Search

Simulating higher-order topological materials in synthetic quantum matter is an active research frontier for its theoretical significance in fundamental physics and promising applications in quantum technologies. Here we experimentally implement two-dimensional (2D) momentum lattices with highly programmable ability using ultracold 87Rb atoms. Through precise control of experimental parameters, we simulate a 2D Su-Schrieffer-Heeger model with this technique, and observe the characteristic dynamics of corner and edge-bound states, where the corner state is identified as a higher-order topological bound state in the continuum. We further study the adiabatic preparation of the corner state by engineering evolutions with time-dependent Hamiltonians. We also demonstrate the higher-order topological phase transition by measuring both the bulk topological invariant and the topological corner state. Our new platform opens the avenue for exploring the exotic dynamics and topology in higher synthetic dimensions, making use of the rich degrees of freedom of cold atoms systems., Comment: 7pages,4figures

To preserve user privacy in recommender systems, federated recommendation (FR) based on federated learning (FL) emerges, keeping the personal data on the local client and updating a model collaboratively. Unlike FL, FR has a unique sparse aggregation mechanism, where the embedding of each item is updated by only partial clients, instead of full clients in a dense aggregation of general FL. Recently, as an essential principle of FL, model security has received increasing attention, especially for Byzantine attacks, where malicious clients can send arbitrary updates. The problem of exploring the Byzantine robustness of FR is particularly critical since in the domains applying FR, e.g., e-commerce, malicious clients can be injected easily by registering new accounts. However, existing Byzantine works neglect the unique sparse aggregation of FR, making them unsuitable for our problem. Thus, we make the first effort to investigate Byzantine attacks on FR from the perspective of sparse aggregation, which is non-trivial: it is not clear how to define Byzantine robustness under sparse aggregations and design Byzantine attacks under limited knowledge/capability. In this paper, we reformulate the Byzantine robustness under sparse aggregation by defining the aggregation for a single item as the smallest execution unit. Then we propose a family of effective attack strategies, named Spattack, which exploit the vulnerability in sparse aggregation and are categorized along the adversary's knowledge and capability. Extensive experimental results demonstrate that Spattack can effectively prevent convergence and even break down defenses under a few malicious clients, raising alarms for securing FR systems., Comment: accepted by AAAI 2025

The optical tweezer experiment with neutral atoms is a focal topic in cold atom physics due to its significant potential in quantum computing and simulation. Here, we present the realization of a dual-species optical tweezer for both Rb and K atoms, marking the first step towards creating a polar molecule optical tweezer array. Initially, Rb and K atoms are collected using a dual magneto-optical trap (MOT) and further cooled to 7 $\mu$K for Rb and 10 $\mu$K for K. By employing 850 nm tweezer beams, we demonstrate the ability to capture individual Rb or K atoms. The filling ratios of Rb and K can be finely adjusted by controlling the atomic densities of both species. Utilizing the post-selection technique, we can create a deterministic array of two-species atoms, paving the way for future polar molecule array formation., Comment: 6 pages, 4 figures

We present a framework for learning to generate background music from video inputs. Unlike existing works that rely on symbolic musical annotations, which are limited in quantity and diversity, our method leverages large-scale web videos accompanied by background music. This enables our model to learn to generate realistic and diverse music. To accomplish this goal, we develop a generative video-music Transformer with a novel semantic video-music alignment scheme. Our model uses a joint autoregressive and contrastive learning objective, which encourages the generation of music aligned with high-level video content. We also introduce a novel video-beat alignment scheme to match the generated music beats with the low-level motions in the video. Lastly, to capture fine-grained visual cues in a video needed for realistic background music generation, we introduce a new temporal video encoder architecture, allowing us to efficiently process videos consisting of many densely sampled frames. We train our framework on our newly curated DISCO-MV dataset, consisting of 2.2M video-music samples, which is orders of magnitude larger than any prior datasets used for video music generation. Our method outperforms existing approaches on the DISCO-MV and MusicCaps datasets according to various music generation evaluation metrics, including human evaluation. Results are available at https://genjib.github.io/project_page/VMAs/index.html, Comment: Project Page: https://genjib.github.io/project_page/VMAs/index.html

