Search

The explosive growth of various types of big data and advances in AI technologies have catalyzed a new type of workloads called multi-modal DNNs. Multi-modal DNNs are capable of interpreting and reasoning about information from multiple modalities, making them more applicable to real-world AI scenarios. In recent research, multi-modal DNNs have outperformed the best uni-modal DNN in a wide range of distributed computing applications from traditional multimedia systems to emerging autonomous edge systems. However, despite their importance and superiority, very limited research attention has been devoted to understand the characteristics of multi-modal DNNs and their implications on current computing software/hardware platforms. Existing benchmarks either target uni-modal DNNs or only focus on the algorithm characteristics of multi-modal DNNs. There lacks representative benchmark suites that provide comprehensive system and architecture level analysis of multi-modal networks. To advance the understanding of these multi-modal DNN workloads and facilitate related research, we present MMBench, an open-source, end-to-end benchmark suite consisting of a set of real-world multi-modal DNN workloads with relevant performance metrics for evaluation. We then use MMBench to conduct an in-depth analysis on the characteristics of multi-modal DNNs. We demonstrate their unique characteristics of clear multi-stage execution, frequent synchronization and high heterogeneity, which distinguish them from conventional uni-modal DNNs. Finally, we conduct a case study and extend our benchmark to edge devices. We hope that our work can provide insights for future software/hardware design and optimization to underpin multi-modal DNNs on both cloud and edge computing platforms.

Multi-modal neural networks have become increasingly pervasive in many machine learning application domains due to their superior accuracy by fusing various modalities. However, they present many unique characteristics such as multi-stage execution, frequent synchronization and high heterogeneity, which are not well understood in the system and architecture community. In this article, we first present and characterize a set of multi-modal neural network workloads of different sizes at inference stage. We then explore their important implications from system and architecture aspects. We hope that our work can help guide future software/hardware design and optimization for efficient inference of multi-modal DNN applications.

The explosive growth of various types of big data and advances in AI technologies have catalyzed a new type of workloads called multi-modal DNNs. Multi-modal DNNs are capable of interpreting and reasoning about information from multiple modalities, making them more applicable to real-world AI scenarios. In recent research, multi-modal DNNs have outperformed the best uni-modal DNN in a wide range of distributed computing applications from traditional multimedia systems to emerging autonomous edge systems. However, despite their importance and superiority, very limited research attention has been devoted to understand the characteristics of multi-modal DNNs and their implications on current computing software/hardware platforms. Existing benchmarks either target uni-modal DNNs or only focus on the algorithm characteristics of multi-modal DNNs. There lacks representative benchmark suites that provide comprehensive system and architecture level analysis of multi-modal networks. To advance the understanding of these multi-modal DNN workloads and facilitate related research, we present MMBench, an open-source, end-to-end benchmark suite consisting of a set of real-world multi-modal DNN workloads with relevant performance metrics for evaluation. We then use MMBench to conduct an in-depth analysis on the characteristics of multi-modal DNNs. We demonstrate their unique characteristics of clear multi-stage execution, frequent synchronization and high heterogeneity, which distinguish them from conventional uni-modal DNNs. Finally, we conduct a case study and extend our benchmark to edge devices. We hope that our work can provide insights for future software/hardware design and optimization to underpin multi-modal DNNs on both cloud and edge computing platforms.

Dissecting site-specific functions of O-glycosylation requires simultaneous identification and quantification of differentially expressed O-glycopeptides by mass spectrometry. However, different dissociation methods have not been systematically compared in their performance in terms of identification, glycosite localization, and quantification with isobaric labeling. Here, we conducted this comparison on highly enriched unlabeled O-glycopeptides with higher-energy collision dissociation (HCD), electron-transfer/collision-induced dissociation (ETciD), and electron transfer/higher-energy collisional dissociation (EThcD), concluding that ETciD and EThcD with optimal supplemental activation resulted in superior identification of glycopeptides and unambiguous site localizations than HCD in a database search by Sequest HT. We later described a pseudo-EThcD strategy that in silica concatenates the electron transfer dissociation spectrum with the paired HCD spectrum acquired sequentially for the same precursor ions, which combines the identification advantage of ETciD/EThcD with the superior reporter ion quality of HCD. We demonstrated its improvements in identification and quantification of isobaric mass tag-labeled O-glycopeptides and showcased the discovery of the specific glycosites of GalNAc transferase 11 (GALNT11) in HepG2 cells.

Simin Deng,1,* Lingyu Sun,2,* Weijie Chen,1,* Xiaorong Liu,2 Shangjie Chen3 1The Second School of Clinical Medicine, Southern Medical University, Dongguan Tung Wah Hospital, Guangdong, People’s Republic of China; 2Department of Rehabilitation Medicine, Dongguan Tung Wah Hospital, Guangdong, People’s Republic of China; 3Department of Rehabilitation Medicine, Affiliated Baoan Hospital of Shenzhen, Southern Medical University, Guangdong, People’s Republic of China*These authors contributed equally to this workCorrespondence: Shangjie ChenDepartment of Rehabilitation Medicine, Affiliated Baoan Hospital of Shenzhen, Southern Medical University, No. 118 Longjing Second Road, Baoan District, Shenzhen, Guangdong, People’s Republic of ChinaTel +86-755-27788311Fax +86-755-29791989Email csjme@163.comPurpose: Subjective cognitive decline (SCD) is the earliest symptom stage of Alzheimer’s disease (AD), and the APOEϵ4 allele is the strongest genetic risk factor for sporadic AD. Based on graph theory, the resting state functional connectivity (rsFC) in SCD patients with APOEϵ4 was studied to explore the effect of APOEϵ4 on the rsFC network properties of SCD patients.Patients and Methods: This cross-sectional study included MRI image data from 19 SCD patients with APOEϵ4 (SCD+), 29 SCD patients without APOEϵ4 (SCD−), and 30 normal control (NC−) individuals without APOEϵ4. We generated a binary matrix based on anatomical automatic labeling (AAL) 90 atlas to construct the functional network. We then calculated the whole brain network characteristics and intracerebral node characteristics by graph theory.Results: For the whole brain network characteristics, all three groups showed small-worldness. The SCD+ group had increased compensatory information transfer speed and enhanced integration capability. This group also had high heterogeneity for intracerebral node characteristics, m

