Single-task Networks Research Articles

Learning quantum properties from short-range correlations using multi-task networks

Characterizing multipartite quantum systems is crucial for quantum computing and many-body physics. The problem, however, becomes challenging when the system size is large and the properties of interest involve correlations among a large number of particles. Here we introduce a neural network model that can predict various quantum properties of many-body quantum states with constant correlation length, using only measurement data from a small number of neighboring sites. The model is based on the technique of multi-task learning, which we show to offer several advantages over traditional single-task approaches. Through numerical experiments, we show that multi-task learning can be applied to sufficiently regular states to predict global properties, like string order parameters, from the observation of short-range correlations, and to distinguish between quantum phases that cannot be distinguished by single-task networks. Remarkably, our model appears to be able to transfer information learnt from lower dimensional quantum systems to higher dimensional ones, and to make accurate predictions for Hamiltonians that were not seen in the training.

SDC-Net++: End-to-End Crash Detection and Action Control for Self-Driving Car Deep-IoT-Based System.

Few prior works study self-driving cars by deep learning with IoT collaboration. SDC-Net, which is an end-to-end multitask self-driving car camera cocoon IoT-based system, is one of the research areas that tackles this direction. However, by design, SDC-Net is not able to identify the accident locations; it only classifies whether a scene is a crash scene or not. In this work, we introduce an enhanced design for the SDC-Net system by (1) replacing the classification network with a detection one, (2) adapting our benchmark dataset labels built on the CARLA simulator to include the vehicles' bounding boxes while keeping the same training, validation, and testing samples, and (3) modifying the shared information via IoT to include the accident location. We keep the same path planning and automatic emergency braking network, the digital automation platform, and the input representations to formulate the comparative study. The SDC-Net++ system is proposed to (1) output the relevant control actions, especially in case of accidents: accelerate, decelerate, maneuver, and brake, and (2) share the most critical information to the connected vehicles via IoT, especially the accident locations. A comparative study is also conducted between SDC-Net and SDC-Net++ with the same input representations: front camera only, panorama and bird's eye views, and with single-task networks, crash avoidance only, and multitask networks. The multitask network with a BEV input representation outperforms the nearest representation in precision, recall, f1-score, and accuracy by more than 15.134%, 12.046%, 13.593%, and 5%, respectively. The SDC-Net++ multitask network with BEV outperforms SDC-Net multitask with BEV in precision, recall, f1-score, accuracy, and average MSE by more than 2.201%, 2.8%, 2.505%, 2%, and 18.677%, respectively.

Fast Aerodynamic Prediction of Airfoil with Trailing Edge Flap Based on Multi-Task Deep Learning

Conventional methods for solving Navier–Stokes (NS) equations to analyze flow fields and aerodynamic forces of airfoils with trailing edge flaps (TEFs) are known for their significant time cost. This study presents a Multi-Task Swin Transformer (MT-Swin-T) deep learning framework tailored for swift prediction of velocity fields and aerodynamic coefficients of TEF-equipped airfoils. The proposed model combines a Swin Transformer (Swin-T) for flow field prediction with a multi-layer perceptron (MLP) dedicated to lift coefficient prediction. Both networks undergo gradient updates through the shared encoder component of the Swin Transformer. Such a trained network model for computational fluid dynamics simulations is both effective and robust, significantly improving the efficiency of complex aerodynamic shape design optimization and flow control. The study further investigates the impact of integrating multi-task learning loss functions, skip connections, and the network’s structural design on prediction accuracy. Additionally, the effectiveness of deep learning in improving the aerodynamic simulation efficiency of airfoils with TEF is examined. Results demonstrate that the multi-task deep learning approach provides accurate predictions for TEF airfoil flow fields and lift coefficients. The strategic combination of these tasks during network training, along with the optimal selection of loss functions, significantly enhances prediction accuracy compared with the single-task network. In a specific case study, the MT-Swin-T model demonstrated a prediction time that was 1/7214 of the time necessitated by CFD simulation.

