Multitasking Manner Research Articles

CAM based fine-grained spatial feature supervision for hierarchical yoga pose classification using multi-stage transfer learning

In this paper, we propose a technique for hierarchical yoga pose classification (YPC) in a multi-stage multi-tasking framework. We propose a three-stage transfer learning based end-to-end training methodology. Novelty lies in (a) proposed supervised contrastive combined loss function for stage-1 training, (b) proposed Encoder–Decoder network architecture with attention mechanism for stage-3 training, (c) proposed spatial context aware multi-tasking combined loss function for stage-3 training. Firstly, for stage-1 training, we propose the usage of linear combination of three loss functions: cross-entropy, self-supervised contrastive loss and supervised contrastive loss in a multi-tasking manner. We introduce radial and cosine margin into the formulation of self-supervised and supervised contrastive loss to pull feature embeddings of same class closer together compared to feature embeddings of different classes. Weights learned over stage-1 training are subsequently fine-tuned over cross-entropy multi-tasking loss in stage-2. These stage-2 weights are transfer learned and are further fine-tuned in stage-3 training. For stage-3 training, we propose the usage of spatial context aware multi-tasking combined loss function. This loss function leverages on the fine-grained spatial features obtained from HiResCAM. These are processed in parallel with features obtained from XGrad-CAM (sensitivity and conservation axioms satisfying features) to further supervise the learning of hierarchical yoga pose classifier. We exemplify our methodology on the publicly available Yoga-82 large-scale dataset. We report peak Top-1 YPC accuracy of 95.89% over 6 pose classes (Yoga-6), 93.85% over 20 pose classes (Yoga-20) and 90.0% over 82 pose classes (Yoga-82). Our proposed method achieves 6.1% improvement over Top-1 classification accuracy in Yoga-6 hierarchy, 9.3% improvement in Yoga-20 hierarchy and 10.9% improvement in Yoga-82 hierarchy in comparison with state-of-the-art (SOTA) methodology. We achieve the current best Top-1 classification accuracies in all the three YPC hierarchies.

Intent Classification and Slot Filling Model for In-Vehicle Services in Korean

Since understanding a user’s request has become a critical task for the artificial intelligence speakers, capturing intents and finding correct slots along with corresponding slot value is significant. Despite various studies concentrating on a real-life situation, dialogue system that is adaptive to in-vehicle services are limited. Moreover, the Korean dialogue system specialized in an vehicle domain rarely exists. We propose a dialogue system that captures proper intent and activated slots for Korean in-vehicle services in a multi-tasking manner. We implement our model with a pre-trained language model, and it includes an intent classifier, slot classifier, slot value predictor, and value-refiner. We conduct the experiments on the Korean in-vehicle services dataset and show 90.74% of joint goal accuracy. Also, we analyze the efficacy of each component of our model and inspect the prediction results with qualitative analysis.

A review of artificial intelligence applications in anterior segment ocular diseases

Background: Artificial intelligence (AI) has great potential for interpreting and analyzing images and processing large amounts of data. There is a growing interest in investigating the applications of AI in anterior segment ocular diseases. This narrative review aims to assess the use of different AI-based algorithms for diagnosing and managing anterior segment entities.
 Methods: We reviewed the applications of different AI-based algorithms in the diagnosis and management of anterior segment entities, including keratoconus, corneal dystrophy, corneal grafts, corneal transplantation, refractive surgery, pterygium, infectious keratitis, cataracts, and disorders of the corneal nerves, conjunctiva, tear film, anterior chamber angle, and iris. The English-language databases PubMed/MEDLINE, Scopus, and Google Scholar were searched using the following keywords: artificial intelligence, deep learning, machine learning, neural network, anterior eye segment diseases, corneal disease, keratoconus, dry eye, refractive surgery, pterygium, infectious keratitis, anterior chamber, and cataract. Relevant articles were compared based on the use of AI models in the diagnosis and treatment of anterior segment diseases. Furthermore, we prepared a summary of the diagnostic performance of the AI-based methods for anterior segment ocular entities.
 Results: Various AI methods based on deep and machine learning can analyze data obtained from corneal imaging modalities with acceptable diagnostic performance. Currently, complicated and time-consuming manual methods are available for diagnosing and treating eye diseases. However, AI methods could save time and prevent vision impairment in eyes with anterior segment diseases. Because many anterior segment diseases can cause irreversible complications and even vision loss, sufficient confidence in the results obtained from the designed model is crucial for decision-making by experts.
 Conclusions: AI-based models could be used as surrogates for analyzing manual data with improveddiagnostic performance. These methods could be reliable tools for diagnosing and managing anterior segmentocular diseases in the near future in remote areas. It is expected that future studies can design algorithms thatuse less data in a multitasking manner for the detection and management of anterior segment diseases.