Global forest cover is critical to the provision of certain ecosystem services. With the advent of the google earth engine cloud platform, fine resolution global land cover mapping task could be accomplished in a matter of days instead of years. The amount of global forest cover (GFC) products has been steadily increasing in the last decades. However, it's hard for users to select suitable one due to great differences between these products, and the accuracy of these GFC products has not been verified on global scale. To provide guidelines for users and producers, it is urgent to produce a validation sample set at the global level. However, this labeling task is time and labor consuming, which has been the main obstacle to the progress of global land cover mapping. In this research, a labor-efficient semi-automatic framework is introduced to build a biggest ever Forest Sample Set (FSS) contained 395280 scattered samples categorized as forest, shrubland, grassland, impervious surface, etc. On the other hand, to provide guidelines for the users, we comprehensively validated the local and global mapping accuracy of all existing 30m GFC products, and analyzed and mapped the agreement of them. Moreover, to provide guidelines for the producers, optimal sampling strategy was proposed to improve the global forest classification. Furthermore, a new global forest cover named GlobeForest2020 has been generated, which proved to improve the previous highest state-of-the-art accuracies (obtained by Gong et al., 2017) by 2.77% in uncertain grids and by 1.11% in certain grids.

In this paper, a distinct world based on the Internet was presented, that is, the virtual world. With recognizing the importance of the three parties in virtual currency market, the government, Internet companies and netizens, this paper decided to focus on the aspect of netizens, and designed a model to regulate the virtual currency, that is, the virtual currency exchange platform. The major aim of this paper was to find out the feasibility of regulating the virtual currency in China from the perspectives of netizens and improved the suggested model. The method of focus group is adopted and three main findings have been explained. Firstly the regulating should be executed gradually in case of confusion. Secondly it would be best to find a balance point between the physical qualities and psychological satisfaction treasured by netizens. Thirdly a government agency such as the central bank, with the features of reliability, authenticity and benefit the market seems to be appropriate to undertake the role of managing the virtual currency exchange. This paper expects to expand the topic of virtual currency to the financial field, and act as a pioneer, whose job is to guide the studies and desire further researches.

In this paper, a distinct world based on the Internet was presented, that is, the virtual world. With recognizing the importance of the three parties in virtual currency market, the government, Internet companies and netizens, this paper decided to focus on the aspect of netizens, and designed a model to regulate the virtual currency, that is, the virtual currency exchange platform. The major aim of this paper was to find out the feasibility of regulating the virtual currency in China from the perspectives of netizens and improved the suggested model. The method of focus group is adopted and three main findings have been explained. Firstly the regulating should be executed gradually in case of confusion. Secondly it would be best to find a balance point between the physical qualities and psychological satisfaction treasured by netizens. Thirdly a government agency such as the central bank, with the features of reliability, authenticity and benefit the market seems to be appropriate to undertake the role of managing the virtual currency exchange. This paper expects to expand the topic of virtual currency to the financial field, and act as a pioneer, whose job is to guide the studies and desire further researches.

In this paper, a distinct world based on the Internet was presented, that is, the virtual world. With recognizing the importance of the three parties in virtual currency market, the government, Internet companies and netizens, this paper decided to focus on the aspect of netizens, and designed a model to regulate the virtual currency, that is, the virtual currency exchange platform. The major aim of this paper was to find out the feasibility of regulating the virtual currency in China from the perspectives of netizens and improved the suggested model. The method of focus group is adopted and three main findings have been explained. Firstly the regulating should be executed gradually in case of confusion. Secondly it would be best to find a balance point between the physical qualities and psychological satisfaction treasured by netizens. Thirdly a government agency such as the central bank, with the features of reliability, authenticity and benefit the market seems to be appropriate to undertake the role of managing the virtual currency exchange. This paper expects to expand the topic of virtual currency to the financial field, and act as a pioneer, whose job is to guide the studies and desire further researches.

In this paper, a distinct world based on the Internet was presented, that is, the virtual world. With recognizing the importance of the three parties in virtual currency market, the government, Internet companies and netizens, this paper decided to focus on the aspect of netizens, and designed a model to regulate the virtual currency, that is, the virtual currency exchange platform. The major aim of this paper was to find out the feasibility of regulating the virtual currency in China from the perspectives of netizens and improved the suggested model. The method of focus group is adopted and three main findings have been explained. Firstly the regulating should be executed gradually in case of confusion. Secondly it would be best to find a balance point between the physical qualities and psychological satisfaction treasured by netizens. Thirdly a government agency such as the central bank, with the features of reliability, authenticity and benefit the market seems to be appropriate to undertake the role of managing the virtual currency exchange. This paper expects to expand the topic of virtual currency to the financial field, and act as a pioneer, whose job is to guide the studies and desire further researches.