Comprehensive visual information acquisition for tomato picking robot based on multitask convolutional neural network

The tomato picking robot's vision system faces two difficult tasks: precise tomato pose acquisition and stem location. Tomato pose and stem location can help determine the end effector pose and achieve collision-free picking. To realise efficient crop picking, the tasks of target location, pose detection, and obstacle semantic segmentation should be completed in one model to obtain comprehensive visual information. Therefore, the multitask convolutional neural network YOLO-MCNN is proposed, a new method to complete the above tasks in one model. By fusing multi-scale features and determining the optimal locations for the semantic segmentation branch, four strategies are proposed for enhancing the segmentation ability. The experiment results show that fusing the semantic segmentation branch with the second layer of shallow feature maps and placing the branch after the 17th layer can result in the best segmentation performance. Fusing shallow feature maps improves small target detection while merging multi-scale feature maps enhances semantic segmentation performance. Moreover, ablation experiments are conducted to understand the influence between multitask convolutional and single task networks. It proves that running multiple tasks on the same backbone network does not affect their performance. The YOLO-MCNN's target detection performance F1 is 87.8%, the semantic segmentation performance mIoU is 74.8%, the keypoint detection performance dlmk is 6.95 pixels, the network size is 15.2 MB, and the inference speed is 19.9ms. Compared with other target detection and semantic segmentation networks, it shows that the comprehensive performance of the YOLO-MCNN is the best. The method provides theoretical foundation for constructing multitask convolutional neural networks.

A multi-task network for speaker and command recognition in industrial environments

In industrial environments, it is crucial to establish a strong collaboration between humans and robots to enhance productivity. However, the nature of the work demands that workers have the authority to provide specific instructions to the robots. The scientific community has extensively investigated these dual requirements, aiming to develop advanced systems capable of recognizing voice commands and implementing speaker authentication. Nevertheless, in the industrial context, these tasks should be executed simultaneously on low-cost and low-power embedded devices that can be mounted on board the robotic platform. To overcome this challenge, we propose a multi-task network for Speech-Command Recognition and Speaker Identification. Additionally, we employ the GradNorm adaptive algorithm to address the issue of task imbalance. To evaluate the proposed system, we introduce a new dataset, MIVIA-ISC, consisting of 20,857 samples uttered by 562 speakers for 31 distinct commands. Our approach significantly reduces the network size by 47% and its execution time by 48% compared to the commonly used methodology, which employs one network for each task. Furthermore, our approach demonstrates a significant improvement in the accuracy of the Speaker Identification task, achieving an 11% increase compared to the corresponding single-task network. Importantly, this enhancement is achieved without compromising the accuracy of the Speech-Command Recognition task, which experiences only a minimal 3% decrease in performance.

LidarMultiNet: Towards a Unified Multi-Task Network for LiDAR Perception

LiDAR-based 3D object detection, semantic segmentation, and panoptic segmentation are usually implemented in specialized networks with distinctive architectures that are difficult to adapt to each other. This paper presents LidarMultiNet, a LiDAR-based multi-task network that unifies these three major LiDAR perception tasks. Among its many benefits, a multi-task network can reduce the overall cost by sharing weights and computation among multiple tasks. However, it typically underperforms compared to independently combined single-task models. The proposed LidarMultiNet aims to bridge the performance gap between the multi-task network and multiple single-task networks. At the core of LidarMultiNet is a strong 3D voxel-based encoder-decoder architecture with a Global Context Pooling (GCP) module extracting global contextual features from a LiDAR frame. Task-specific heads are added on top of the network to perform the three LiDAR perception tasks. More tasks can be implemented simply by adding new task-specific heads while introducing little additional cost. A second stage is also proposed to refine the first-stage segmentation and generate accurate panoptic segmentation results. LidarMultiNet is extensively tested on both Waymo Open Dataset and nuScenes dataset, demonstrating for the first time that major LiDAR perception tasks can be unified in a single strong network that is trained end-to-end and achieves state-of-the-art performance. Notably, LidarMultiNet reaches the official 1 place in the Waymo Open Dataset 3D semantic segmentation challenge 2022 with the highest mIoU and the best accuracy for most of the 22 classes on the test set, using only LiDAR points as input. It also sets the new state-of-the-art for a single model on the Waymo 3D object detection benchmark and three nuScenes benchmarks.

