Deep learning has emerged as a powerful tool for predicting the remaining useful life(RUL)of batteries,contingent upon access to ample data.However,the inherent limitations of data availability from traditional or acc...Deep learning has emerged as a powerful tool for predicting the remaining useful life(RUL)of batteries,contingent upon access to ample data.However,the inherent limitations of data availability from traditional or accelerated life testing pose significant challenges.To mitigate the prediction accuracy issues arising from small sample sizes in existing intelligent methods,we introduce a novel data augmentation framework for RUL prediction.This framework harnesses the inherent high coincidence of degradation patterns exhibited by lithium-ion batteries to pinpoint the knee point,a critical juncture marking a significant shift in the degradation trajectory.By focusing on this critical knee point,we leverage the power of normalizing flow models to generate virtual data,effectively augmenting the training sample size.Additionally,we integrate a Bayesian Long Short-Term Memory network,optimized with Box-Cox transformation,to address the inherent uncertainty associated with predictions based on augmented data.This integration allows for a more nuanced understanding of RUL prediction uncertainties,offering valuable confidence intervals.The efficacy and superiority of the proposed framework are validated through extensive experiments on the CS2 dataset from the University of Maryland and the CrFeMnNiCo dataset from our laboratory.The results clearly demonstrate a substantial improvement in the confidence interval of RUL predictions compared to pre-optimization,highlighting the ability of the framework to achieve high-precision RUL predictions even with limited data.展开更多
DDoS attacks represent one of the most pervasive and evolving threats in cybersecurity,capable of crippling critical infrastructures and disrupting services globally.As networks continue to expand and threats become m...DDoS attacks represent one of the most pervasive and evolving threats in cybersecurity,capable of crippling critical infrastructures and disrupting services globally.As networks continue to expand and threats become more sophisticated,there is an urgent need for Intrusion Detection Systems(IDS)capable of handling these challenges effectively.Traditional IDS models frequently have difficulties in detecting new or changing attack patterns since they heavily depend on existing characteristics.This paper presents a novel approach for detecting unknown Distributed Denial of Service(DDoS)attacks by integrating Sliced Iterative Normalizing Flows(SINF)into IDS.SINF utilizes the Sliced Wasserstein distance to repeatedly modify probability distributions,enabling better management of high-dimensional data when there are only a few samples available.The unique architecture of SINF ensures efficient density estimation and robust sample generation,enabling IDS to adapt dynamically to emerging threats without relying heavily on predefined signatures or extensive retraining.By incorporating Open-Set Recognition(OSR)techniques,this method improves the system’s ability to detect both known and unknown attacks while maintaining high detection performance.The experimental evaluation on CICIDS2017 and CICDDoS2019 datasets demonstrates that the proposed system achieves an accuracy of 99.85%for known attacks and an F1 score of 99.99%after incremental learning for unknown attacks.The results clearly demonstrate the system’s strong generalization capability across unseen attacks while maintaining the computational efficiency required for real-world deployment.展开更多
Unsupervised methods based on density representation have shown their abilities in anomaly detection,but detection performance still needs to be improved.Specifically,approaches using normalizing flows can accurately ...Unsupervised methods based on density representation have shown their abilities in anomaly detection,but detection performance still needs to be improved.Specifically,approaches using normalizing flows can accurately evaluate sample distributions,mapping normal features to the normal distribution and anomalous features outside it.Consequently,this paper proposes a Normalizing Flow-based Bidirectional Mapping Residual Network(NF-BMR).It utilizes pre-trained Convolutional Neural Networks(CNN)and normalizing flows to construct discriminative source and target domain feature spaces.Additionally,to better learn feature information in both domain spaces,we propose the Bidirectional Mapping Residual Network(BMR),which maps sample features to these two spaces for anomaly detection.The two detection spaces effectively complement each other’s deficiencies and provide a comprehensive feature evaluation from two perspectives,which leads to the improvement of detection performance.Comparative experimental results on the MVTec AD and DAGM datasets against the Bidirectional Pre-trained Feature Mapping Network(B-PFM)and other state-of-the-art methods demonstrate that the proposed approach achieves superior performance.On the MVTec AD dataset,NF-BMR achieves an average AUROC of 98.7%for all 15 categories.Especially,it achieves 100%optimal detection performance in five categories.On the DAGM dataset,the average AUROC across ten categories is 98.7%,which is very close to supervised methods.展开更多
Monocular 3D object detection is challenging due to the lack of accurate depth information.Some methods estimate the pixel-wise depth maps from off-the-shelf depth estimators and then use them as an additional input t...Monocular 3D object detection is challenging due to the lack of accurate depth information.Some methods estimate the pixel-wise depth maps from off-the-shelf depth estimators and then use them as an additional input to augment the RGB images.Depth-based methods attempt to convert estimated depth maps to pseudo-LiDAR and then use LiDAR-based object detectors or focus on the perspective of image and depth fusion learning.However,they demonstrate limited performance and efficiency as a result of depth inaccuracy and complex fusion mode with convolutions.Different from these approaches,our proposed depth-guided vision transformer with a normalizing flows(NF-DVT)network uses normalizing flows to build priors in depth maps to achieve more accurate depth information.Then we develop a novel Swin-Transformer-based backbone with a fusion module to process RGB image patches and depth map patches with two separate branches and fuse them using cross-attention to exchange information with each other.Furthermore,with the help of pixel-wise relative depth values in depth maps,we develop new relative position embeddings in the cross-attention mechanism to capture more accurate sequence ordering of input tokens.Our method is the first Swin-Transformer-based backbone architecture for monocular 3D object detection.The experimental results on the KITTI and the challenging Waymo Open datasets show the effectiveness of our proposed method and superior performance over previous counterparts.展开更多
Scenario generation is a critical step in stochastic programming for energy systems applications,where accurate representation of uncertainty directly impacts the decision quality.Normalizing flows(NFs),a class of inv...Scenario generation is a critical step in stochastic programming for energy systems applications,where accurate representation of uncertainty directly impacts the decision quality.Normalizing flows(NFs),a class of invertible deep generative models,offer flexibility in learning complex distributions by maximizing the likelihood,but often suffer from limited accuracy in reproducing key statistical properties of real-world data.In this work we propose a moments-informed Normalizing Flows(MI-NF)framework,in which moment constraints are incorporated into the NF training process to improve the accuracy of scenario-based probabilistic forecasts.thermore,Fur-Gaussian Processes(GPs)are employed to adaptively determine the moment regularization weight.Case studies on the open-access dataset of the Global Energy Forecasting Competition 2014 demonstrate that scenarios generated by the MI-NF model achieve over 40%lower mean absolute error on the testing set.When applied within a stochastic programming framework for a local electricity-hydrogen market,the improved scenario accuracy leads to more cost-effective and robust operational decisions under uncertainty.展开更多
