Nuclear mass based on the multi-task learning neural network method

The global nuclear mass based on the macroscopic-microscopic model was studied by applying a newly designed multi-task learning artificial neural network(MTL-ANN). First, the reported nuclear binding energies of 2095 nuclei(Z ≥ 8, N ≥ 8) released in the latest Atomic Mass Evaluation AME2020 and the deviations between the fitting result of the liquid drop model(LDM)and data from AME2020 for each nucleus were obtained.To compensate for the deviations and investigate the possible ignored physics in the LDM, the MTL-ANN method was introduced in the model. Compared to the single-task learning(STL) method, this new network has a powerful ability to simultaneously learn multi-nuclear properties,such as the binding energies and single neutron and proton separation energies. Moreover, it is highly effective in reducing the risk of overfitting and achieving better predictions. Consequently, good predictions can be obtained using this nuclear mass model for both the training and validation datasets and for the testing dataset. In detail, the global root mean square(RMS) of the binding energy is effectively reduced from approximately 2.4 MeV of LDM to the current 0.2 MeV, and the RMS of Sn, Spcan also reach approximately 0.2 MeV. Moreover, compared to STL, for the training and validation sets, 3-9% improvement can be achieved with the binding energy, and 20-30% improvement for S_(n), S_(p);for the testing sets, the reduction in deviations can even reach 30-40%, which significantly illustrates the advantage of the current MTL.

MMLUP: Multi-Source & Multi-Task Learning for User Profiles in Social Network

With the rapid development of the mobile Internet,users generate massive data in different forms in social network every day,and different characteristics of users are reflected by these social media data.How to integrate multiple heterogeneous information and establish user profiles from multiple perspectives plays an important role in providing personalized services,marketing,and recommendation systems.In this paper,we propose Multi-source&Multi-task Learning for User Profiles in Social Network which integrates multiple social data sources and contains a multi-task learning framework to simultaneously predict various attributes of a user.Firstly,we design their own feature extraction models for multiple heterogeneous data sources.Secondly,we design a shared layer to fuse multiple heterogeneous data sources as general shared representation for multi-task learning.Thirdly,we design each task’s own unique presentation layer for discriminant output of specific-task.Finally,we design a weighted loss function to improve the learning efficiency and prediction accuracy of each task.Our experimental results on more than 5000 Sina Weibo users demonstrate that our approach outperforms state-of-the-art baselines for inferring gender,age and region of social media users.

Spatio-Temporal Cellular Network Traffic Prediction Using Multi-Task Deep Learning for AI-Enabled 6G

Spatio-temporal cellular network traffic prediction at wide-area level plays an important role in resource reconfiguration,traffic scheduling and intrusion detection,thus potentially supporting connected intelligence of the sixth generation of mobile communications technology(6G).However,the existing studies just focus on the spatio-temporal modeling of traffic data of single network service,such as short message,call,or Internet.It is not conducive to accurate prediction of traffic data,characterised by diverse network service,spatio-temporality and supersize volume.To address this issue,a novel multi-task deep learning framework is developed for citywide cellular network traffic prediction.Functionally,this framework mainly consists of a dual modular feature sharing layer and a multi-task learning layer(DMFS-MT).The former aims at mining long-term spatio-temporal dependencies and local spatio-temporal fluctuation trends in data,respectively,via a new combination of convolutional gated recurrent unit(ConvGRU)and 3-dimensional convolutional neural network(3D-CNN).For the latter,each task is performed for predicting service-specific traffic data based on a fully connected network.On the real-world Telecom Italia dataset,simulation results demonstrate the effectiveness of our proposal through prediction performance measure,spatial pattern comparison and statistical distribution verification.

Multi-task Learning of Semantic Segmentation and Height Estimation for Multi-modal Remote Sensing Images

Deep learning based methods have been successfully applied to semantic segmentation of optical remote sensing images.However,as more and more remote sensing data is available,it is a new challenge to comprehensively utilize multi-modal remote sensing data to break through the performance bottleneck of single-modal interpretation.In addition,semantic segmentation and height estimation in remote sensing data are two tasks with strong correlation,but existing methods usually study individual tasks separately,which leads to high computational resource overhead.To this end,we propose a Multi-Task learning framework for Multi-Modal remote sensing images(MM_MT).Specifically,we design a Cross-Modal Feature Fusion(CMFF)method,which aggregates complementary information of different modalities to improve the accuracy of semantic segmentation and height estimation.Besides,a dual-stream multi-task learning method is introduced for Joint Semantic Segmentation and Height Estimation(JSSHE),extracting common features in a shared network to save time and resources,and then learning task-specific features in two task branches.Experimental results on the public multi-modal remote sensing image dataset Potsdam show that compared to training two tasks independently,multi-task learning saves 20%of training time and achieves competitive performance with mIoU of 83.02%for semantic segmentation and accuracy of 95.26%for height estimation.

