Deep learning algorithm featuring continuous learning for modulation classifications in wireless networks

Although modulation classification based on deep neural network can achieve high Modulation Classification(MC)accuracies,catastrophic forgetting will occur when the neural network model continues to learn new tasks.In this paper,we simulate the dynamic wireless communication environment and focus on breaking the learning paradigm of isolated automatic MC.We innovate a research algorithm for continuous automatic MC.Firstly,a memory for storing representative old task modulation signals is built,which is employed to limit the gradient update direction of new tasks in the continuous learning stage to ensure that the loss of old tasks is also in a downward trend.Secondly,in order to better simulate the dynamic wireless communication environment,we employ the mini-batch gradient algorithm which is more suitable for continuous learning.Finally,the signal in the memory can be replayed to further strengthen the characteristics of the old task signal in the model.Simulation results verify the effectiveness of the method.

Comprehensive classifications for the endovascular recanalization of vertebral artery stump syndrome

Background:and purpose:To share our single-center vertebral artery stump syndrome(VASS)treatment experience and assess the role of comprehensive classification based on anatomic development,proximal conditions,and distal conditions(PAD).Materials and methods:Data were retrospectively collected from patients who underwent endovascular thrombectomy(EVT)at the Stroke Center of the First Hospital of Jilin University between January 2016 and December2021.Among patients with acute ischemic stroke in the posterior circulation,those with acute occlusion of the intracranial arteries and occlusion at the origin of the vertebral artery confirmed by digital subtraction angiography were selected.The clinical data were summarized and analyzed.Results:Fifteen patients with VASS were enrolled in the study.The overall success rate of surgical recanalization was 80%.The successful proximal recanalization rate was 70.6%,and the recanalization rates for P1,P2,P3,and P4 were 100%,71.4%,50%,and 66.67%,respectively.The mean operation times for the A1 and A2 types were124 and 120 min,respectively.The successful distal recanalization rate was 91.7%,and the recanalization rates for types D1,D2,D3,and D4 were 100%,83.3%,100%,and 100%,respectively.Five patients experienced perioperative complications(incidence rate:33.3%).Distal embolism occurred in three patients(incidence rate:20%).No dissection or subarachnoid hemorrhage occurred in any patient.Conclusion:EVT is a technically feasible treatment for VASS,and comprehensive PAD classification can,to a certain extent,help initially estimate the difficulty of surgery and provide guidance for interventional procedures.

Kinematic Analysis and Rock Mass Classifications for Rock Slope Failure at USAID Highways

Rock slope kinematic analysis and rock mass classifications has been conducted at the 17^(th) km to 26^(th) km of USAID(United States Agency for International Development)highway in Indonesia.This research aimed to examine the type of rock slope failures and the quality of rock mass as well.The scan-line method was performed in six slopes by using a geological compass to determine rock mass structure on the rock slope,and the condition of joints such as persistence,aperture,roughness,infilling material,weathering and groundwater conditions.Slope kinematic analysis was performed employing a stereographic projection.The rock slope quality and stability were investigated based on RMR(rock mass rating)and SMR(slope mass rating)parameters.The rock slope kinematic analysis revealed that planar failure was likely to occur in Slope 1,3,and 4,the wedge failure in Slope 1 and 6,and toppling failure in Slope 2,5,and 6.The RMR rating is ranging from 57 to 64 and can be categorized as Fair to Good rock.The SMR rating revealed that the failure probability of Slope 3 was 90%,while it was from 40%to 60%for others.Despite the uniform RMR for all slopes,the SMR was significantly different.The detailed quantitative consideration of orientation of joint sets and geometry of the slope contributed to such differences in outcomes.

Geotechnical Behavior of Eastern Senegal Rock Mass Slopes by Geomechanical Classifications and Applications

In Senegal, rock mechanic studies began in last four years and were a very new engineering domain. So, the first stage is to characterize geological materials. It is on this way that this work is done. This paper analyzes geomechanical parameters of sandstones of Dindifélo and basalt of Bafoundou belonging to the Proterozoic rocks domain of eastern Senegal. By Rock Mass Rating (RMR) and Geological Strength Index (GSI), sandstones of the Dindifello Cliff show fair to poor characteristics while basalts of the hills of Bafoundou are fair quality. In addition to the qualities of rock mass, Young moduli, uniaxial compressive strengths and tensile strength of rock mass are also defined using Rock Quality Designation (RQD) GSI, RMR. Hoek-Brown parameters m and a, depend both on the fracturation and the content fine in the rock. Values of mechanical parameters are different when deduced from RMR, GSI and from intact rock laboratory test. Those differences are due to variables taken account. The variation depends also on the quality of the rock. Statistical analysis shows possible unstabilities which depend on rock mass parameters but with acceptable probability of failure. Probability of failure is the highest when deducing from Mohr criterion than from Hoek-Brown criterion.