We propose a normalizing flow based on the wavelet framework for super-resolution(SR)called WDFSR.It learns the conditional distribution mapping between low-resolution images in the RGB domain and high-resolution imag...We propose a normalizing flow based on the wavelet framework for super-resolution(SR)called WDFSR.It learns the conditional distribution mapping between low-resolution images in the RGB domain and high-resolution images in the wavelet domain to simultaneously generate high-resolution images of different styles.To address the issue of some flowbased models being sensitive to datasets,which results in training fluctuations that reduce the mapping ability of the model and weaken generalization,we designed a method that combines a T-distribution and QR decomposition layer.Our method alleviates this problem while maintaining the ability of the model to map different distributions and produce higher-quality images.Good contextual conditional features can promote model training and enhance the distribution mapping capabilities for conditional distribution mapping.Therefore,we propose a Refinement layer combined with an attention mechanism to refine and fuse the extracted condition features to improve image quality.Extensive experiments on several SR datasets demonstrate that WDFSR outperforms most general CNN-and flow-based models in terms of PSNR value and perception quality.We also demonstrated that our framework works well for other low-level vision tasks,such as low-light enhancement.The pretrained models and source code with guidance for reference are available at https://github.com/Lisbegin/WDFSR.展开更多
In this work,we propose an adaptive learning approach based on temporal normalizing flows for solving time-dependent Fokker-Planck(TFP)equations.It is well known that solutions of such equations are probability densit...In this work,we propose an adaptive learning approach based on temporal normalizing flows for solving time-dependent Fokker-Planck(TFP)equations.It is well known that solutions of such equations are probability density functions,and thus our approach relies on modelling the target solutions with the temporal normalizing flows.The temporal normalizing flow is then trained based on the TFP loss function,without requiring any labeled data.Being a machine learning scheme,the proposed approach is mesh-free and can be easily applied to high dimensional problems.We present a variety of test problems to show the effectiveness of the learning approach.展开更多
Glitches represent a category of non-Gaussian and transient noise that frequently intersects with gravitational wave(GW)signals,thereby exerting a notable impact on the processing of GW data.The inference of GW parame...Glitches represent a category of non-Gaussian and transient noise that frequently intersects with gravitational wave(GW)signals,thereby exerting a notable impact on the processing of GW data.The inference of GW parameters,crucial for GW astronomy research,is particularly susceptible to such interference.In this study,we pioneer the utilization of a temporal and time-spectral fusion normalizing flow for likelihood-free inference of GW parameters,seamlessly integrating the high temporal resolution of the time domain with the frequency separation characteristics of both time and frequency domains.Remarkably,our findings indicate that the accuracy of this inference method is comparable to that of traditional non-glitch sampling techniques.Furthermore,our approach exhibits a greater efficiency,boasting processing times on the order of milliseconds.In conclusion,the application of a normalizing flow emerges as pivotal in handling GW signals affected by transient noises,offering a promising avenue for enhancing the field of GW astronomy research.展开更多
We introduce a novel method using a new generative model that automatically learns effective representations of the target and background appearance to detect,segment and track each instance in a video sequence.Differ...We introduce a novel method using a new generative model that automatically learns effective representations of the target and background appearance to detect,segment and track each instance in a video sequence.Differently from current discriminative tracking-by-detection solutions,our proposed hierarchical structural embedding learning can predict more highquality masks with accurate boundary details over spatio-temporal space via the normalizing flows.We formulate the instance inference procedure as a hierarchical spatio-temporal embedded learning across time and space.Given the video clip,our method first coarsely locates pixels belonging to a particular instance with Gaussian distribution and then builds a novel mixing distribution to promote the instance boundary by fusing hierarchical appearance embedding information in a coarse-to-fine manner.For the mixing distribution,we utilize a factorization condition normalized flow fashion to estimate the distribution parameters to improve the segmentation performance.Comprehensive qualitative,quantitative,and ablation experiments are performed on three representative video instance segmentation benchmarks(i.e.,YouTube-VIS19,YouTube-VIS21,and OVIS)and the effectiveness of the proposed method is demonstrated.More impressively,the superior performance of our model on an unsupervised video object segmentation dataset(i.e.,DAVIS19)proves its generalizability.Our algorithm implementations are publicly available at https://github.com/zyqin19/HEVis.展开更多
It is essential to understand the water consumption characteristics and physiological adjustments of tree species under drought conditions,as well as the effects of pure and mixed plantations on these characteristics ...It is essential to understand the water consumption characteristics and physiological adjustments of tree species under drought conditions,as well as the effects of pure and mixed plantations on these characteristics in semi-arid regions.In this study,the normalized sap flow(SFn),leaf water potential,stomatal conductance(gs),and photosynthetic rate(Pr)were monitored for two dominant species,i.e.,Pinus tabuliformis and Hippophae rhamnoides,in both pure and mixed plantations in a semi-arid region of Chinese Loess Plateau.A threshold-delay model showed that the lower rainfall thresholds(RL)for P.tabuliformis and H.rhamnoides in pure plantations were 9.6 and 11.0 mm,respectively,and the time lags(τ)after rainfall were 1.15 and 1.76 d for corresponding species,respectively.The results indicated that P.tabuliformis was more sensitive to rainfall pulse than H.rhamnoides.In addition,strong stomatal control allowed P.tabuliformis to experience low gsand Prin response to drought,while maintaining a high midday leaf water potential(Ψm).However,H.rhamnoides maintained high gsand Prat a lowΨmexpense.Therefore,P.tabuliformis and H.rhamnoides can be considered as isohydric and anisohydric species,respectively.In mixed plantation,the values of RLfor P.tabuliformis and H.rhamnoides were 6.5 and 8.9 mm,respectively;and the values ofτwere 0.86 and 1.61 d for corresponding species,respectively,which implied that mixed afforestation enhanced the rainfall pulse sensitivity for both two species,especially for P.tabuliformis.In addition,mixed afforestation significantly reduced SFn,gs,and Prfor P.tabuliformis(P<0.05),while maintaining a high leaf water potential status.However,no significant effect of mixed afforestation of H.rhamnoides was observed at the expense of leaf water potential status in response to drought.Although inconsistent physiological responses were adopted by these species,the altered water consumption characteristics,especially for P.tabuliformis indicated that the mixed afforestation requires further investigation.展开更多
Gradient vector flow （GVF） is an effective external force for active contours, but its iso- tropic nature handicaps its performance. The recently proposed gradient vector flow in the normal direction （NGVF） is ani...Gradient vector flow （GVF） is an effective external force for active contours, but its iso- tropic nature handicaps its performance. The recently proposed gradient vector flow in the normal direction （NGVF） is anisotropic since it only keeps the diffusion along the normal direction of the isophotes; however, it has difficulties forcing a snake into long, thin boundary indentations. In this paper, a novel external force for active contours called normally generalized gradient vector flow （NGGVF） is proposed, which generalizes the NGVF formulation to include two spatially varying weighting functions. Consequently, the proposed NGGVF snake is anisotropic and would improve ac- tive contour convergence into long, thin boundary indentations while maintaining other desirable properties of the NGVF snake, such as enlarged capture range, initialization insensitivity and good convergence at concavities. The advantages on synthetic and real images are demonstrated.展开更多