Robust multi-task distributed estimation based on generalized maximum correntropy criterion

False data injection(FDI) attacks are common in the distributed estimation of multi-task network environments, so an attack detection strategy is designed by combining the generalized maximum correntropy criterion. Based on this, we propose a diffusion least-mean-square algorithm based on the generalized maximum correntropy criterion(GMCC-DLMS)for multi-task networks. The algorithm achieves gratifying estimation results. Even more, compared to the related work,it has better robustness when the number of attacked nodes increases. Moreover, the assumption about the number of attacked nodes is relaxed, which is applicable to multi-task environments. In addition, the performance of the proposed GMCC-DLMS algorithm is analyzed in the mean and mean-square senses. Finally, simulation experiments confirm the performance and effectiveness against FDI attacks of the algorithm.

A General Linguistic Steganalysis Framework Using Multi-Task Learning

Prevailing linguistic steganalysis approaches focus on learning sensitive features to distinguish a particular category of steganographic texts from non-steganographic texts,by performing binary classification.While it remains an unsolved problem and poses a significant threat to the security of cyberspace when various categories of non-steganographic or steganographic texts coexist.In this paper,we propose a general linguistic steganalysis framework named LS-MTL,which introduces the idea of multi-task learning to deal with the classification of various categories of steganographic and non-steganographic texts.LS-MTL captures sensitive linguistic features from multiple related linguistic steganalysis tasks and can concurrently handle diverse tasks with a constructed model.In the proposed framework,convolutional neural networks(CNNs)are utilized as private base models to extract sensitive features for each steganalysis task.Besides,a shared CNN is built to capture potential interaction information and share linguistic features among all tasks.Finally,LS-MTL incorporates the private and shared sensitive features to identify the detected text as steganographic or non-steganographic.Experimental results demonstrate that the proposed framework LS-MTL outperforms the baseline in the multi-category linguistic steganalysis task,while average Acc,Pre,and Rec are increased by 0.5%,1.4%,and 0.4%,respectively.More ablation experimental results show that LS-MTL with the shared module has robust generalization capability and achieves good detection performance even in the case of spare data.

Design space exploration of neural network accelerator based on transfer learning

With the increasing demand of computational power in artificial intelligence(AI)algorithms,dedicated accelerators have become a necessity.However,the complexity of hardware architectures,vast design search space,and complex tasks of accelerators have posed significant challenges.Tra-ditional search methods can become prohibitively slow if the search space continues to be expanded.A design space exploration(DSE)method is proposed based on transfer learning,which reduces the time for repeated training and uses multi-task models for different tasks on the same processor.The proposed method accurately predicts the latency and energy consumption associated with neural net-work accelerator design parameters,enabling faster identification of optimal outcomes compared with traditional methods.And compared with other DSE methods by using multilayer perceptron(MLP),the required training time is shorter.Comparative experiments with other methods demonstrate that the proposed method improves the efficiency of DSE without compromising the accuracy of the re-sults.

Multi-Task Visual Semantic Embedding Network for Image-Text Retrieval

Image-text retrieval aims to capture the semantic correspondence between images and texts,which serves as a foundation and crucial component in multi-modal recommendations,search systems,and online shopping.Existing mainstream methods primarily focus on modeling the association of image-text pairs while neglecting the advantageous impact of multi-task learning on image-text retrieval.To this end,a multi-task visual semantic embedding network(MVSEN)is proposed for image-text retrieval.Specifically,we design two auxiliary tasks,including text-text matching and multi-label classification,for semantic constraints to improve the generalization and robustness of visual semantic embedding from a training perspective.Besides,we present an intra-and inter-modality interaction scheme to learn discriminative visual and textual feature representations by facilitating information flow within and between modalities.Subsequently,we utilize multi-layer graph convolutional networks in a cascading manner to infer the correlation of image-text pairs.Experimental results show that MVSEN outperforms state-of-the-art methods on two publicly available datasets,Flickr30K and MSCOCO,with rSum improvements of 8.2%and 3.0%,respectively.

