In multimodal sentiment analysis, modality missingness and quality degradation are common. Existing methods often rely on batch-level modality generation, generation but neglect sample-level missingness, hence their flexibility is limited severely in real-world scenarios. To address this, Sample-specific Modality Diagnosis and Cross-modal Enhancement for Incomplete Multimodal Representations (SMCIR) is proposed. Specifically, The Dynamic Multi-feature Fusion Detector (DMFD) is presented, which detects missingness and severity at the sample-level using indicators such as information entropy, modality similarity, and mutual information. Unlike batch-based methods, the DMFD provides fine-grained detection and adaptive responses, improving sensitivity to modality disturbances. Meanwhile, the Context-aware Modality Completion Generator (CMCG) is developed to restore missing modalities through context-guided reconstruction using multiscale feature fusion and cross-modal attention. In this way, the proposed CMCG method can avoid redundancy and inconsistency, enhancing the consistency and discriminativity of the fused representation. In CMCG, the text modality serves as a stable guide to improve context consistency. Experiments on the CMU-MOSI and CMU-MOSEI datasets show that SMCIR outperforms existing full-modal and non-recovery-based methods, well validating its efficacy and superiority in multimodal learning.

This paper presents SANA-Sprint, an efficient diffusion model for ultra-fast text-to-image (T2I) generation. SANA-Sprint is built on a pre-trained foundation model and augmented with hybrid distillation, dramatically reducing inference steps from 20 to 1-4. We introduce three key innovations: (1) We propose a training-free approach that transforms a pre-trained flow-matching model for continuous-time consistency distillation (sCM), eliminating costly training from scratch and achieving high training efficiency. Our hybrid distillation strategy combines sCM with latent adversarial distillation (LADD): sCM ensures alignment with the teacher model, while LADD enhances single-step generation fidelity. (2) SANA-Sprint is a unified step-adaptive model that achieves high-quality generation in 1-4 steps, eliminating step-specific training and improving efficiency. (3) We integrate ControlNet with SANA-Sprint for real-time interactive image generation, enabling instant visual feedback for user interaction. SANA-Sprint establishes a new Pareto frontier in speed-quality tradeoffs, achieving state-of-the-art performance with 7.59 FID and 0.74 GenEval in only 1 step - outperforming FLUX-schnell (7.94 FID / 0.71 GenEval) while being 10x faster (0.1s vs 1.1s on H100). It also achieves 0.1s (T2I) and 0.25s (ControlNet) latency for 1024 x 1024 images on H100, and 0.31s (T2I) on an RTX 4090, showcasing its exceptional efficiency and potential for AI-powered consumer applications (AIPC). Code and pre-trained models will be open-sourced.

We present DC-AE 1.5, a new family of deep compression autoencoders for high-resolution diffusion models. Increasing the autoencoder's latent channel number is a highly effective approach for improving its reconstruction quality. However, it results in slow convergence for diffusion models, leading to poorer generation quality despite better reconstruction quality. This issue limits the quality upper bound of latent diffusion models and hinders the employment of autoencoders with higher spatial compression ratios. We introduce two key innovations to address this challenge: i) Structured Latent Space, a training-based approach to impose a desired channel-wise structure on the latent space with front latent channels capturing object structures and latter latent channels capturing image details; ii) Augmented Diffusion Training, an augmented diffusion training strategy with additional diffusion training objectives on object latent channels to accelerate convergence. With these techniques, DC-AE 1.5 delivers faster convergence and better diffusion scaling results than DC-AE. On ImageNet 512x512, DC-AE-1.5-f64c128 delivers better image generation quality than DC-AE-f32c32 while being 4x faster. Code: https://github.com/dc-ai-projects/DC-Gen.