Extreme-mass-ratio inspiral(EMRI)signals pose significant challenges to gravitational wave(GW)data analysis,mainly owing to their highly complex waveforms and high-dimensional parameter space.Given their extended time...Extreme-mass-ratio inspiral(EMRI)signals pose significant challenges to gravitational wave(GW)data analysis,mainly owing to their highly complex waveforms and high-dimensional parameter space.Given their extended timescales of months to years and low signal-to-noise ratios,detecting and analyzing EMRIs with confidence generally relies on long-term observations.Besides the length of data,parameter estimation is particularly challenging due to non-local parameter degeneracies,arising from multiple local maxima,as well as flat regions and ridges inherent in the likelihood function.These factors lead to exceptionally high time complexity for parameter analysis based on traditional matched filtering and random sampling methods.To address these challenges,the present study explores a machine learning approach to Bayesian posterior estimation of EMRI signals,leveraging the recently developed flow matching technique based on ordinary differential equation neural networks.To our knowledge,this is also the first instance of applying continuous normalizing flows to EMRI analysis.Our approach demonstrates an increase in computational efficiency by several orders of magnitude compared to the traditional Markov chain Monte Carlo(MCMC)methods,while preserving the unbiasedness of results.However,we note that the posterior distributions generated by FMPE may exhibit broader uncertainty ranges than those obtained through full Bayesian sampling,requiring subsequent refinement via methods such as MCMC.Notably,when searching from large priors,our model rapidly approaches the true values while MCMC struggles to converge to the global maximum.Our findings highlight that machine learning has the potential to efficiently handle the vast EMRI parameter space of up to seventeen dimensions,offering new perspectives for advancing space-based GW detection and GW astronomy.展开更多
Using neural networks for supervised learning means learning a function that maps input <em>x</em> to output <em>y</em>. However, in many applications, the inverse learning is also wanted, <...Using neural networks for supervised learning means learning a function that maps input <em>x</em> to output <em>y</em>. However, in many applications, the inverse learning is also wanted, <em>i.e.</em>, inferring <em>y</em> from <em>x</em>, which requires invertibility of the learning. Since the dimension of input is usually much higher than that of the output, there is information loss in the forward learning from input to output. Thus, creating invertible neural networks is a difficult task. However, recent development of invertible learning techniques such as normalizing flows has made invertible neural networks a reality. In this work, we applied flow-based invertible neural networks as generative models to inverse molecule design. In this context, the forward learning is to predict chemical properties given a molecule, and the inverse learning is to infer the molecules given the chemical properties. Trained on 100 and 1000 molecules, respectively, from a benchmark dataset QM9, our model identified novel molecules that had chemical property values well exceeding the limits of the training molecules as well as the limits of the whole QM9 of 133,885 molecules, moreover our generative model could easily sample many molecules (<em>x</em> values) from any one chemical property value (<em>y</em> value). Compared with the previous method in the literature that could only optimize one molecule for one chemical property value at a time, our model could be trained once and then be sampled any multiple times and for any chemical property values without the need of retraining. This advantage comes from treating inverse molecule design as an inverse regression problem. In summary, our main contributions were two: 1) our model could generalize well from the training data and was very data efficient, 2) our model could learn bidirectional correspondence between molecules and their chemical properties, thereby offering the ability to sample any number of molecules from any <em>y</em> values. In conclusion, our findings revealed the efficiency and effectiveness of using invertible neural networks as generative models in inverse molecule design.展开更多
The traditional stochastic homogenization method can obtain homogenized solutions of elliptic problems with stationary random coefficients.However,many random composite materials in scientific and engineering computin...The traditional stochastic homogenization method can obtain homogenized solutions of elliptic problems with stationary random coefficients.However,many random composite materials in scientific and engineering computing do not satisfy the stationary assumption.To overcome the difficulty,we propose a normalizing field flow induced two-stage stochastic homogenization method to efficiently solve the random elliptic problem with non-stationary coefficients.By applying the two-stage stochastic homogenization method,the original elliptic equation with random and fast oscillatory coefficients is approximated as an equivalent elliptic equation,where the equivalent coefficients are obtained by solving a set of cell problems.Without the stationary assumption,the number of cell problems is large and the corresponding computational cost is high.To improve the efficiency,we apply the normalizing field flow model to learn a reference Gaussian field for the random equivalent coefficients based on a small amount of data,which is obtained by solving the cell problems with the finite element method.Numerical results demonstrate that the newly proposed method is efficient and accurate in tackling high dimensional partial differential equations in composite materials with complex random microstructures.展开更多
Objective To evaluate in vivo stability of ethylenedylbis cysteine diethylester (ECD) brain SPECT. Methods Each of 13 normal volunteers (31. 2 ± 11. 8 years) has 12 dynamic SPECK scans ac-quired in 60min 1h after...Objective To evaluate in vivo stability of ethylenedylbis cysteine diethylester (ECD) brain SPECT. Methods Each of 13 normal volunteers (31. 2 ± 11. 8 years) has 12 dynamic SPECK scans ac-quired in 60min 1h after an injection of 99mTc-ECD using a triple headed gamma camera equipped with ultra high resolution fan beam collimators. Average counts per pixel were measured from frontal, temporal, parie-tal, occipital regions, cerebellum, basal ganglia, thalamus and white matter. Regional ECD clearance rates, regional gray-to-white matter (G/W) ratios and the change of the G/W ratio were calculated. Results The average ECD clearance rate was 4. 2% /h, ranged from 3. 03% /h to 5. 41% /h corresponding to white matter and occipital. There was no significant difference between regional ECD clearance rates. Regional G 7W ratio was between 1.27 to 1.75. The G/W ratio of temporal lobe was lower than the occipital ( P <0.05). The change of regional G/W ratio with time is slow. Conclusion Regional ECD distribution is stable in normal brain. ECD clearance from brain is slow and no significant regional difference.展开更多
The application of multiple unmanned aerial vehicles(UAVs)for the pursuit and capture of unauthorized UAVs has emerged as a novel approach to ensuring the safety of urban airspace.However,pursuit UAVs necessitate the ...The application of multiple unmanned aerial vehicles(UAVs)for the pursuit and capture of unauthorized UAVs has emerged as a novel approach to ensuring the safety of urban airspace.However,pursuit UAVs necessitate the utilization of their own sensors to proactively gather information from the unauthorized UAV.Considering the restricted sensing range of sensors,this paper proposes a multi-UAV with limited visual field pursuit-evasion(MUV-PE)problem.Each pursuer has a visual field characterized by limited perception distance and viewing angle,potentially obstructed by buildings.Only when the unauthorized UAV,i.e.,the evader,enters the visual field of any pursuer can its position be acquired.The objective of the pursuers is to capture the evader as soon as possible without collision.To address this problem,we propose the normalizing flow actor with graph attention critic(NAGC)algorithm,a multi-agent reinforcement learning(MARL)approach.NAGC executes normalizing flows to augment the flexibility of policy network,enabling the agent to sample actions from more intricate distributions rather than common distributions.To enhance the capability of simultaneously comprehending spatial relationships among multiple UAVs and environmental obstacles,NAGC integrates the“obstacle-target”graph attention networks,significantly aiding pursuers in supporting search or pursuit activities.Extensive experiments conducted in a high-precision simulator validate the promising performance of the NAGC algorithm.展开更多