Target Detection Algorithm in Foggy Scenes Based on Dual Subnets

Under the influence of air humidity,dust,aerosols,etc.,in real scenes,haze presents an uneven state.In this way,the image quality and contrast will decrease.In this case,It is difficult to detect the target in the image by the universal detection network.Thus,a dual subnet based on multi-task collaborative training(DSMCT)is proposed in this paper.Firstly,in the training phase,the Gated Context Aggregation Network(GCANet)is used as the supervisory network of YOLOX to promote the extraction of clean information in foggy scenes.In the test phase,only the YOLOX branch needs to be activated to ensure the detection speed of the model.Secondly,the deformable convolution module is used to improve GCANet to enhance the model’s ability to capture details of non-homogeneous fog.Finally,the Coordinate Attention mechanism is introduced into the Vision Transformer and the backbone network of YOLOX is redesigned.In this way,the feature extraction ability of the network for deep-level information can be enhanced.The experimental results on artificial fog data set FOG_VOC and real fog data set RTTS show that the map value of DSMCT reached 86.56%and 62.39%,respectively,which was 2.27%and 4.41%higher than the current most advanced detection model.The DSMCT network has high practicality and effectiveness for target detection in real foggy scenes.

Modeling and Control of Nonlinear Discrete-time Systems Based on Compound Neural Networks

An adaptive inverse controller for nonliear discrete-time system is proposed in this paper. A compound neural network is constructed to identify the nonlinear system, which includes a linear part to approximate the nonlinear system and a recurrent neural network to minimize the difference between the linear model and the real nonlinear system. Because the current control input is not included in the input vector of recurrent neural network (RNN), the inverse control law can be calculated directly. This scheme can be used in real-time nonlinear single-input single-output (SISO) and multi-input multi-output (MIMO) system control with less computation work. Simulation studies have shown that this scheme is simple and affects good control accuracy and robustness.

Research on Facial Expression Capture Based on Two-Stage Neural Network

To generate realistic three-dimensional animation of virtual character,capturing real facial expression is the primary task.Due to diverse facial expressions and complex background,facial landmarks recognized by existing strategies have the problem of deviations and low accuracy.Therefore,a method for facial expression capture based on two-stage neural network is proposed in this paper which takes advantage of improved multi-task cascaded convolutional networks(MTCNN)and high-resolution network.Firstly,the convolution operation of traditional MTCNN is improved.The face information in the input image is quickly filtered by feature fusion in the first stage and Octave Convolution instead of the original ones is introduced into in the second stage to enhance the feature extraction ability of the network,which further rejects a large number of false candidates.The model outputs more accurate facial candidate windows for better landmarks recognition and locates the faces.Then the images cropped after face detection are input into high-resolution network.Multi-scale feature fusion is realized by parallel connection of multi-resolution streams,and rich high-resolution heatmaps of facial landmarks are obtained.Finally,the changes of facial landmarks recognized are tracked in real-time.The expression parameters are extracted and transmitted to Unity3D engine to drive the virtual character’s face,which can realize facial expression synchronous animation.Extensive experimental results obtained on the WFLW database demonstrate the superiority of the proposed method in terms of accuracy and robustness,especially for diverse expressions and complex background.The method can accurately capture facial expression and generate three-dimensional animation effects,making online entertainment and social interaction more immersive in shared virtual space.

A multi-input and multi-output design on automotive engine management system

Lookup table is widely used in automotive industry for the design of engine control units(ECU).Together with a proportional-integral controller,a feed-forward and feedback control scheme is often adopted for automotive engine management system(EMS).Usually,an ECU has a structure of multi-input and single-output(MISO).Therefore,if there are multiple objectives proposed in EMS,there would be corresponding numbers of ECUs that need to be designed.In this situation,huge efforts and time were spent on calibration.In this work,a multi-input and multi-out(MIMO) approach based on model predictive control(MPC) was presented for the automatic cruise system of automotive engine.The results show that the tracking of engine speed command and the regulation of air/fuel ratio(AFR) can be achieved simultaneously under the new scheme.The mean absolute error(MAE) for engine speed control is 0.037,and the MAE for air fuel ratio is 0.069.

Cooperative MIMO MAC Transmission Using Space Codes in Wireless Sensor Network

Wireless sensor network （WSN） requires robust and efficient communication protocols to minimise delay and save energy. The lifetime of WSN can be maximised by selecting proper medium access control （MAC） scheme depending on the contention level of the network. The throughput of WSN however reduces due to channel fading effects even with the proper design of MAC protocol. Hence this paper proposes a new MAC scheme for enabling packet transmission using cooperative multi-input multi-output （MIMO） utilising space time codes（STC） such as space time block code （STBC）, space time trellis code （STTC） to achieve higher energy savings and lower delay by allowing nodes to transmit and receive information jointly. The performance of the proposed MAC protocol is evaluated in terms of transmission error probability, energy consumption and delay. Simulation results show that the proposed cooperative MIMO MAC protocol provides reliable and efficient transmission by leveraging MIMO diversity gains.

