with a balanced distribution. We release the DSG-1k prompts and the corresponding DSG questions.