Abstract Extreme multi-label text classification (XMTC) annotates related labels for unknown from large-scale label sets. Transformer-based methods have become the dominant approach solving XMTC task due to their effective representation capabilities. However, existing fail effectively exploit correlation between in task. To address this shortcoming, we propose a novel model called TLC-XML, i.e., Transformer with extreme classification. TLC-XML comprises three modules: Partition, Matcher and...

Single-cell ribonucleic acid sequencing (scRNA-seq) is a high-throughput genomic technique that utilized to investigate single-cell transcriptomes. Cluster analysis can effectively reveal the heterogeneity and diversity of cells in scRNA-seq data, but existing clustering algorithms struggle with inherent high dimensionality, noise, sparsity data. To overcome these limitations, we propose algorithm: Dual Correlation Reduction network-based Extreme Learning Machine (DCRELM). First, DCRELM...

Due to the atmospheric scattering and absorption, hazy weather often occurs in our everyday life, thus reducing visibility of scenes. Single image dehazing is considered as an ill-posed challenging problem computer vision. To restore inclement weather, we propose attention-to-attention generative adversarial network (AAGAN) whose motivation human visual perceptual mechanism. More specifically, a dense channel attention model embedded into encoder. Moreover, its output projected forward...

Abstract Fuzzy extreme learning machine (FELM) is an effective algorithm for dealing with classification problems noises, which uses a membership function to effectively suppress noise in data. However, FELM has the following drawbacks: (a) The degree of samples constructed by considering only distance between and class center, not local information samples. It easy mistake some boundary noises. (b) least squares loss function, leads sensitivity feature instability re-sampling. To address...

Silent speech decoding (SSD), based on articulatory neuromuscular activities, has become a prevalent task of brain-computer interfaces (BCIs) in recent years. Many works have been devoted to surface electromyography (sEMG) from activities. However, restoring silent tonal languages such as Mandarin Chinese is still difficult. This paper proposes an optimized sequence-to-sequence (Seq2Seq) approach synthesize voice the sEMG-based speech. We extract duration information regulate using audio...

Summary In unmanned aerial vehicles navigation, path planning is aimed at obtaining the optimal safety between start and destination locations. The efficiency optimality criterion depend on environment method adopted. this paper, a general fast framework proposed for navigation. Standard A* search performed online roadmap, which consists of segments that are pre‐computed offline with aid multi‐resolution grid terminate somewhere along boundary adjacent cells. Fast marching (FMM) was employed...

The emerging high efficient video coding (HEVC) standard adopts quad-tree structure to partition the unit (CU) which is flexible and efficient. However, it causes enormous computational complexity. In this paper, a fast CU partitioning algorithm in inter prediction of HEVC proposed. Firstly, based on visual saliency map detection, depth proposed alleviate Furthermore, an early termination bits presented reduce unnecessary modes. Experimental results show that our method reduces complexity...

This thesis, subject to the obstacles exist throughout process of tuning and optimization prevailing DC Double close-loop PID controller parameters, explains realize parameters by introducing Artificial Bee Colony Algorithm on basis Nectar Collecting Behavior system. The comparison analysis, processed step response between consequences Engineering Methods, indicate that improved can help us obtain a better dynamic performance index, faster following behaviors robustness, due which practical...

Scratchpad memory (SPM) is widely utilized in many embedded systems as a software- controlled on-chip to replace the traditional cache. New non-volatile (NVM) has emerged promising candidate SRAM SPM, due its significant benefits, such low-power consumption and high performance. In particular, several representative NVMs, PCM, ReRAM, STT-RAM can build multiple-level cells (MLC) achieve even higher density. Nevertheless, this triggers off energy overhead longer access latency compared with...

Multi-label twin support vector machine (MLTSVM), being an effective multi-label classifier based on (TSVM), has been widely studied and applied due to its excellent classification performance. However, there are some disadvantages in classical MLTSVM: (a) MLTSVM needs solve a series of quadratic programming problems (QPPs), which makes learning speed lower. (b) For problems, it is very difficult obtain all labels samples. In fact, the datasets that we can only contain small amount labeled...

A novel approach to quantitatively recognize the intensity of primary taste stimuli was explored based on surface electromyography (sEMG). We captured sEMG samples under with different intensities and recognized preprocessed Support Vector Machine (SVM). The feasibility recognizing Sour, Bitter, Salty verified. signals were acquired citric acid (aq), sucrose magnesium chloride sodium glutamate (aq) concentrations, for five types tastes: Sweet, Salty, Umami, whose order fixed in this article....

Multi-label learning is a meaningful supervised task in which each sample may belong to multiple labels simultaneously. Due this characteristic, multi-label more complicated and difficult than mul... | Find, read cite all the research you need on Tech Science Press

To overcome the disadvantages of least squares twin support vector hypersphere (LS-TSVH), some improvements are proposed in this paper. First, LS-TSVH ignores local sample information; it treats each equally when constructing separating hyperspheres, which causes to be highly sensitive noisy samples. solve problem, we introduce density information into and propose a weighted (WLSTSVH) approach. Then, use Newton downhill algorithm efficiently. Furthermore, limitation that is suitable only for...

In this paper, a saliency aware fast intra coding algorithm for HEVC is proposed consists of perceptual and prediction mode decision algorithm. Firstly, based on the visual detection, an adaptive CU splitting method to reduce encoding complexity. Furthermore, quantization parameter adaptively adjusted at level according relative importance each distortion efficiently controlled. Secondly, with step halving rough early modes pruning presented selectively check potential effectively complexity...

P2P communities, is a method for arranging large numbers of peers in self configuring peer relationship based on declared attributes (or interests) the participating peers. This expected to have an impact sharing resources and pruning search spaces interests clients. Current peer- to-peer systems are targeted information sharing, file storage, searching indexing often using overlay network. In this paper we expand scope peer-to-peer include concept business environment analogous "stock...

In the past, people control robots by remote-control or keyboard which is quite difficult and not efficient. This method needs to occupy people's hands totally unfriendly disabled. So we developed a spherical robot remote system based on silent speech recognition technology. After surface EMG signals are collected equipment, preprocessed extracted features, results produced random forest sent make it move. By this way, can move according thought without hands' movements. paper, reliability...