Evaluation of Linear Precoding Schemes for Cooperative Multi-Cell MU MIMO in Future Mobile Communication Systems

In Mobile Communication Systems, inter-cell interference becomes one of the challenges that degrade the system’s performance, especially in the region with massive mobile users. The linear precoding schemes were proposed to mitigate interferences between the base stations (inter-cell). These schemes are categorized into linear and non-linear;this study focused on linear precoding schemes, which are grounded into three types, namely Zero Forcing (ZF), Block Diagonalization (BD), and Signal Leakage Noise Ratio (SLNR). The study included the Cooperative Multi-cell Multi Input Multi Output (MIMO) System, whereby each Base Station serves more than one mobile station and all Base Stations on the system are assisted by each other by shared the Channel State Information (CSI). Based on the Multi-Cell Multiuser MIMO system, each Base Station on the cell is intended to maximize the data transmission rate by its mobile users by increasing the Signal Interference to Noise Ratio after the interference has been mitigated due to the usefully of linear precoding schemes on the transmitter. Moreover, these schemes used different approaches to mitigate interference. This study mainly concentrates on evaluating the performance of these schemes through the channel distribution models such as Ray-leigh and Rician included in the presence of noise errors. The results show that the SLNR scheme outperforms ZF and BD schemes overall scenario. This implied that when the value of SNR increased the performance of SLNR increased by 21.4% and 45.7% for ZF and BD respectively.

Multi-Task Learning for Food Identification and Analysis with Deep Convolutional Neural Networks

In this paper, we proposed a multi-task system that can identify dish types, food ingredients, and cooking methods from food images with deep convolutional neural networks. We built up a dataset of 360 classes of different foods with at least 500 images for each class. To reduce the noises of the data, which was collected from the Internet, outlier images were detected and eliminated through a one-class SVM trained with deep convolutional features. We simultaneously trained a dish identifier, a cooking method recognizer, and a multi-label ingredient detector. They share a few low-level layers in the deep network architecture. The proposed framework shows higher accuracy than traditional method with handcrafted features, and the cooking method recognizer and ingredient detector can be applied to dishes which are not included in the training dataset to provide reference information for users.

结合多任务迁移学习与知识蒸馏的人脸美丽预测研究

目前,人脸美丽预测存在数据样本少、评价指标不明确和人脸外观变化大等问题。多任务迁移学习能有效利用相关任务和源域任务额外的有用信息,知识蒸馏可将教师模型的部分知识蒸馏到学生模型,降低模型复杂性和大小。本文将多任务迁移学习与知识蒸馏相结合,用于人脸美丽预测,以大规模亚洲人脸美丽数据库(Large Scale Asia Facial Beauty Database,LSAFBD)中人脸美丽预测为主任务,以SCUT-FBP5500数据库中性别识别为辅任务。首先,构建多输入多任务的人脸美丽教师模型和学生模型;其次,训练多任务教师模型并计算其软目标;最后,结合多任务教师模型的软目标和学生模型的软、硬目标进行知识蒸馏。实验结果表明,多任务教师模型在人脸美丽预测任务中取得68.23%的准确率,其结构较复杂,参数量达14793K;而多任务学生模型通过知识蒸馏后分类准确率为67.39%,但其结构简单、参数量仅1366K。本方法多任务教师模型分类准确率比其他方法高,多任务学生模型分类准确率虽然略低一点,但其模型更简单、参数量更少,更有利于用更轻量的网络模型进行人脸美丽预测。

Information networks fusion based on multi-task coordination

Information networks provide a powerful representation of entities and the relationships between them.Information networks fusion is a technique for information fusion that jointly reasons about entities,links and relations in the presence of various sources.However,existing methods for information networks fusion tend to rely on a single task which might not get enough evidence for reasoning.In order to solve this issue,in this paper,we present a novel model called MC-INFM(information networks fusion model based on multi-task coordination).Different from traditional models,MC-INFM casts the fusion problem as a probabilistic inference problem,and collectively performs multiple tasks(including entity resolution,link prediction and relation matching)to infer the final result of fusion.First,we define the intra-features and the inter-features respectively and model them as factor graphs,which can provide abundant evidence to infer.Then,we use conditional random field(CRF)to learn the weight of each feature and infer the results of these tasks simultaneously by performing the maximum probabilistic inference.Experiments demonstrate the effectiveness of our proposed model.