In this paper,the invariant geometric flows for hypersurfaces in centro-affine geometry are explored.We first present evolution equations of the centro-affine invariants corresponding to the geometric flows.Based on t...In this paper,the invariant geometric flows for hypersurfaces in centro-affine geometry are explored.We first present evolution equations of the centro-affine invariants corresponding to the geometric flows.Based on these fundamental evolution equations,we show that the centro-affine heat flow for hypersurfaces is equivalent to a system of ordinary differential equations,which can be solved explicitly.Finally,the centro-affine invariant normal flows for hypersurfaces are investigated,and two specific flows are provided to illustrate the behaviour of the flows.展开更多
In generating adversarial examples,the conventional black-box attack methods rely on sufficient feedback from the to-be-attacked models by repeatedly querying until the attack is successful,which usually results in th...In generating adversarial examples,the conventional black-box attack methods rely on sufficient feedback from the to-be-attacked models by repeatedly querying until the attack is successful,which usually results in thousands of trials during an attack.This may be unacceptable in real applications since Machine Learning as a Service Platform(MLaaS)usually only returns the final result(i.e.,hard-label)to the client and a system equipped with certain defense mechanisms could easily detect malicious queries.By contrast,a feasible way is a hard-label attack that simulates an attacked action being permitted to conduct a limited number of queries.To implement this idea,in this paper,we bypass the dependency on the to-be-attacked model and benefit from the characteristics of the distributions of adversarial examples to reformulate the attack problem in a distribution transform manner and propose a distribution transform-based attack(DTA).DTA builds a statistical mapping from the benign example to its adversarial counterparts by tackling the conditional likelihood under the hard-label black-box settings.In this way,it is no longer necessary to query the target model frequently.A well-trained DTA model can directly and efficiently generate a batch of adversarial examples for a certain input,which can be used to attack un-seen models based on the assumed transferability.Furthermore,we surprisingly find that the well-trained DTA model is not sensitive to the semantic spaces of the training dataset,meaning that the model yields acceptable attack performance on other datasets.Extensive experiments validate the effectiveness of the proposed idea and the superiority of DTA over the state-of-the-art.展开更多
Extracting knowledge from high-dimensional data has been notoriously difficult,primarily due to the so-called"curse of dimensionality"and the complex joint distributions of these dimensions.This is a particu...Extracting knowledge from high-dimensional data has been notoriously difficult,primarily due to the so-called"curse of dimensionality"and the complex joint distributions of these dimensions.This is a particularly profound issue for high-dimensional gravitational wave data analysis where one requires to conduct Bayesian inference and estimate joint posterior distributions.In this study,we incorporate prior physical knowledge by sampling from desired interim distributions to develop the training dataset.Accordingly,the more relevant regions of the high-dimensional feature space are covered by additional data points,such that the model can learn the subtle but important details.We adapt the normalizing flow method to be more expressive and trainable,such that the information can be effectively extracted and represented by the transformation between the prior and target distributions.Once trained,our model only takes approximately 1 s on one V100 GPU to generate thousands of samples for probabilistic inference purposes.The evaluation of our approach confirms the efficacy and efficiency of gravitational wave data inferences and points to a promising direction for similar research.The source code,specifications,and detailed procedures are publicly accessible on GitHub.展开更多
To generate dance that temporally and aesthetically matches the music is a challenging problem in three aspects.First,the generated motion should be beats-aligned to the local musical features.Second,the global aesthe...To generate dance that temporally and aesthetically matches the music is a challenging problem in three aspects.First,the generated motion should be beats-aligned to the local musical features.Second,the global aesthetic style should be matched between motion and music.And third,the generated motion should be diverse and non-self-repeating.To address these challenges,we propose ReChoreoNet,which re-choreographs high-quality dance motion for a given piece of music.A data-driven learning strategy is proposed to efficiently correlate the temporal connections between music and motion in a progressively learned cross-modality embedding space.The beats-aligned content motion will be subsequently used as autoregressive context and control signal to control a normalizing-flow model,which transfers the style of a prototype motion to the final generated dance.In addition,we present an aesthetically labelled music-dance repertoire(MDR)for both efficient learning of the cross-modality embedding,and understanding of the aesthetic connections between music and motion.We demonstrate that our repertoire-based framework is robustly extensible in both content and style.Both quantitative and qualitative experiments have been carried out to validate the efficiency of our proposed model.展开更多
基金supported by the National Natural Science Foundation of China(Grant No.62227814,52205040,22279070,and U21A20170)the Natural Science Basic Research Program of Shaanxi(2023-JC-QN-0140)+3 种基金the Young Talent Fund of Xi’an Association for Science and Technology(Grant No.959202313096)the Key Projects of the Shaanxi Province Natural Science Foundation(Grant No.2025JC-QYXQ-038)the Open Foundation of the State Key Laboratory of Fluid Power and Mechatronic Systems(Grant No.GZKF-202430)the National Key Research and Development Program of China(Grant No.2024YFB3311204)。
文摘Deep learning has emerged as a powerful tool for predicting the remaining useful life(RUL)of batteries,contingent upon access to ample data.However,the inherent limitations of data availability from traditional or accelerated life testing pose significant challenges.To mitigate the prediction accuracy issues arising from small sample sizes in existing intelligent methods,we introduce a novel data augmentation framework for RUL prediction.This framework harnesses the inherent high coincidence of degradation patterns exhibited by lithium-ion batteries to pinpoint the knee point,a critical juncture marking a significant shift in the degradation trajectory.By focusing on this critical knee point,we leverage the power of normalizing flow models to generate virtual data,effectively augmenting the training sample size.Additionally,we integrate a Bayesian Long Short-Term Memory network,optimized with Box-Cox transformation,to address the inherent uncertainty associated with predictions based on augmented data.This integration allows for a more nuanced understanding of RUL prediction uncertainties,offering valuable confidence intervals.The efficacy and superiority of the proposed framework are validated through extensive experiments on the CS2 dataset from the University of Maryland and the CrFeMnNiCo dataset from our laboratory.The results clearly demonstrate a substantial improvement in the confidence interval of RUL predictions compared to pre-optimization,highlighting the ability of the framework to achieve high-precision RUL predictions even with limited data.