Facial movements play a crucial role in conveying altitude and intentions, and facial optical flow provides a dynamic and detailed representation of it. However, the scarcity of datasets and a modern baseline hinders the progress in facial optical flow research. This paper proposes FacialFlowNet (FFN), a novel large-scale facial optical flow dataset, and the Decomposed Facial Flow Model (DecFlow), the first method capable of decomposing facial flow. FFN comprises 9,635 identities and 105,970 image pairs, offering unprecedented diversity for detailed facial and head motion analysis. DecFlow features a facial semantic-aware encoder and a decomposed flow decoder, excelling in accurately estimating and decomposing facial flow into head and expression components. Comprehensive experiments demonstrate that FFN significantly enhances the accuracy of facial flow estimation across various optical flow methods, achieving up to an 11% reduction in Endpoint Error (EPE) (from 3.91 to 3.48). Moreover, DecFlow, when coupled with FFN, outperforms existing methods in both synthetic and real-world scenarios, enhancing facial expression analysis. The decomposed expression flow achieves a substantial accuracy improvement of 18% (from 69.1% to 82.1%) in micro-expressions recognition. These contributions represent a significant advancement in facial motion analysis and optical flow estimation. Codes and datasets can be found., Comment: ACMMM2024

The melting point of a material constitutes a pivotal property with profound implications across various disciplines of science, engineering, and technology. Recent advancements in machine learning potentials have revolutionized the field, enabling ab initio predictions of materials' melting points through atomic-scale simulations. However, a universal simulation methodology that can be universally applied to any material remains elusive. In this paper, we present a generic, fully automated workflow designed to predict the melting points of materials utilizing molecular dynamics simulations. This workflow incorporates two tailored simulation modalities, each addressing scenarios with and without elemental partitioning between solid and liquid phases. When the compositions of both phases remain unchanged upon melting or solidification, signifying the absence of partitioning, the melting point is identified as the temperature at which these phases coexist in equilibrium. Conversely, in cases where elemental partitioning occurs, our workflow estimates both the nominal melting point, marking the initial transition from solid to liquid, and the nominal solidification point, indicating the reverse process. To ensure precision in determining these critical temperatures, we employ an innovative temperature-volume data fitting technique, suitable for a diverse range of materials exhibiting notable volume disparities between their solid and liquid states. This comprehensive approach offers a robust and versatile solution for predicting melting points, fostering advancements in materials science and technology.

Self-supervised pre-training medical foundation models on large-scale datasets demonstrate exceptional performance. However, recent research questions this traditional notion, exploring whether an increase in pre-training data always leads to enhanced model performance. To address this issue, data-effective learning approaches have been introduced to select valuable samples for foundation model pre-training. Notably, current methods in this area lack a clear standard for sample selection, and the underlying theoretical foundation remains unknown. As the first attempt to address this limitation, we leverage V-information in self-supervised pre-training of foundation models. Our theoretical derivation confirms that by optimizing V-information, sample selection can be framed as an optimization problem where choosing diverse and challenging samples enhances model performance even under limited training data. Under this guidance, we develop an optimal data-effective learning method (OptiDEL) to optimize V-information in real-world medical domains. The OptiDEL method generates more diverse and harder samples to achieve or even exceed the performance of models trained on the full dataset while using substantially less data. We compare the OptiDEL method with state-of-the-art approaches finding that OptiDEL consistently outperforms existing approaches across eight different datasets, with foundation models trained on only 5% of the pre-training data surpassing the performance of those trained on the full dataset.

Differentially private graph analysis is a powerful tool for deriving insights from diverse graph data while protecting individual information. Designing private analytic algorithms for different graph queries often requires starting from scratch. In contrast, differentially private synthetic graph generation offers a general paradigm that supports one-time generation for multiple queries. Although a rich set of differentially private graph generation algorithms has been proposed, comparing them effectively remains challenging due to various factors, including differing privacy definitions, diverse graph datasets, varied privacy requirements, and multiple utility metrics. To this end, we propose PGB (Private Graph Benchmark), a comprehensive benchmark designed to enable researchers to compare differentially private graph generation algorithms fairly. We begin by identifying four essential elements of existing works as a 4-tuple: mechanisms, graph datasets, privacy requirements, and utility metrics. We discuss principles regarding these elements to ensure the comprehensiveness of a benchmark. Next, we present a benchmark instantiation that adheres to all principles, establishing a new method to evaluate existing and newly proposed graph generation algorithms. Through extensive theoretical and empirical analysis, we gain valuable insights into the strengths and weaknesses of prior algorithms. Our results indicate that there is no universal solution for all possible cases. Finally, we provide guidelines to help researchers select appropriate mechanisms for various scenarios., Comment: 18 pages, accepted by ICDE 2025