Multitasking neural network to jointly map discrete fracture structures and matrix transmissivity by inverting hydraulic data acquired in 2D fractured aquifers. XNET-fracture

In this paper, we propose a new approach to simultaneously map the fracture/conduit network and the equivalent transmissivity of the rock matrix from hydraulic head measurements acquired during pumping/injection tests in fractured aquifers. The algorithm relies on the use of a multitask neural network to directly approximate the joint inversion operator that links hydraulic head data to aquifer hydraulic properties. In which, the hydraulic head responses are used as input data while the output consists of the fracture structure and matrix transmissivity field. The multitasking is formed by fusing two single-task neural networks, both based on convolutional encoder-decoder architectures, one of which processes the fracture map inversion and the other the transmissivity map. Training neural network then relies on a large database, which consists of thousands of synthetic aquifers characterized by the presence of the fracture/conduit network and a heterogeneous transmissivity field attributed to the rock matrix, both randomly generated. In these aquifer models, the groundwater flow equation is solved in a discrete-continuum concept to numerically simulate the hydraulic head responses associated with sequential pumping/injection tests.Multitasking succeeds in reconstructing the fracture architecture and matrix transmissivity field with higher accuracy in comparison to conventional single-task networks. The difference is due to the transfer mechanism involved in the multi-task network, where the information shared between the tasks effectively enhances the accuracy of both reconstructions. However, as with other inversion techniques, the result accuracy depends on the quality and quantity of the hydraulic head data used in the inversion, as well as the prior information involved in featuring the training models.

Multitask Learning for Activity Detection in Neovascular Age-Related Macular Degeneration.

The purpose of this study was to provide a comparison of performance and explainability of a multitask convolutional deep neuronal network to single-task networks for activity detection in neovascular age-related macular degeneration (nAMD). From 70 patients (46 women and 24 men) who attended the University Eye Hospital Tübingen, 3762 optical coherence tomography B-scans (right eye = 2011 and left eye = 1751) were acquired with Heidelberg Spectralis, Heidelberg, Germany. B-scans were graded by a retina specialist and an ophthalmology resident, and then used to develop a multitask deep learning model to predict disease activity in neovascular age-related macular degeneration along with the presence of sub- and intraretinal fluid. We used performance metrics for comparison to single-task networks and visualized the deep neural network (DNN)-based decision with t-distributed stochastic neighbor embedding and clinically validated saliency mapping techniques. The multitask model surpassed single-task networks in accuracy for activity detection (94.2% vs. 91.2%). The area under the curve of the receiver operating curve was 0.984 for the multitask model versus 0.974 for the single-task model. Furthermore, compared to single-task networks, visualizations via t-distributed stochastic neighbor embedding and saliency maps highlighted that multitask networks' decisions for activity detection in neovascular age-related macular degeneration were highly consistent with the presence of both sub- and intraretinal fluid. Multitask learning increases the performance of neuronal networks for predicting disease activity, while providing clinicians with an easily accessible decision control, which resembles human reasoning. By improving nAMD activity detection performance and transparency of automated decisions, multitask DNNs can support the translation of machine learning research into clinical decision support systems for nAMD activity detection.

Multi-task learning-based histologic subtype classification of non-small cell lung cancer.

In clinical applications, accurate histologic subtype classification of lung cancer is important for determining appropriate treatment plans. The purpose of this paper is to evaluate the role of multi-task learning in the classification of adenocarcinoma and squamous cell carcinoma. In this paper, we propose a novel multi-task learning model for histologic subtype classification of non-small cell lung cancer based on computed tomography (CT) images. The model consists of a histologic subtype classification branch and a staging branch, which share a part of the feature extraction layers and are simultaneously trained. By optimizing on the two tasks simultaneously, our model could achieve high accuracy in histologic subtype classification of non-small cell lung cancer without relying on physician's precise labeling of tumor areas. In this study, 402 cases from The Cancer Imaging Archive (TCIA) were used in total, and they were split into training set (n = 258), internal test set (n = 66) and external test set (n = 78). Compared with the radiomics method and single-task networks, our multi-task model could reach an AUC of 0.843 and 0.732 on internal and external test set, respectively. In addition, multi-task network can achieve higher accuracy and specificity than single-task network. Compared with the radiomics methods and single-task networks, our multi-task learning model could improve the accuracy of histologic subtype classification of non-small cell lung cancer by sharing network layers, which no longer relies on the physician's precise labeling of lesion regions and could further reduce the manual workload of physicians.