基金supported by the National Science and Technology Council,Taiwan with grant numbers NSTC 112-2221-E-992-045,112-2221-E-992-057-MY3,and 112-2622-8-992-009-TD1.
文摘DDoS attacks represent one of the most pervasive and evolving threats in cybersecurity,capable of crippling critical infrastructures and disrupting services globally.As networks continue to expand and threats become more sophisticated,there is an urgent need for Intrusion Detection Systems(IDS)capable of handling these challenges effectively.Traditional IDS models frequently have difficulties in detecting new or changing attack patterns since they heavily depend on existing characteristics.This paper presents a novel approach for detecting unknown Distributed Denial of Service(DDoS)attacks by integrating Sliced Iterative Normalizing Flows(SINF)into IDS.SINF utilizes the Sliced Wasserstein distance to repeatedly modify probability distributions,enabling better management of high-dimensional data when there are only a few samples available.The unique architecture of SINF ensures efficient density estimation and robust sample generation,enabling IDS to adapt dynamically to emerging threats without relying heavily on predefined signatures or extensive retraining.By incorporating Open-Set Recognition(OSR)techniques,this method improves the system’s ability to detect both known and unknown attacks while maintaining high detection performance.The experimental evaluation on CICIDS2017 and CICDDoS2019 datasets demonstrates that the proposed system achieves an accuracy of 99.85%for known attacks and an F1 score of 99.99%after incremental learning for unknown attacks.The results clearly demonstrate the system’s strong generalization capability across unseen attacks while maintaining the computational efficiency required for real-world deployment.
基金This work was supported in part by the National Key R&D Program of China 2021YFE0110500in part by the National Natural Science Foundation of China under Grant 62062021in part by the Guiyang Scientific Plan Project[2023]48-11.
文摘Unsupervised methods based on density representation have shown their abilities in anomaly detection,but detection performance still needs to be improved.Specifically,approaches using normalizing flows can accurately evaluate sample distributions,mapping normal features to the normal distribution and anomalous features outside it.Consequently,this paper proposes a Normalizing Flow-based Bidirectional Mapping Residual Network(NF-BMR).It utilizes pre-trained Convolutional Neural Networks(CNN)and normalizing flows to construct discriminative source and target domain feature spaces.Additionally,to better learn feature information in both domain spaces,we propose the Bidirectional Mapping Residual Network(BMR),which maps sample features to these two spaces for anomaly detection.The two detection spaces effectively complement each other’s deficiencies and provide a comprehensive feature evaluation from two perspectives,which leads to the improvement of detection performance.Comparative experimental results on the MVTec AD and DAGM datasets against the Bidirectional Pre-trained Feature Mapping Network(B-PFM)and other state-of-the-art methods demonstrate that the proposed approach achieves superior performance.On the MVTec AD dataset,NF-BMR achieves an average AUROC of 98.7%for all 15 categories.Especially,it achieves 100%optimal detection performance in five categories.On the DAGM dataset,the average AUROC across ten categories is 98.7%,which is very close to supervised methods.
基金supported in part by the Major Project for New Generation of AI (2018AAA0100400)the National Natural Science Foundation of China (61836014,U21B2042,62072457,62006231)the InnoHK Program。
文摘Monocular 3D object detection is challenging due to the lack of accurate depth information.Some methods estimate the pixel-wise depth maps from off-the-shelf depth estimators and then use them as an additional input to augment the RGB images.Depth-based methods attempt to convert estimated depth maps to pseudo-LiDAR and then use LiDAR-based object detectors or focus on the perspective of image and depth fusion learning.However,they demonstrate limited performance and efficiency as a result of depth inaccuracy and complex fusion mode with convolutions.Different from these approaches,our proposed depth-guided vision transformer with a normalizing flows(NF-DVT)network uses normalizing flows to build priors in depth maps to achieve more accurate depth information.Then we develop a novel Swin-Transformer-based backbone with a fusion module to process RGB image patches and depth map patches with two separate branches and fuse them using cross-attention to exchange information with each other.Furthermore,with the help of pixel-wise relative depth values in depth maps,we develop new relative position embeddings in the cross-attention mechanism to capture more accurate sequence ordering of input tokens.Our method is the first Swin-Transformer-based backbone architecture for monocular 3D object detection.The experimental results on the KITTI and the challenging Waymo Open datasets show the effectiveness of our proposed method and superior performance over previous counterparts.
基金support from the Engineering&Physical Sciences Re-search Council(EPSRC),UK under the projects EP/T022930/1,EP/W003317/1 and EP/V051008/1,is gratefully acknowledged.
文摘Scenario generation is a critical step in stochastic programming for energy systems applications,where accurate representation of uncertainty directly impacts the decision quality.Normalizing flows(NFs),a class of invertible deep generative models,offer flexibility in learning complex distributions by maximizing the likelihood,but often suffer from limited accuracy in reproducing key statistical properties of real-world data.In this work we propose a moments-informed Normalizing Flows(MI-NF)framework,in which moment constraints are incorporated into the NF training process to improve the accuracy of scenario-based probabilistic forecasts.thermore,Fur-Gaussian Processes(GPs)are employed to adaptively determine the moment regularization weight.Case studies on the open-access dataset of the Global Energy Forecasting Competition 2014 demonstrate that scenarios generated by the MI-NF model achieve over 40%lower mean absolute error on the testing set.When applied within a stochastic programming framework for a local electricity-hydrogen market,the improved scenario accuracy leads to more cost-effective and robust operational decisions under uncertainty.
基金grateful to Zhejiang Gongshang University for its valuable computing resources and outstanding laboratory facilities,and support from the National Natural Science Foundation of China(Grant No.62172366)the Zhejiang Provincial Natural Science Foundation of China(Grant No.LY22F020013)+1 种基金“Pioneer”and“Leading Goose”R&D Program of Zhejiang Province(Grant No.2023C01150),Major Sci-Tech Innovation Project of Hangzhou City(Grant No.2022AIZD0110)“Digital+”Discipline Construction Project of Zhejiang Gongshang University(Grant No.SZJ2022B009).