We introduce DC-AR, a novel masked autoregressive (AR) text-to-image generation framework that delivers superior image generation quality with exceptional computational efficiency. Due to the tokenizers' limitations, prior masked AR models have lagged behind diffusion models in terms of quality or efficiency. We overcome this limitation by introducing DC-HT - a deep compression hybrid tokenizer for AR models that achieves a 32x spatial compression ratio while maintaining high reconstruction fidelity and cross-resolution generalization ability. Building upon DC-HT, we extend MaskGIT and create a new hybrid masked autoregressive image generation framework that first produces the structural elements through discrete tokens and then applies refinements via residual tokens. DC-AR achieves state-of-the-art results with a gFID of 5.49 on MJHQ-30K and an overall score of 0.69 on GenEval, while offering 1.5-7.9x higher throughput and 2.0-3.5x lower latency compared to prior leading diffusion and autoregressive models.

We present Deep Compression Autoencoder (DC-AE), a new family of autoencoders for accelerating high-resolution diffusion models. Existing autoencodes have demonstrated impressive results at a moderate spatial compression ratio (e.g., 8x), but fail to maintain satisfactory reconstruction accuracy for high spatial compression ratios (e.g., 64x). We address this challenge by introducing two key techniques: (1) Residual Autoencoding, where we design our models to learn residuals based on the space-to-channel transformed features to alleviate the optimization difficulty of high spatial-compression autoencoders; (2) Decoupled High-Resolution Adaptation, an efficient decoupled three-phase training strategy for mitigating the generalization penalty of high spatial-compression autoencoders. With these designs, we improve the autoencoder's spatial compression ratio up to 128 while maintaining the reconstruction quality. Applying our DC-AE to latent diffusion models, we achieve significant speedup without accuracy drop. For example, on ImageNet 512x512, our DC-AE provides 19.1x inference speedup and 17.9x training speedup on H100 GPU for UViT-H while achieving a better FID, compared with the widely used SD-VAE-f8 autoencoder.

We introduce Hybrid Autoregressive Transformer (HART), the first autoregressive (AR) visual generation model capable of directly generating 1024x1024 images, rivaling diffusion models in image generation quality. Existing AR models face limitations due to the poor image reconstruction quality of their discrete tokenizers and the prohibitive training costs associated with generating 1024px images. To address these challenges, we present the hybrid tokenizer, which decomposes the continuous latents from the autoencoder into two components: discrete tokens representing the big picture and continuous tokens representing the residual components that cannot be represented by the discrete tokens. The discrete component is modeled by a scalable-resolution discrete AR model, while the continuous component is learned with a lightweight residual diffusion module with only 37M parameters. Compared with the discrete-only VAR tokenizer, our hybrid approach improves reconstruction FID from 2.11 to 0.30 on MJHQ-30K, leading to a 31% generation FID improvement from 7.85 to 5.38. HART also outperforms state-of-the-art diffusion models in both FID and CLIP score, with 4.5-7.7$\times$ higher throughput and 6.9-13.4$\times$ lower MACs. Our code is open sourced at https://github.com/mit-han-lab/hart.

We introduce Sana, a text-to-image framework that can efficiently generate images up to 4096$\times$4096 resolution. Sana can synthesize high-resolution, high-quality images with strong text-image alignment at a remarkably fast speed, deployable on laptop GPU. Core designs include: (1) Deep compression autoencoder: unlike traditional AEs, which compress images only 8$\times$, we trained an AE that can compress images 32$\times$, effectively reducing the number of latent tokens. (2) Linear DiT: we replace all vanilla attention in DiT with linear attention, which is more efficient at high resolutions without sacrificing quality. (3) Decoder-only text encoder: we replaced T5 with modern decoder-only small LLM as the text encoder and designed complex human instruction with in-context learning to enhance the image-text alignment. (4) Efficient training and sampling: we propose Flow-DPM-Solver to reduce sampling steps, with efficient caption labeling and selection to accelerate convergence. As a result, Sana-0.6B is very competitive with modern giant diffusion model (e.g. Flux-12B), being 20 times smaller and 100+ times faster in measured throughput. Moreover, Sana-0.6B can be deployed on a 16GB laptop GPU, taking less than 1 second to generate a 1024$\times$1024 resolution image. Sana enables content creation at low cost. Code and model will be publicly released upon publication.