Multi-Tasking Memcapacitive Networks

Recent studies have shown that networks of memcapacitive devices provide an ideal computing platform of low power consumption for reservoir computing systems. Random, crossbar, or small-world power-law (SWPL) structures are common topologies for reservoir substrates to compute single tasks. However, neurological studies have shown that the interconnections of cortical brain regions associated with different functions form a rich-club structure. This structure allows human brains to perform multiple activities simultaneously. So far, memcapacitive reservoirs can perform only single tasks. Here, we propose, for the first time, cluster networks functioning as memcapacitive reservoirs to perform multiple tasks simultaneously. Our results illustrate that cluster networks surpassed crossbar and SWPL networks by factors of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$4.1\times, 5.2\times $ </tex-math></inline-formula> , and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1.7\times $ </tex-math></inline-formula> on three tasks: Isolated Spoken Digits, MNIST, and CIFAR-10. Compared to single-task networks in our previous and published results, multitasking cluster networks could accomplish similar accuracies of 86%, 94.4%, and 27.9% for MNIST, Isolated Spoken Digits, and CIFAR-10. Our extended simulations reveal that both the input signal amplitudes and the inter-cluster connections contribute to the accuracy of cluster networks. Selecting optimal values for signal amplitudes and inter-cluster links is key to obtaining high classification accuracy and low power consumption. Our results illustrate the promise of memcapacitive brain-inspired cluster networks and their capability to solve multiple tasks simultaneously. Such novel computing architectures have the potential to make edge applications more efficient and allow systems that cannot be reconfigured to solve multiple tasks.

A Multitask Deep Learning Model for Parsing Bridge Elements and Segmenting Defect in Bridge Inspection Images

The vast network of bridges in the United States has a high requirement for maintenance and rehabilitation. The massive cost of manual visual inspection to assess bridge conditions is a burden to some extent. Advanced robots have been used to automate inspection data collection. Automating the segmentations of multiclass elements and surface defects on the elements in the large volume of inspection image data would facilitate an efficient and effective assessment of bridge conditions. Training separate single-task networks for element parsing (i.e., semantic segmentation of multiclass elements) and defect segmentation fails to incorporate the close connection between these two tasks. Both recognizable structural elements and apparent surface defects are present in the inspection images. This paper is motivated to develop a multitask deep learning model that fully uses such interdependence between bridge elements and defects to boost the model’s task performance and generalization. Furthermore, the study investigated the effectiveness of the proposed model designs for improving task performance, including feature decomposition, cross-talk sharing, and multi-objective loss function. A data set with pixel-level labels of bridge elements and corrosion was developed for model training and testing. Quantitative and qualitative results from evaluating the developed multitask deep model demonstrate its advantages over the single-task-based model not only in performance (2.59% higher mIoU on bridge parsing and 1.65% on corrosion segmentation) but also in computational time and implementation capability.

Physics-Guided Multitask Learning for Estimating Power Generation and CO2 Emissions From Satellite Imagery

Fossil fuel combustion produces large quantities of carbon dioxide (CO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ), a major Greenhouse Gas (GHG), which is one of the main drivers of Climate Change. A quantitative assessment of GHG emissions is fundamental to predicting climate change effects, enforcing emission regulations, and monitoring pollution trading schemes. Unfortunately, the reporting of GHG emissions is only required in some countries, resulting in insufficient global coverage. At the same time, the transition from fossil fuels to zero carbon in order to limit climate change is at the heart of several ecological movements, hence the need for quantifying energy production, as well. In this work, we propose an end-to-end method to estimate power generation rates for fossil fuel power plants from satellite images, based on which we approximate GHG (CO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ) emission rates. We present a physics-guided multitask deep learning approach able to simultaneously predict from a single satellite image of a power plant: (i) the pixel-area covered by plumes, (ii) the type of fired fuel, and (iii) the power generation rate. To ensure physically realistic predictions from our model we account for environmental conditions and empirical physical constraints. We then convert the predicted power generation rate into estimates for the rate at which CO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> is being emitted, using a fuel-dependent conversion factor. Experimental results show that our multitask learning approach improves the power generation estimation MAE by 23% compared to a single-task network trained on the same dataset. Code <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> and dataset <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> utilized in this work are publicly available.

An Efficient End-to-End Multitask Network Architecture for Defect Inspection.