Deep convolutional neural networks have made significant breakthroughs in medical image classification, under the assumption that training samples from all classes are simultaneously available. However, in real-world medical scenarios, there's a common need to continuously learn about new diseases, leading to the emerging field of class incremental learning (CIL) in the medical domain. Typically, CIL suffers from catastrophic forgetting when trained on new classes. This phenomenon is mainly caused by the imbalance between old and new classes, and it becomes even more challenging with imbalanced medical datasets. In this work, we introduce two simple yet effective plug-in methods to mitigate the adverse effects of the imbalance. First, we propose a CIL-balanced classification loss to mitigate the classifier bias toward majority classes via logit adjustment. Second, we propose a distribution margin loss that not only alleviates the inter-class overlap in embedding space but also enforces the intra-class compactness. We evaluate the effectiveness of our method with extensive experiments on three benchmark datasets (CCH5000, HAM10000, and EyePACS). The results demonstrate that our approach outperforms state-of-the-art methods., Comment: Accepted by ACM MM 2024

As a heavy molecule, barium monofluoride (BaF) presents itself as a promising candidate for measuring permanent electric dipole moment. The precision of such measurements can be significantly enhanced by utilizing a cold molecular sample. Here we report the realization of three-dimensional magneto-optical trapping (MOT) of BaF molecules. Through the repumping of all the vibrational states up to $v=3$, and rotational states up to $N=3$, we effectively close the transition to a leakage level lower than $10^{-5}$. This approach enables molecules to scatter a sufficient number of photons required for laser cooling and trapping. By employing a technique that involves chirping the slowing laser frequency, BaF molecules are decelerated to near-zero velocity, resulting in the capture of approximately $3\times 10^3$ molecules in a dual-frequency MOT setup. Our findings represent a significant step towards the realization of ultracold BaF molecules and the conduct of precision measurements with cold molecules., Comment: 5 pages, 4 figures

Graph condensation (GC), which reduces the size of a large-scale graph by synthesizing a small-scale condensed graph as its substitution, has benefited various graph learning tasks. However, existing GC methods rely on centralized data storage, which is unfeasible for real-world decentralized data distribution, and overlook data holders' privacy-preserving requirements. To bridge this gap, we propose and study the novel problem of federated graph condensation (FGC) for graph neural networks (GNNs). Specifically, we first propose a general framework for FGC, where we decouple the typical gradient matching process for GC into client-side gradient calculation and server-side gradient matching, integrating knowledge from multiple clients' subgraphs into one smaller condensed graph. Nevertheless, our empirical studies show that under the federated setting, the condensed graph will consistently leak data membership privacy, i.e., the condensed graph during federated training can be utilized to steal training data under the membership inference attack (MIA). To tackle this issue, we innovatively incorporate information bottleneck principles into the FGC, which only needs to extract partial node features in one local pre-training step and utilize the features during federated training. Theoretical and experimental analyses demonstrate that our framework consistently protects membership privacy during training. Meanwhile, it can achieve comparable and even superior performance against existing centralized GC and federated graph learning (FGL) methods., Comment: 14 pages. Accepted by AAAI 2025

Traditional audio-visual methods rely on independent audio and visual backbones, which is costly and not scalable. In this work, we investigate using an audio-visual siamese network (AVSiam) for efficient and scalable audio-visual pretraining. Our framework uses a single shared vision transformer backbone to process audio and visual inputs, improving its parameter efficiency, reducing the GPU memory footprint, and allowing us to scale our method to larger datasets and model sizes. We pretrain our model using a contrastive audio-visual matching objective with a multi-ratio random masking scheme, which enables our model to process larger audio-visual instance batches, helpful for contrastive learning. Unlike prior audio-visual methods, our method can robustly handle audio, visual, and audio-visual inputs with a single shared ViT backbone. Furthermore, despite using the shared backbone for both modalities, AVSiam achieves competitive or even better results than prior methods on AudioSet and VGGSound for audio-visual classification and retrieval. Our code is available at https://github.com/GenjiB/AVSiam