This paper presents SANA-1.5, a linear Diffusion Transformer for efficient scaling in text-to-image generation. Building upon SANA-1.0, we introduce three key innovations: (1) Efficient Training Scaling: A depth-growth paradigm that enables scaling from 1.6B to 4.8B parameters with significantly reduced computational resources, combined with a memory-efficient 8-bit optimizer. (2) Model Depth Pruning: A block importance analysis technique for efficient model compression to arbitrary sizes with minimal quality loss. (3) Inference-time Scaling: A repeated sampling strategy that trades computation for model capacity, enabling smaller models to match larger model quality at inference time. Through these strategies, SANA-1.5 achieves a text-image alignment score of 0.72 on GenEval, which can be further improved to 0.80 through inference scaling, establishing a new SoTA on GenEval benchmark. These innovations enable efficient model scaling across different compute budgets while maintaining high quality, making high-quality image generation more accessible.

Cloud computing eliminates the limitations of local hardware architecture while also enabling rapid data sharing between healthcare institutions. Encryption of electronic medical records (EMRs) before uploading to cloud servers is necessary for privacy. However, encryption brings challenges for computation. Public Key Encryption with Equality Test (PKEET) allows cloud servers to test the underlying message equality without decryption. Therefore, it can be used to classify the encrypted EMRs corresponding to different medical symptoms. However, traditional PKEETs have limitations in testing the similarity between the ciphertexts. Undoubtedly, it can not handle EMR classification with similar medical symptoms efficiently. In this work, we propose a lightweight public key encryption with similarity test (PKEST) for the EMR classification shared in medical consortia. Our scheme can resist offline message recovery attacks, which may be launched by the insider manager, and the traditional paring computation is not necessary. Our experiment simulation shows that the similarity error between ciphertext and plaintext is tiny when the parameters are set properly. Compared to previous works, our scheme not only achieves the classification of similar encrypted EMRs but is also more efficient than traditional PKEETs since our construction does not need paring computation anymore.

Safety is the primary priority of autonomous driving. Nevertheless, no published dataset currently supports the direct and explainable safety evaluation for autonomous driving. In this work, we propose DeepAccident, a large-scale dataset generated via a realistic simulator containing diverse accident scenarios that frequently occur in real-world driving. The proposed DeepAccident dataset includes 57K annotated frames and 285K annotated samples, approximately 7 times more than the large-scale nuScenes dataset with 40k annotated samples. In addition, we propose a new task, end-to-end motion and accident prediction, which can be used to directly evaluate the accident prediction ability for different autonomous driving algorithms. Furthermore, for each scenario, we set four vehicles along with one infrastructure to record data, thus providing diverse viewpoints for accident scenarios and enabling V2X (vehicle-to-everything) research on perception and prediction tasks. Finally, we present a baseline V2X model named V2XFormer that demonstrates superior performance for motion and accident prediction and 3D object detection compared to the single-vehicle model.

Compared with traditional RGB-only visual tracking, few datasets have been constructed for RGB-D tracking. In this paper, we propose ARKitTrack, a new RGB-D tracking dataset for both static and dynamic scenes captured by consumer-grade LiDAR scanners equipped on Apple's iPhone and iPad. ARKitTrack contains 300 RGB-D sequences, 455 targets, and 229.7K video frames in total. Along with the bounding box annotations and frame-level attributes, we also annotate this dataset with 123.9K pixel-level target masks. Besides, the camera intrinsic and camera pose of each frame are provided for future developments. To demonstrate the potential usefulness of this dataset, we further present a unified baseline for both box-level and pixel-level tracking, which integrates RGB features with bird's-eye-view representations to better explore cross-modality 3D geometry. In-depth empirical analysis has verified that the ARKitTrack dataset can significantly facilitate RGB-D tracking and that the proposed baseline method compares favorably against the state of the arts. The code and dataset is available at https://arkittrack.github.io.