Recently, computer vision-based methods have been successfully applied in many industrial fields. Nevertheless, automated detection of steel surface defects remains a challenge due to the complexity of surface defects. To solve this problem, many models have been proposed, but these models are not good enough to detect all defects. After analyzing the previous research, we believe that the single-task network cannot fully meet the actual detection needs owing to its own characteristics. To address this problem, an end-to-end multi-task network has been proposed. It consists of one encoder and two decoders. The encoder is used for feature extraction, and the two decoders are used for object detection and semantic segmentation, respectively. In an effort to deal with the challenge of changing defect scales, we propose the Depthwise Separable Atrous Spatial Pyramid Pooling module. This module can obtain dense multi-scale features at a very low computational cost. After that, Residually Connected Depthwise Separable Atrous Convolutional Blocks are used to extract spatial information under low computation for better segmentation prediction. Furthermore, we investigate the impact of training strategies on network performance. The performance of the network can be optimized by adopting the strategy of training the segmentation task first and using the deep supervision training method. At length, the advantages of object detection and semantic segmentation are tactfully combined. Our model achieves mIOU 79.37% and mAP@0.5 78.38% on the NEU dataset. Comparative experiments demonstrate that this method has apparent advantages over other models. Meanwhile, the speed of detection amount to 85.6 FPS on a single GPU, which is acceptable in the practical detection process.

A Prior Semantic Network for Large-Scale Landcover Change of Landsat Imagery

Landcover change can reflect changes in the natural environment and the impact of human activities. Remotely sensed big data with large-scale and multi-temporal key characteristics provide the data support for landcover change information extraction. The development of deep learning provides technical method support for information extraction from remotely sensed big data. However, the current mainstream deep learning change detection methods only establish the changing relationship between two phases of images. They cannot directly extract the ground object categories before and after the change. It is easily affected by pseudo-changes caused by the color difference of multi-temporal images, resulting in many false detections. In this paper, we propose a prior semantic network and a difference enhancement block module to establish prior guidance and constraints on changing features to solve the pseudo-change problem. We propose a semantic-change integrated single-task network, which can simultaneously extract multi-temporal landcover classification and landcover change. On the self-made, large-scale multi-temporal Landsat dataset, we have performed multi-temporal landcover change information extraction, reaching an overall accuracy of 83.1% and achieving state-of-the-art performance. Finally, we thoroughly analyzed the landcover change results in the study area from 2005 to 2020.

A synchronous detection-segmentation method for oversized gangue on a coal preparation plant based on multi-task learning

Online identification and sorting for coal and gangue has always been a hot issue in the field of coal processing intelligence. Existing research has focused on materials with particle sizes below 300 mm, and its front-end algorithms are dedicated to achieving image classification or object detection. The lack of detailed shape information of materials in these methods enables them to be not suitable for sorting oversized gangue. In this work, we proposed a synchronous detection-segmentation method for oversized gangue, which was implemented as a joint network based on the multi-task learning theory. The loss function of joint network and the feature interaction channels between the shared encoding module and the parallel decoding branches were designed to efficiently achieve object detection and semantic segmentation for oversized gangue. The proposed method has been evaluated in a comprehensive manner using huge amounts of coal-gangue images taken in an actual production process. The superiority of our joint network based on multi-task learning was verified by comparing several experimental results of them with the classical single-task networks. The issue of convergence synchronization between the multi-task branches was investigated to further optimize the segmentation results. Meanwhile, the effectiveness of the proposed method in improving the sorting capability of the manipulator was explained through a qualitative analysis for a case of sorting oversized gangue.

Multitask Learning Based on Improved Uncertainty Weighted Loss for Multi-Parameter Meteorological Data Prediction

With the exponential growth in the amount of available data, traditional meteorological data processing algorithms have become overwhelmed. The application of artificial intelligence in simultaneous prediction of multi-parameter meteorological data has attracted much attention. However, existing single-task network models are generally limited by the data correlation dependence problem. In this paper, we use a priori knowledge for network design and propose a multitask model based on an asymmetric sharing mechanism, which effectively solves the correlation dependence problem in multi-parameter meteorological data prediction and achieves simultaneous prediction of multiple meteorological parameters with complex correlations for the first time. The performance of the multitask model depends largely on the relative weights among the task losses, and manually adjusting these weights is a difficult and expensive process, which makes it difficult for multitask learning to achieve the expected results in practice. In this paper, we propose an improved multitask loss processing method based on the assumptions of homoscedasticity uncertainty and the Laplace loss distribution and validate it using the German Jena dataset. The results show that the method can automatically balance the losses of each subtask and has better performance and robustness.