文摘We propose a normalizing flow based on the wavelet framework for super-resolution(SR)called WDFSR.It learns the conditional distribution mapping between low-resolution images in the RGB domain and high-resolution images in the wavelet domain to simultaneously generate high-resolution images of different styles.To address the issue of some flowbased models being sensitive to datasets,which results in training fluctuations that reduce the mapping ability of the model and weaken generalization,we designed a method that combines a T-distribution and QR decomposition layer.Our method alleviates this problem while maintaining the ability of the model to map different distributions and produce higher-quality images.Good contextual conditional features can promote model training and enhance the distribution mapping capabilities for conditional distribution mapping.Therefore,we propose a Refinement layer combined with an attention mechanism to refine and fuse the extracted condition features to improve image quality.Extensive experiments on several SR datasets demonstrate that WDFSR outperforms most general CNN-and flow-based models in terms of PSNR value and perception quality.We also demonstrated that our framework works well for other low-level vision tasks,such as low-light enhancement.The pretrained models and source code with guidance for reference are available at https://github.com/Lisbegin/WDFSR.
基金supported by the NSF of China(under grant numbers 12288201 and 11731006)the National Key R&D Program of China(2020YFA0712000)the Strategic Priority Research Program of Chinese Academy of Sciences(Grant No.XDA25010404).
文摘In this work,we propose an adaptive learning approach based on temporal normalizing flows for solving time-dependent Fokker-Planck(TFP)equations.It is well known that solutions of such equations are probability density functions,and thus our approach relies on modelling the target solutions with the temporal normalizing flows.The temporal normalizing flow is then trained based on the TFP loss function,without requiring any labeled data.Being a machine learning scheme,the proposed approach is mesh-free and can be easily applied to high dimensional problems.We present a variety of test problems to show the effectiveness of the learning approach.
基金the National SKA Program of China(2022SKA0110200,2022SKA0110203)the National Natural Science Foundation of China(11975072,11875102,11835009)the National 111 Project(B16009)。
文摘Glitches represent a category of non-Gaussian and transient noise that frequently intersects with gravitational wave(GW)signals,thereby exerting a notable impact on the processing of GW data.The inference of GW parameters,crucial for GW astronomy research,is particularly susceptible to such interference.In this study,we pioneer the utilization of a temporal and time-spectral fusion normalizing flow for likelihood-free inference of GW parameters,seamlessly integrating the high temporal resolution of the time domain with the frequency separation characteristics of both time and frequency domains.Remarkably,our findings indicate that the accuracy of this inference method is comparable to that of traditional non-glitch sampling techniques.Furthermore,our approach exhibits a greater efficiency,boasting processing times on the order of milliseconds.In conclusion,the application of a normalizing flow emerges as pivotal in handling GW signals affected by transient noises,offering a promising avenue for enhancing the field of GW astronomy research.
基金supported in part by the National Natural Science Foundation of China(62176139,62106128,62176141)the Major Basic Research Project of Shandong Natural Science Foundation(ZR2021ZD15)+4 种基金the Natural Science Foundation of Shandong Province(ZR2021QF001)the Young Elite Scientists Sponsorship Program by CAST(2021QNRC001)the Open Project of Key Laboratory of Artificial Intelligence,Ministry of Educationthe Shandong Provincial Natural Science Foundation for Distinguished Young Scholars(ZR2021JQ26)the Taishan Scholar Project of Shandong Province(tsqn202103088)。
文摘We introduce a novel method using a new generative model that automatically learns effective representations of the target and background appearance to detect,segment and track each instance in a video sequence.Differently from current discriminative tracking-by-detection solutions,our proposed hierarchical structural embedding learning can predict more highquality masks with accurate boundary details over spatio-temporal space via the normalizing flows.We formulate the instance inference procedure as a hierarchical spatio-temporal embedded learning across time and space.Given the video clip,our method first coarsely locates pixels belonging to a particular instance with Gaussian distribution and then builds a novel mixing distribution to promote the instance boundary by fusing hierarchical appearance embedding information in a coarse-to-fine manner.For the mixing distribution,we utilize a factorization condition normalized flow fashion to estimate the distribution parameters to improve the segmentation performance.Comprehensive qualitative,quantitative,and ablation experiments are performed on three representative video instance segmentation benchmarks(i.e.,YouTube-VIS19,YouTube-VIS21,and OVIS)and the effectiveness of the proposed method is demonstrated.More impressively,the superior performance of our model on an unsupervised video object segmentation dataset(i.e.,DAVIS19)proves its generalizability.Our algorithm implementations are publicly available at https://github.com/zyqin19/HEVis.
基金supported by the National Key R&D Program of China (2017YFA0604801)the National Natural Science Foundation of China (41501576)+1 种基金the China Special Fund for Meteorological Research in the Public Interest (Major Projects) (GYHY201506001-3)the Fundamental Research Funds for the Central Universities (2452016105)
文摘It is essential to understand the water consumption characteristics and physiological adjustments of tree species under drought conditions,as well as the effects of pure and mixed plantations on these characteristics in semi-arid regions.In this study,the normalized sap flow(SFn),leaf water potential,stomatal conductance(gs),and photosynthetic rate(Pr)were monitored for two dominant species,i.e.,Pinus tabuliformis and Hippophae rhamnoides,in both pure and mixed plantations in a semi-arid region of Chinese Loess Plateau.A threshold-delay model showed that the lower rainfall thresholds(RL)for P.tabuliformis and H.rhamnoides in pure plantations were 9.6 and 11.0 mm,respectively,and the time lags(τ)after rainfall were 1.15 and 1.76 d for corresponding species,respectively.The results indicated that P.tabuliformis was more sensitive to rainfall pulse than H.rhamnoides.In addition,strong stomatal control allowed P.tabuliformis to experience low gsand Prin response to drought,while maintaining a high midday leaf water potential(Ψm).However,H.rhamnoides maintained high gsand Prat a lowΨmexpense.Therefore,P.tabuliformis and H.rhamnoides can be considered as isohydric and anisohydric species,respectively.In mixed plantation,the values of RLfor P.tabuliformis and H.rhamnoides were 6.5 and 8.9 mm,respectively;and the values ofτwere 0.86 and 1.61 d for corresponding species,respectively,which implied that mixed afforestation enhanced the rainfall pulse sensitivity for both two species,especially for P.tabuliformis.In addition,mixed afforestation significantly reduced SFn,gs,and Prfor P.tabuliformis(P<0.05),while maintaining a high leaf water potential status.However,no significant effect of mixed afforestation of H.rhamnoides was observed at the expense of leaf water potential status in response to drought.Although inconsistent physiological responses were adopted by these species,the altered water consumption characteristics,especially for P.tabuliformis indicated that the mixed afforestation requires further investigation.