Lane-changing trajectory prediction based on multi-task learning

As a complex driving behaviour,lane-changing(LC)behaviour has a great influence on traffic flow.Improper lane-changing behaviour often leads to traffic accidents.Numerous studies are currently being conducted to predict lane-change trajectories to minimize dangers.However,most of their models focus on how to optimize input variables without considering the interaction between output variables.This study proposes an LC trajectory prediction model based on a multi-task deep learning framework to improve driving safety.Concretely,in this work,the coupling effect of lateral and longitudinal movement is considered in the L.C process.Trajectory changes in two directions will be modelled separately,and the information interaction is completed under the multi-task learing framework.In addition,the trajectory fragents are clustered by the driving features,and trajectory type recognition is added to the trajectory prediction framework as an auxiliary task.Finally,the prediction process of lateral and longitudinal trajectory and LC style is completed by long short-term memory(LSTM).The model training and testing are conducted with the data collected by the driving simulator,and the proposed method expresses better performance in LC trjectory prediction compared with several traditional models.The results of this study can enhance the trajectory prediction accuracy of advanced driving assistance systems(ADASs)and reduce the traffic accidents caused by lane changes.

Multi-user rate and power analysis in a cognitive radio network with massive multi-input multi-output

This paper discusses transmission performance and power allocation strategies in an underlay cognitive radio （CR） network that contains relay and massive multi-input multi-output （MIMO）. The downlink transmission performance of a relay-aided massive MIMO network without CR is derived. By using the power distribution criteria, the kth user＇s asymptotic signal to interference and noise ratio （SINR） is independent of fast fading. When the ratio between the base station （BS） antennas and the relay antennas becomes large enough, the transmission performance of the whole system is independent of BS-to-relay channel parameters and relates only to the relay-to-users stage. Then cognitive transmission performances of primary users （PUs） and secondary users （SUs） in an underlay CR network with massive MIMO are derived under perfect and imperfect channel state information （CSI）, including the end-to-end SINR and achievable sum rate. When the numbers of primary base station （PBS） antennas, secondary base station （SBS） antennas, and relay antennas become infinite, the asymptotic SINR of the kth PU and SU is independent of fast fading. The interference between the primary network and secondary network can be canceled asymptotically.Transmission performance does not include the interference temperature. The secondary network can use its peak power to transmit signals without causing any interference to the primary network. Interestingly, when the antenna ratio becomes large enough, the asymptotic sum rate equals half of the rate of a single-hop single-antenna K-user system without fast fading. Next, the PUs＇ utility function is defined. The optimal relay power is derived to maximize the utility function. The numerical results verify our analysis. The relationships between the transmission rate and the antenna nunber, relay power, and antenna ratio are simulated. We show that the massive MIMO with linear pre-coding can mitigate asymptotically the interference in a multi-user underlay CR network. The primary and secondary networks can operate independently.

Power flow forecasts at transmission grid nodes using Graph Neural Networks

The increasing share of renewable energy in the electricity grid and progressing changes in power consumption have led to fluctuating,and weather-dependent power flows.To ensure grid stability,grid operators rely on power forecasts which are crucial for grid calculations and planning.In this paper,a Multi-Task Learning approach is combined with a Graph Neural Network(GNN)to predict vertical power flows at transformers connecting high and extra-high voltage levels.The proposed method accounts for local differences in power flow characteristics by using an Embedding Multi-Task Learning approach.The use of a Bayesian embedding to capture the latent node characteristics allows to share the weights across all transformers in the subsequent node-invariant GNN while still allowing the individual behavioral patterns of the transformers to be distinguished.At the same time,dependencies between transformers are considered by the GNN architecture which can learn relationships between different transformers and thus take into account that power flows in an electricity network are not independent from each other.The effectiveness of the proposed method is demonstrated through evaluation on two real-world data sets provided by two of four German Transmission System Operators,comprising large portions of the operated German transmission grid.The results show that the proposed Multi-Task Graph Neural Network is a suitable representation learner for electricity networks with a clear advantage provided by the preceding embedding layer.It is able to capture interconnections between correlated transformers and indeed improves the performance in power flow prediction compared to standard Neural Networks.A sign test shows that the proposed model reduces the test RMSE on both data sets compared to the benchmark models significantly.