Medical image segmentation plays a vital role in clinic disease diagnosis and medical image analysis. However, labeling medical images for segmentation task is tough due to the indispensable domain expertise of radiologists. Furthermore, considering the privacy and sensitivity of medical images, it is impractical to build a centralized segmentation dataset from different medical institutions. Federated learning aims to train a shared model of isolated clients without local data exchange which aligns well with the scarcity and privacy characteristics of medical data. To solve the problem of labeling hard, many advanced semi-supervised methods have been proposed in a centralized data setting. As for federated learning, how to conduct semi-supervised learning under this distributed scenario is worth investigating. In this work, we propose a novel federated semi-supervised learning framework for medical image segmentation. The intra-client and inter-client consistency learning are introduced to smooth predictions at the data level and avoid confirmation bias of local models. They are achieved with the assistance of a Variational Autoencoder (VAE) trained collaboratively by clients. The added VAE model plays three roles: 1) extracting latent low-dimensional features of all labeled and unlabeled data; 2) performing a novel type of data augmentation in calculating intra-client consistency loss; 3) utilizing the generative ability of itself to conduct inter-client consistency distillation. The proposed framework is compared with other federated semi-supervised or self-supervised learning methods. The experimental results illustrate that our method outperforms the state-of-the-art method while avoiding a lot of computation and communication overhead., Comment: Working in progress

$k$-core is a subgraph where every node has at least $k$ neighbors within the subgraph. The $k$-core subgraphs has been employed in large platforms like Network Repository to comprehend the underlying structures and dynamics of the network. Existing studies have primarily focused on finding $k$-core groups without considering their size, despite the relevance of solution sizes in many real-world scenarios. This paper addresses this gap by introducing the size-prescribed $k$-core search (SPCS) problem, where the goal is to find a subgraph of a specified size that has the highest possible core number. We propose two algorithms, namely the {\it TSizeKcore-BU} and the {\it TSizeKcore-TD}, to identify cohesive subgraphs that satisfy both the $k$-core requirement and the size constraint. Our experimental results demonstrate the superiority of our approach in terms of solution quality and efficiency. The {\it TSizeKcore-BU} algorithm proves to be highly efficient in finding size-prescribed $k$-core subgraphs on large datasets, making it a favorable choice for such scenarios. On the other hand, the {\it TSizeKcore-TD} algorithm is better suited for small datasets where running time is less critical.

We present a parameter-efficient method for continual video question-answering (VidQA) learning. Our method, named DAM, uses the proposed Dynamic Adapter Merging to (i) mitigate catastrophic forgetting, (ii) enable efficient adaptation to continually arriving datasets, (iii) handle inputs from unknown datasets during inference, and (iv) enable knowledge sharing across similar dataset domains. Given a set of continually streaming VidQA datasets, we sequentially train dataset-specific adapters for each dataset while freezing the parameters of a large pretrained video-language backbone. During inference, given a video-question sample from an unknown domain, our method first uses the proposed non-parametric router function to compute a probability for each adapter, reflecting how relevant that adapter is to the current video-question input instance. Subsequently, the proposed dynamic adapter merging scheme aggregates all the adapter weights into a new adapter instance tailored for that particular test sample to compute the final VidQA prediction, mitigating the impact of inaccurate router predictions and facilitating knowledge sharing across domains. Our DAM model outperforms prior state-of-the-art continual learning approaches by 9.1% while exhibiting 1.9% less forgetting on 6 VidQA datasets spanning various domains. We further extend DAM to continual image classification and image QA and outperform prior methods by a large margin. The code is publicly available at: https://github.com/klauscc/DAM, Comment: The first two authors contribute equally

Foundation models, pre-trained on massive datasets, have achieved unprecedented generalizability. However, is it truly necessary to involve such vast amounts of data in pre-training, consuming extensive computational resources? This paper introduces data-effective learning, aiming to use data in the most impactful way to pre-train foundation models. This involves strategies that focus on data quality rather than quantity, ensuring the data used for training has high informational value. Data-effective learning plays a profound role in accelerating foundation model training, reducing computational costs, and saving data storage, which is very important as the volume of medical data in recent years has grown beyond many people's expectations. However, due to the lack of standards and comprehensive benchmarks, research on medical data-effective learning is poorly studied. To address this gap, our paper introduces a comprehensive benchmark specifically for evaluating data-effective learning in the medical field. This benchmark includes a dataset with millions of data samples from 31 medical centers (DataDEL), a baseline method for comparison (MedDEL), and a new evaluation metric (NormDEL) to objectively measure data-effective learning performance. Our extensive experimental results show the baseline MedDEL can achieve performance comparable to the original large dataset with only 5% of the data. Establishing such an open data-effective learning benchmark is crucial for the medical foundation model research community because it facilitates efficient data use, promotes collaborative breakthroughs, and fosters the development of cost-effective, scalable, and impactful healthcare solutions.