Self-adaptive multifactorial evolutionary algorithm for multitasking production optimization

The economic and efficient development of petroleum resources can be realized by dynamically adjusting the reservoir development scheme for the sake of higher recovery efficiency with low development cost. As an important part of the closed-loop reservoir management (CLRM), production optimization has gained increasing research interests. A number of sophisticated production optimization techniques have been proposed in the recent years. It is noteworthy that almost all of these existing methods optimize multiple distinct problems independently and neglect the latent synergies among them. However, seldom real-world problems exist in isolation. A number of studies in the community of computational intelligence demonstrated that the latent similarities among multiple distinct optimization tasks can be utilized to achieve knowledge transfer and thus significantly improve the overall optimization performance. With this in mind, a novel multitasking optimization method named multifactorial evolutionary algorithm (MFEA) is introduced to solve production optimization problems in this study. Different production optimization problems are seen as multiple distinct tasks in a multi-tasking environment thus the given problems can be solved in a multi-tasking manner. To the best of our knowledge, this is the first inspirational application of knowledge transfer to reservoir production optimization. Unfortunately, without any prior knowledge about inter-task similarity, a prespecified transfer intensity parameter adopted by the MFEA can potentially lead to performance slowdowns on some unrelated problems. This phenomenon is also known as the negative transfer. To address this issue, a novel self-adaptive multifactorial evolutionary algorithm (SA-MFEA) is proposed in this study. The transfer intensity parameter is estimated online based on a novel inter-task similarity measurement mechanism. The positive transfer between the problems with high degree of relatedness can be greatly boosted by estimating a higher transfer intensity, while the negative transfer between the distinct problems with low similarity can be effectively curbed by adopting a low value of transfer intensity. At last, the efficacy of the proposed method is experimentally verified on a synthetic multitasking problem with three distinct benchmark functions and two multitasking production optimization problems with distinct component reservoir models.

Body-Part-Aware and Multitask-Aware Single-Image-Based Action Recognition

Action recognition is an application that, ideally, requires real-time results. We focus on single-image-based action recognition instead of video-based because of improved speed and lower cost of computation. However, a single image contains limited information, which makes single-image-based action recognition a difficult problem. To get an accurate representation of action classes, we propose three feature-stream-based shallow sub-networks (image-based, attention-image-based, and part-image-based feature networks) on the deep pose estimation network in a multitasking manner. Moreover, we design the multitask-aware loss function, so that the proposed method can be adaptively trained with heterogeneous datasets where only human pose annotations or action labels are included (instead of both pose and action information), which makes it easier to apply the proposed approach to new data on behavioral analysis on intelligent systems. In our extensive experiments, we showed that these streams represent complementary information and, hence, the fused representation is robust in distinguishing diverse fine-grained action classes. Unlike other methods, the human pose information was trained using heterogeneous datasets in a multitasking manner; nevertheless, it achieved 91.91% mean average precision on the Stanford 40 Actions Dataset. Moreover, we demonstrated the proposed method can be flexibly applied to multi-labels action recognition problem on the V-COCO Dataset.

Embedded Control System with SCADA Modbus Network for Monitoring Heat Treating Furnace

In order to control heat treating furnace temperature and transfer real time data, an embedded control system has been built. The system’s hardware was mainly made up of microcontroller, analog, digital and RS485 interfaces. It has the ability of exchanging data with an industrial PC via SCADA Modbus network. According to the features of control process and the hardware’s performance, the system’s software was divided into Modbus task, control task, I/O scan task, panel task, and alarm task. The RTX51 Tiny was used as software kernel. By running in cooperative multitasking manner, the five tasks handle various functions of the system. Each of the five tasks is introduced. The task scheduling is explained. It is shown that the system can meet the requirements of process control and data acquisition. With the advantages of multitasking structure, simple hardware frame, reliable network communication and low cost, the design method we mentioned is valuable for developing embedded control systems.

Configuration Reusing in On-Line Task Scheduling for Reconfigurable Computing Systems

Reconfigurable computing systems can be reconfigured at runtime and support partial reconfigurability which makes us able to execute tasks in a true multitasking manner. To manage such systems at runtime, a reconfigurable operating system is needed. The main part of this operating system is resource management unit which performs on-line scheduling and placement of hardware tasks at runtime. Reconfiguration overhead is an important obstacle that limits the performance of on-line scheduling algorithms in reconfigurable computing systems and increases the overall execution time. Configuration reusing (task reusing) can decrease reconfiguration overhead considerably, particularly in periodic applications or the applications in which the probability of tasks recurrence is high. In this paper, we present a technique called reusing-based scheduling (RBS), for on-line scheduling and placement in which configuration reusing is considered as a main characteristic in order to reduce reconfiguration overhead and decrease total execution time of the tasks. Several experiments have been conducted on the proposed algorithm. Obtained results show considerable improvement in overall execution time of the tasks.

Operating systems for reconfigurable embedded platforms: online scheduling of real-time tasks

Today's reconfigurable hardware devices have huge densities and are partially reconfigurable, allowing for the configuration and execution of hardware tasks in a true multitasking manner. This makes reconfigurable platforms an ideal target for many modern embedded systems that combine high computation demands with dynamic task sets. A rather new line of research is engaged in the construction of operating systems for reconfigurable embedded platforms. Such an operating system provides a minimal programming model and a runtime system. The runtime system performs online task and resource management. In this paper, we first discuss design issues for reconfigurable hardware operating systems. Then, we focus on a runtime system for guarantee-based scheduling of hard real-time tasks. We formulate the scheduling problem for the 1D and 2D resource models and present two heuristics, the horizon and the stuffing technique, to tackle it. Simulation experiments conducted with synthetic workloads evaluate the performance and the runtime efficiency of the proposed schedulers. The scheduling performance for the 1D resource model is strongly dependent on the aspect ratios of the tasks. Compared to the 1D model, the 2D resource model is clearly superior. Finally, the runtime overhead of the scheduling algorithms is shown to be acceptably low.