基金Supported by the National Natural Science Foundation of China(60805004)the State Key Lab of Space Medicine Fundamen-tals and Application(SMFA09A16)
文摘Gradient vector flow （GVF） is an effective external force for active contours, but its iso- tropic nature handicaps its performance. The recently proposed gradient vector flow in the normal direction （NGVF） is anisotropic since it only keeps the diffusion along the normal direction of the isophotes; however, it has difficulties forcing a snake into long, thin boundary indentations. In this paper, a novel external force for active contours called normally generalized gradient vector flow （NGGVF） is proposed, which generalizes the NGVF formulation to include two spatially varying weighting functions. Consequently, the proposed NGGVF snake is anisotropic and would improve ac- tive contour convergence into long, thin boundary indentations while maintaining other desirable properties of the NGVF snake, such as enlarged capture range, initialization insensitivity and good convergence at concavities. The advantages on synthetic and real images are demonstrated.
基金supported by the National Key Research and Development Program of China(Grant Nos.2021YFC2201901,2021YFC2203004,2020YFC2200100 and 2021YFC2201903)International Partnership Program of the Chinese Academy of Sciences(Grant No.025GJHZ2023106GC)+4 种基金the financial support from Brazilian agencies Funda??o de AmparoàPesquisa do Estado de S?o Paulo(FAPESP)Funda??o de Amparoà Pesquisa do Estado do Rio Grande do Sul(FAPERGS)Fundacao de Amparoà Pesquisa do Estado do Rio de Janeiro(FAPERJ)Conselho Nacional de Desenvolvimento Científico e Tecnológico(CNPq)Coordenacao de Aperfeicoamento de Pessoal de Nível Superior(CAPES)。
文摘Extreme-mass-ratio inspiral(EMRI)signals pose significant challenges to gravitational wave(GW)data analysis,mainly owing to their highly complex waveforms and high-dimensional parameter space.Given their extended timescales of months to years and low signal-to-noise ratios,detecting and analyzing EMRIs with confidence generally relies on long-term observations.Besides the length of data,parameter estimation is particularly challenging due to non-local parameter degeneracies,arising from multiple local maxima,as well as flat regions and ridges inherent in the likelihood function.These factors lead to exceptionally high time complexity for parameter analysis based on traditional matched filtering and random sampling methods.To address these challenges,the present study explores a machine learning approach to Bayesian posterior estimation of EMRI signals,leveraging the recently developed flow matching technique based on ordinary differential equation neural networks.To our knowledge,this is also the first instance of applying continuous normalizing flows to EMRI analysis.Our approach demonstrates an increase in computational efficiency by several orders of magnitude compared to the traditional Markov chain Monte Carlo(MCMC)methods,while preserving the unbiasedness of results.However,we note that the posterior distributions generated by FMPE may exhibit broader uncertainty ranges than those obtained through full Bayesian sampling,requiring subsequent refinement via methods such as MCMC.Notably,when searching from large priors,our model rapidly approaches the true values while MCMC struggles to converge to the global maximum.Our findings highlight that machine learning has the potential to efficiently handle the vast EMRI parameter space of up to seventeen dimensions,offering new perspectives for advancing space-based GW detection and GW astronomy.
文摘Using neural networks for supervised learning means learning a function that maps input <em>x</em> to output <em>y</em>. However, in many applications, the inverse learning is also wanted, <em>i.e.</em>, inferring <em>y</em> from <em>x</em>, which requires invertibility of the learning. Since the dimension of input is usually much higher than that of the output, there is information loss in the forward learning from input to output. Thus, creating invertible neural networks is a difficult task. However, recent development of invertible learning techniques such as normalizing flows has made invertible neural networks a reality. In this work, we applied flow-based invertible neural networks as generative models to inverse molecule design. In this context, the forward learning is to predict chemical properties given a molecule, and the inverse learning is to infer the molecules given the chemical properties. Trained on 100 and 1000 molecules, respectively, from a benchmark dataset QM9, our model identified novel molecules that had chemical property values well exceeding the limits of the training molecules as well as the limits of the whole QM9 of 133,885 molecules, moreover our generative model could easily sample many molecules (<em>x</em> values) from any one chemical property value (<em>y</em> value). Compared with the previous method in the literature that could only optimize one molecule for one chemical property value at a time, our model could be trained once and then be sampled any multiple times and for any chemical property values without the need of retraining. This advantage comes from treating inverse molecule design as an inverse regression problem. In summary, our main contributions were two: 1) our model could generalize well from the training data and was very data efficient, 2) our model could learn bidirectional correspondence between molecules and their chemical properties, thereby offering the ability to sample any number of molecules from any <em>y</em> values. In conclusion, our findings revealed the efficiency and effectiveness of using invertible neural networks as generative models in inverse molecule design.
基金supported by the National Natural Science Foundation of China grant(12131002,51739007,12271409)Strategic Priority Research Program of the Chinese Academy of Sciences(XDC06030101)+2 种基金the National Key R&D Program of China with the grant(2020YFA-0713603)Natural Science Foundation of Shanghai grant(21ZR1465800)the Interdisciplinary Project in Ocean Research of Tongji University and the Fundamental Research Funds for the Central Universities..
文摘The traditional stochastic homogenization method can obtain homogenized solutions of elliptic problems with stationary random coefficients.However,many random composite materials in scientific and engineering computing do not satisfy the stationary assumption.To overcome the difficulty,we propose a normalizing field flow induced two-stage stochastic homogenization method to efficiently solve the random elliptic problem with non-stationary coefficients.By applying the two-stage stochastic homogenization method,the original elliptic equation with random and fast oscillatory coefficients is approximated as an equivalent elliptic equation,where the equivalent coefficients are obtained by solving a set of cell problems.Without the stationary assumption,the number of cell problems is large and the corresponding computational cost is high.To improve the efficiency,we apply the normalizing field flow model to learn a reference Gaussian field for the random equivalent coefficients based on a small amount of data,which is obtained by solving the cell problems with the finite element method.Numerical results demonstrate that the newly proposed method is efficient and accurate in tackling high dimensional partial differential equations in composite materials with complex random microstructures.