Surface currents emerge in superconductors exposed to magnetic fields, and are a key signature of the Meissner effect. Analogously, chiral dynamics were observed in quantum simulators under synthetic Abelian gauge fields. The flexible control of these simulators also facilitates the engineering of non-Abelian gauge fields, but their impact on the chiral dynamics remains elusive. Here, by employing the cutting-edge momentum-lattice technique, we implement a synthetic SU(2) gauge field in a spinful 1D ladder and study the rich chiral dynamics therein. We confirm the non-Abelian nature of the synthetic potential by observing the non-Abelian Aharonov-Bohm effect on a single plaquette. More importantly, the chiral current along the two legs of the ladder is observed to be spin-dependent and highly tunable through the parameters of the gauge potential. We experimentally map out different dynamic regimes of the chiral current, and reveal the underlying competition between overlaying flux ladders with distinct spin compositions. Our experiment demonstrates the dramatic impact of non-Abelian gauge fields on the system dynamics, paving the way for future studies of exotic synthetic gauge fields on the versatile platform of momentum lattices., Comment: 6 pages, 4 figures

The manipulation of particle transport in synthetic quantum matter is an active research frontier for its theoretical importance and potential applications. Here we experimentally demonstrate an engineered topological transport in a synthetic flat-band lattice of ultracold $^{87}$Rb atoms. We implement a quasi-one-dimensional rhombic chain with staggered flux in the momentum space of the atomic condensate and observe biased local oscillations that originate from the interplay of the staggered flux and flat-band localization under the mechanism of Aharonov-Bohm caging. Based on these features, we design and experimentally confirm a state-dependent chiral transport under the periodic modulation of the synthetic flux. We show that the phenomenon is topologically protected by the winding of the Floquet Bloch bands of a coarse-grained effective Hamiltonian. The observed chiral transport offers a strategy for efficient quantum device design where topological robustness is ensured by fast Floquet driving and flat-band localization., Comment: 5 pages, 3 figures

With the rise of real-time rendering and the evolution of display devices, there is a growing demand for post-processing methods that offer high-resolution content in a high frame rate. Existing techniques often suffer from quality and latency issues due to the disjointed treatment of frame supersampling and extrapolation. In this paper, we recognize the shared context and mechanisms between frame supersampling and extrapolation, and present a novel framework, Space-time Supersampling (STSS). By integrating them into a unified framework, STSS can improve the overall quality with lower latency. To implement an efficient architecture, we treat the aliasing and warping holes unified as reshading regions and put forth two key components to compensate the regions, namely Random Reshading Masking (RRM) and Efficient Reshading Module (ERM). Extensive experiments demonstrate that our approach achieves superior visual fidelity compared to state-of-the-art (SOTA) methods. Notably, the performance is achieved within only 4ms, saving up to 75\% of time against the conventional two-stage pipeline that necessitates 17ms., Comment: Accepted to AAAI 2024

Existing recurrent optical flow estimation networks are computationally expensive since they use a fixed large number of iterations to update the flow field for each sample. An efficient network should skip iterations when the flow improvement is limited. In this paper, we develop a Context-Aware Iteration Policy Network for efficient optical flow estimation, which determines the optimal number of iterations per sample. The policy network achieves this by learning contextual information to realize whether flow improvement is bottlenecked or minimal. On the one hand, we use iteration embedding and historical hidden cell, which include previous iterations information, to convey how flow has changed from previous iterations. On the other hand, we use the incremental loss to make the policy network implicitly perceive the magnitude of optical flow improvement in the subsequent iteration. Furthermore, the computational complexity in our dynamic network is controllable, allowing us to satisfy various resource preferences with a single trained model. Our policy network can be easily integrated into state-of-the-art optical flow networks. Extensive experiments show that our method maintains performance while reducing FLOPs by about 40%/20% for the Sintel/KITTI datasets., Comment: 2024, Association for the Advancement of Artificial Intelligence