Multi-Task Learning-Based Channel Estimation for RIS Assisted Multi-User Communication Systems

In this letter, we propose a multi-task learning (MTL)-based joint channel estimation scheme for reconfigurable intelligent surface (RIS) assisted millimeter-wave communication system, where the direct channel (DC) and cascaded channel (CC) are estimated at the same coherence time by learning the feature of shared pilots. Since the dimension of CC is much larger than the DC, we design a learnable joint loss function based on homoscedastic task uncertainty to balance the training of two subtasks. Meanwhile, the residual shrinkage blocks are introduced into the multi-task network architecture to release the noise effect. Simulation results show that the estimation accuracy of MTL with less pilot overhead outperforms conventional channel estimation scheme, and significantly reduces training overhead compared with the single-task network.

A computer-aid multi-task light-weight network for macroscopic feces diagnosis.

The abnormal traits and colors of feces typically indicate that the patients are probably suffering from tumor or digestive-system diseases. Thus a fast, accurate and automatic health diagnosis system based on feces is urgently necessary for improving the examination speed and reducing the infection risk. The rarity of the pathological images would deteriorate the accuracy performance of the trained models. In order to alleviate this problem, we employ augmentation and over-sampling to expand the samples of the classes that have few samples in the training batch. In order to achieve an impressive recognition performance and leverage the latent correlation between the traits and colors of feces pathological samples, a multi-task network is developed to recognize colors and traits of the macroscopic feces images. The parameter number of a single multi-task network is generally much smaller than the total parameter number of multiple single-task networks, so the storage cost is reduced. The loss function of the multi-task network is the weighted sum of the losses of the two tasks. In this paper, the weights of the tasks are determined according to their difficulty levels that are measured by the fitted linear functions. The sufficient experiments confirm that the proposed method can yield higher accuracies, and the efficiency is also improved.

MBMT-Net: A Multi-Task Learning Based Convolutional Neural Network Architecture for Dense Prediction Tasks

Recently proposed improvements in the field of Computer Vision refer to enhancing the feature processing capabilities of Single-Task Convolutional Neural Networks. A typical Single-Task network consists of a backbone and a head, where the feature extractor is usually optimised using the gradient provided by the head. Inevitably, the backbone specialises for the given task. This sort of approach does not scale well for learning multiple tasks at once while having the same input. As a response, there is an increasing interest in Multi-Task formulations. Since most Multi-Task architectures employ a single shared backbone, when gradients from different tasks are propagated back to it, it can result in its oversaturation. Thus, this problem may be solved using Multi-Backbone feature extractors. Hence, as a strategy proposed to compensate for these shortcomings, we introduce <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MBMT-Net</i> , a Multi-Backbone-Multi-Task-Network architecture based on a development strategy that infuses backbones with more diverse and specialised processing capabilities. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MBMT-Net</i> consists of parallel pre-trained backbones whose outputs are concatenated and offered to the Multi-Task heads that shall benefit from richer and more diverse features with decreased number of network parameters when compared to traditional Multi-Task architectures. Our strategy is architecture independent, and it can be applied to different types of backbones and parsing heads, which greatly extends the domain of configurable features, finally enhancing existing Single- and Multi-Task model building strategies and outperforming them when using the Multi-Backbone design. As a result, while having a deficit of 12.16M parameters, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MBMT-Net</i> reaches state-of-the-art performances and surpasses the previously best semantic segmentation Multi-Task model in terms of Mean Intersection over Union when evaluated on NYUv2 data set.

Robust direction-of-arrival estimation for a target speaker based on multi-task U-net based direct-path dominance test

A dedicated multi-task network for direct-path dominance test has been proposed recently, based on which the robustness of the direction-of-arrival estimation for a target speaker tends to be significantly improved. In this letter, the network is further refined to avoid the original two-stage processing. Moreover, the benefit and generalization of the multi-task network are confirmed by comparison with a single-task network on a significantly larger database. The efficacy of the proposed method in high noisy environments is also validated in real application environments.