Perception systems in modern autonomous driving vehicles typically take inputs from complementary multi-modal sensors, e.g., LiDAR and cameras. However, in real-world applications, sensor corruptions and failures lead to inferior performances, thus compromising autonomous safety. In this paper, we propose a robust framework, called MetaBEV, to address extreme real-world environments, involving overall six sensor corruptions and two extreme sensor-missing situations. In MetaBEV, signals from multiple sensors are first processed by modal-specific encoders. Subsequently, a set of dense BEV queries are initialized, termed meta-BEV. These queries are then processed iteratively by a BEV-Evolving decoder, which selectively aggregates deep features from either LiDAR, cameras, or both modalities. The updated BEV representations are further leveraged for multiple 3D prediction tasks. Additionally, we introduce a new M2oE structure to alleviate the performance drop on distinct tasks in multi-task joint learning. Finally, MetaBEV is evaluated on the nuScenes dataset with 3D object detection and BEV map segmentation tasks. Experiments show MetaBEV outperforms prior arts by a large margin on both full and corrupted modalities. For instance, when the LiDAR signal is missing, MetaBEV improves 35.5% detection NDS and 17.7% segmentation mIoU upon the vanilla BEVFusion [25] model; and when the camera signal is absent, MetaBEV still achieves 69.2% NDS and 53.7% mIoU, which is even higher than previous works that perform on full-modalities. Moreover, MetaBEV performs moderately against previous methods in both canonical perception and multi-task learning settings, refreshing state-of-the-art nuScenes BEV map segmentation with 70.4% mIoU.

Remote sensing images have the essential attribute of large-scale spatial variation and complex scene information, as well as the high similarity between various classes and the significant differences within same class, which are easy to cause misclassification. To solve this problem, an efficient systematic architecture named EFCOMFF-Net (Multi-scale Feature Fusion Network with Enhanced Feature Correlation) is proposed to reduce the gap among multi-scale features and fuse them to improve the representation ability of remote sensing images. Firstly, to strengthen the correlation of multi-scale features, a Feature Correlation Enhancement Module (FCEM) is specifically developed, which takes the features of different stages of the backbone network as input data to obtain multi-scale features with enhanced correlation. Considering the differences between the shallow features and the deep features, the EFCOMFF-Net-v1 related to shallow features and EFCOMFF-Net-v2 related to deep features with different structures are proposed. Secondly, the designed two versions of the deep learning network focus on the global contour information and need to encode more accurate spatial information. A Feature Aggregation Attention Module (FAAM) is designed and embedded into the network to encode the deep features by applying the spatial information aggregation features. Finally, considering that the simple integration strategy cannot reduce the gap between the shallow multi-scale features and the deep features, a Feature Refinement Module (FRM) is presented to optimize the network. ResNet50, DenseNet121, and ResNet152 are selected to conduct a considerable number of experiments on four datasets, which show the superiority of our method compared to recent methods.

Convolutional neural network (CNN) and Transformer-based self-attention models have their advantages in extracting local information and global semantic information, and it is a trend to design a model combining stacked residual convolution blocks (SRCB) and Transformer. How to efficiently integrate the two mechanisms to improve the segmentation effect of remote sensing (RS) images is an urgent problem to be solved. An efficient fusion via SRCB and Transformer (SRCBTFusion-Net) is proposed as a new semantic segmentation architecture for RS images. The SRCBTFusion-Net adopts an encoder-decoder structure, and the Transformer is embedded into SRCB to form a double coding structure, then the coding features are up-sampled and fused with multi-scale features of SRCB to form a decoding structure. Firstly, a semantic information enhancement module (SIEM) is proposed to get global clues for enhancing deep semantic information. Subsequently, the relationship guidance module (RGM) is incorporated to re-encode the decoder’s upsampled feature maps, enhancing the edge segmentation performance. Secondly, a multipath atrous self-attention module (MASM) is developed to enhance the effective selection and weighting of low-level features, effectively reducing the potential confusion introduced by the skip connections between low-level and high-level features. Finally, a multi-scale feature aggregation module (MFAM) is developed to enhance the extraction of semantic and contextual information, thus alleviating the loss of image feature information and improving the ability to identify similar categories. The proposed SRCBTFusion-Net’s performance on the Vaihingen and Potsdam datasets is superior to the state-of-the-art methods. The code will be freely available at https://github.com/js257/SRCBTFusion-Net.