The heterogeneous information network (HIN), which contains rich semantics depicted by meta-paths, has emerged as a potent tool for mitigating data sparsity in recommender systems. Existing HIN-based recommender systems operate under the assumption of centralized storage and model training. However, real-world data is often distributed due to privacy concerns, leading to the semantic broken issue within HINs and consequent failures in centralized HIN-based recommendations. In this paper, we suggest the HIN is partitioned into private HINs stored on the client side and shared HINs on the server. Following this setting, we propose a federated heterogeneous graph neural network (FedHGNN) based framework, which facilitates collaborative training of a recommendation model using distributed HINs while protecting user privacy. Specifically, we first formalize the privacy definition for HIN-based federated recommendation (FedRec) in the light of differential privacy, with the goal of protecting user-item interactions within private HIN as well as users' high-order patterns from shared HINs. To recover the broken meta-path based semantics and ensure proposed privacy measures, we elaborately design a semantic-preserving user interactions publishing method, which locally perturbs user's high-order patterns and related user-item interactions for publishing. Subsequently, we introduce an HGNN model for recommendation, which conducts node- and semantic-level aggregations to capture recovered semantics. Extensive experiments on four datasets demonstrate that our model outperforms existing methods by a substantial margin (up to 34% in HR@10 and 42% in NDCG@10) under a reasonable privacy budget., Comment: Accepted by WWW 2024

Self-supervised representation learning has gained increasing attention for strong generalization ability without relying on paired datasets. However, it has not been explored sufficiently for facial representation. Self-supervised facial representation learning remains unsolved due to the coupling of facial identities, expressions, and external factors like pose and light. Prior methods primarily focus on contrastive learning and pixel-level consistency, leading to limited interpretability and suboptimal performance. In this paper, we propose LatentFace, a novel generative framework for self-supervised facial representations. We suggest that the disentangling problem can be also formulated as generative objectives in space and time, and propose the solution using a 3D-aware latent diffusion model. First, we introduce a 3D-aware autoencoder to encode face images into 3D latent embeddings. Second, we propose a novel representation diffusion model to disentangle 3D latent into facial identity and expression. Consequently, our method achieves state-of-the-art performance in facial expression recognition (FER) and face verification among self-supervised facial representation learning models. Our model achieves a 3.75\% advantage in FER accuracy on RAF-DB and 3.35\% on AffectNet compared to SOTA methods.

We consider excited configuration and multistate configuration of rotating axion boson stars~(RABSs).RABSs are asymptotically flat, stationary, spinning, horizonless solutions of Einstein-Klein-Gordon theory in which the scalar potential depends on scalar field mass $\mu$ and axion decay constant $f_a$. The excited RABSs have two types of solutions, including $^2S$ state and $^2P$ state. The rotating multistate axion boson stars~(RMABSs) consist of coupled fundamental configuration and excited configuration, and also include two types of solutions: $^1S^2S$ state and $^1S^2P$ state. Some differences between RABSs models and rotating mini-boson stars models are discussed. We show the solution space of these models for different values of decay constant $f_a$. We found fundamental RABSs have a higher maximum mass than the excited state in the low decay constant region. Moreover, the multistate configuration allows a higher mass than both the fundamental configuration and the excited configuration at the same frequency. We also found the RMABSs have the second branch in which the ergo-region emerges. This means the second branch may be superradiant unstable., Comment: 18 pages, 9 figures

In recent years, many deep learning-based methods have been proposed to tackle the problem of optical flow estimation and achieved promising results. However, they hardly consider that most videos are compressed and thus ignore the pre-computed information in compressed video streams. Motion vectors, one of the compression information, record the motion of the video frames. They can be directly extracted from the compression code stream without computational cost and serve as a solid prior for optical flow estimation. Therefore, we propose an optical flow model, MVFlow, which uses motion vectors to improve the speed and accuracy of optical flow estimation for compressed videos. In detail, MVFlow includes a key Motion-Vector Converting Module, which ensures that the motion vectors can be transformed into the same domain of optical flow and then be utilized fully by the flow estimation module. Meanwhile, we construct four optical flow datasets for compressed videos containing frames and motion vectors in pairs. The experimental results demonstrate the superiority of our proposed MVFlow, which can reduce the AEPE by 1.09 compared to existing models or save 52% time to achieve similar accuracy to existing models., Comment: Accepted by ACM MM 2023