文摘Objective To evaluate in vivo stability of ethylenedylbis cysteine diethylester (ECD) brain SPECT. Methods Each of 13 normal volunteers (31. 2 ± 11. 8 years) has 12 dynamic SPECK scans ac-quired in 60min 1h after an injection of 99mTc-ECD using a triple headed gamma camera equipped with ultra high resolution fan beam collimators. Average counts per pixel were measured from frontal, temporal, parie-tal, occipital regions, cerebellum, basal ganglia, thalamus and white matter. Regional ECD clearance rates, regional gray-to-white matter (G/W) ratios and the change of the G/W ratio were calculated. Results The average ECD clearance rate was 4. 2% /h, ranged from 3. 03% /h to 5. 41% /h corresponding to white matter and occipital. There was no significant difference between regional ECD clearance rates. Regional G 7W ratio was between 1.27 to 1.75. The G/W ratio of temporal lobe was lower than the occipital ( P <0.05). The change of regional G/W ratio with time is slow. Conclusion Regional ECD distribution is stable in normal brain. ECD clearance from brain is slow and no significant regional difference.
基金supported in part by the National Natural Science Foundation of China(62373380)。
文摘The application of multiple unmanned aerial vehicles(UAVs)for the pursuit and capture of unauthorized UAVs has emerged as a novel approach to ensuring the safety of urban airspace.However,pursuit UAVs necessitate the utilization of their own sensors to proactively gather information from the unauthorized UAV.Considering the restricted sensing range of sensors,this paper proposes a multi-UAV with limited visual field pursuit-evasion(MUV-PE)problem.Each pursuer has a visual field characterized by limited perception distance and viewing angle,potentially obstructed by buildings.Only when the unauthorized UAV,i.e.,the evader,enters the visual field of any pursuer can its position be acquired.The objective of the pursuers is to capture the evader as soon as possible without collision.To address this problem,we propose the normalizing flow actor with graph attention critic(NAGC)algorithm,a multi-agent reinforcement learning(MARL)approach.NAGC executes normalizing flows to augment the flexibility of policy network,enabling the agent to sample actions from more intricate distributions rather than common distributions.To enhance the capability of simultaneously comprehending spatial relationships among multiple UAVs and environmental obstacles,NAGC integrates the“obstacle-target”graph attention networks,significantly aiding pursuers in supporting search or pursuit activities.Extensive experiments conducted in a high-precision simulator validate the promising performance of the NAGC algorithm.
基金This work was supported by National Natural Science Foundation of China(Grant Nos.11631007 and 11971251).
文摘In this paper,the invariant geometric flows for hypersurfaces in centro-affine geometry are explored.We first present evolution equations of the centro-affine invariants corresponding to the geometric flows.Based on these fundamental evolution equations,we show that the centro-affine heat flow for hypersurfaces is equivalent to a system of ordinary differential equations,which can be solved explicitly.Finally,the centro-affine invariant normal flows for hypersurfaces are investigated,and two specific flows are provided to illustrate the behaviour of the flows.
基金supported in part by the National Natural Science Foundation of China under Grant 62162067,62101480 and 62362068Research and Application of Object Detection based on Artificial Intelligence,in part by the Yunnan Province expert workstations under Grant 202305AF150078the Scientific Research Fund Project of Yunnan Provincial Education Department under 2023Y0249.
文摘In generating adversarial examples,the conventional black-box attack methods rely on sufficient feedback from the to-be-attacked models by repeatedly querying until the attack is successful,which usually results in thousands of trials during an attack.This may be unacceptable in real applications since Machine Learning as a Service Platform(MLaaS)usually only returns the final result(i.e.,hard-label)to the client and a system equipped with certain defense mechanisms could easily detect malicious queries.By contrast,a feasible way is a hard-label attack that simulates an attacked action being permitted to conduct a limited number of queries.To implement this idea,in this paper,we bypass the dependency on the to-be-attacked model and benefit from the characteristics of the distributions of adversarial examples to reformulate the attack problem in a distribution transform manner and propose a distribution transform-based attack(DTA).DTA builds a statistical mapping from the benign example to its adversarial counterparts by tackling the conditional likelihood under the hard-label black-box settings.In this way,it is no longer necessary to query the target model frequently.A well-trained DTA model can directly and efficiently generate a batch of adversarial examples for a certain input,which can be used to attack un-seen models based on the assumed transferability.Furthermore,we surprisingly find that the well-trained DTA model is not sensitive to the semantic spaces of the training dataset,meaning that the model yields acceptable attack performance on other datasets.Extensive experiments validate the effectiveness of the proposed idea and the superiority of DTA over the state-of-the-art.
基金supported by the Peng Cheng Laboratory Cloud Brain(No.PCL2021A13)the National Natural Science Foundation of China(Nos.11721303,12075297,and 11690021)the Strategic Priority Research Program of the Chinese Academy of Sciences(No.XDA1502110202)
文摘Extracting knowledge from high-dimensional data has been notoriously difficult,primarily due to the so-called"curse of dimensionality"and the complex joint distributions of these dimensions.This is a particularly profound issue for high-dimensional gravitational wave data analysis where one requires to conduct Bayesian inference and estimate joint posterior distributions.In this study,we incorporate prior physical knowledge by sampling from desired interim distributions to develop the training dataset.Accordingly,the more relevant regions of the high-dimensional feature space are covered by additional data points,such that the model can learn the subtle but important details.We adapt the normalizing flow method to be more expressive and trainable,such that the information can be effectively extracted and represented by the transformation between the prior and target distributions.Once trained,our model only takes approximately 1 s on one V100 GPU to generate thousands of samples for probabilistic inference purposes.The evaluation of our approach confirms the efficacy and efficiency of gravitational wave data inferences and points to a promising direction for similar research.The source code,specifications,and detailed procedures are publicly accessible on GitHub.
基金supported by the Theme-based Research Scheme,Research Grants Council of Hong Kong,China(T45-205/21-N).
文摘To generate dance that temporally and aesthetically matches the music is a challenging problem in three aspects.First,the generated motion should be beats-aligned to the local musical features.Second,the global aesthetic style should be matched between motion and music.And third,the generated motion should be diverse and non-self-repeating.To address these challenges,we propose ReChoreoNet,which re-choreographs high-quality dance motion for a given piece of music.A data-driven learning strategy is proposed to efficiently correlate the temporal connections between music and motion in a progressively learned cross-modality embedding space.The beats-aligned content motion will be subsequently used as autoregressive context and control signal to control a normalizing-flow model,which transfers the style of a prototype motion to the final generated dance.In addition,we present an aesthetically labelled music-dance repertoire(MDR)for both efficient learning of the cross-modality embedding,and understanding of the aesthetic connections between music and motion.We demonstrate that our repertoire-based framework is robustly extensible in both content and style.Both quantitative and qualitative experiments have been carried out to validate the efficiency of our proposed model.