Assessment of the cropland classifications in four global land cover datasets： A case study of Shaanxi Province, China

Accurate and reliable cropland surface information is of vital importance for agricultural planning and food security monitoring. As several global land cover datasets have been independently released, an inter-comparison of these data products on the classification of cropland is highly needed. This paper presents an assessment of cropland classifications in four global land cover datasets, i.e., moderate resolution imaging spectrometer （MODIS） land cover product, global land cover map of 2009 （GlobCover2009）, finer resolution observation and monitoring of global cropland （FROM-GC） and 30-m global land cover dataset （GlobeLand30）. The temporal coverage of these four datasets are circa 2010. One of the typical agricultur- al regions of China, Shaanxi Province, was selected as the study area. The assessment proceeded from three aspects： accuracy, spatial agreement and absolute area. In accuracy assessment, 506 validation samples, which consist of 168 cropland samples and 338 non-cropland ones, were automatically and systematically selected, and manually interpreted by referencing high-resolution images dated from 2009 to 2011 on Google Earth. The results show that the overall accuracy （OA） of four datasets ranges from 61.26 to 80.63%. GlobeLand30 dataset, with the highest accuracy, is the most accurate dataset for cropland classification. The cropland spatial agreement （mainly located in the plain ecotope of Shaanxi） and the non-cropland spatial agreement （sparsely distributed in the south and middle of Shaanxi） of the four datasets only makes up 33.96% of the whole province. FIROM-GC and GlobeLand30, obtaining the highest spatial agreement index of 62.40%, have the highest degree of spatial consistency. In terms of the absolute area, MODIS underestimates the cropland area, while GlobCover2009 significantly overestimates it. These findings are of value in revealing to which extent and on which aspect that these global land cover datasets may agree with each other at small scale on each ecotope region. The approaches taken in this study could be used to derive a fused cropland classification dataset.

Emerging molecular classifications and therapeutic implications for gastric cancer

Gastric cancer(GC) is a highly aggressive and life-threatening malignancy.Even with radical surgical removal and front-line chemotherapy,more than half of GCs locally relapse and metastasize at a distant site.The dismal outcomes reflect the ineffectiveness of a one-size fits-all approach for a highly heterogeneous disease with diverse etiological causes and complex molecular underpinnings.The recent comprehensive genomic and molecular profiling has led to our deepened understanding of GC.The emerging molecular classification schemes based on the genetic,epigenetic,and molecular signatures are providing great promise for the development of more effective therapeutic strategies in a more personalized and precise manner.To this end,the Cancer Genome Atlas(TCGA) research network conducted a comprehensive molecular evaluation of primary GCs and proposed a new molecular classification dividing GCs into four subtypes:Epstein-Barr virus-associated tumors,microsatellite unstable tumors,genomically stable tumors,and tumors with chromosomal instability.This review primarily focuses on the TCGA molecular classification of GCs and discusses the implications on novel targeted therapy strategies.We believe that these fundamental findings will support the future application of targeted therapies and will guide our efforts to develop more efficacious drugs to treat human GCs.

Clinical significance of different periampullary diverticulum classifications for endoscopic retrograde cholangiopancreatography cannulation

BACKGROUND Different types of periampullary diverticulum(PAD) may differentially affect the success of endoscopic retrograde cholangiopancreatography(ERCP) cannulation,but the clinical significance of the two current PAD classifications for cannulation is limited.AIM To verify the clinical value of our newly proposed PAD classification.METHODS A new PAD classification(Li-Tanaka classification) was proposed at our center.All PAD patients with native papillae who underwent ERCP from January 2012 to December 2017 were classified according to three classification systems, and the effects of various types of PAD on ERCP cannulation were compared.RESULTS A total of 3564 patients with native papillae were enrolled, including 967(27.13%)PAD patients and 2597(72.87%) non-PAD patients. In the Li-Tanaka classification, type Ⅰ PAD patients exhibited the highest difficult cannulation rate(23.1%, P = 0.01), and type Ⅱ and Ⅳ patients had the highest cannulation success rates(99.4% in type Ⅱ and 99.3% in type Ⅳ, P < 0.001). In a multivariableadjusted logistic model, the overall successful cannulation rate in PAD patients was higher than that in non-PAD patients [odds ratio(OR) = 1.87, 95% confidence interval(CI): 1.04-3037, P = 0.037]. In addition, compared to the non-PAD group,the difficulty of cannulation in the type Ⅰ PAD group according to the Li-Tanaka classification was greater(OR = 2.04, 95%CI: 1.13-3.68, P = 0.004), and the successful cannulation rate was lower(OR = 0.27, 95%CI: 0.11-0.66, P < 0.001),while it was higher in the type Ⅱ PAD group(OR = 4.44, 95%CI: 1.61-12.29, P <0.01).CONCLUSION Among the three PAD classifications, the Li-Tanaka classification has an obvious clinical advantage for ERCP cannulation, and it is helpful for evaluating potentially difficult and successful cannulation cases among different types of PAD patients.

Classical and nonclassical symmetry classifications of nonlinear wave equation with dissipation

A complete classical symmetry classification and a nonclassical symmetry classification of a class of nonlinear wave equations are given with three arbitrary parameter functions. The obtained results show that such nonlinear wave equations admit richer classical and nonclassical symmetries, leading to the conservation laws and the reduction of the wave equations. Some exact solutions of the considered wave equations for particular cases are derived.

Relationship between subtypes of MRI-confirmed acute ischemic stroke based on OCSP and TOAST classifications

BACKGROUND： Two classification systems exist for subtypes of acute cerebral infarction. One was developed for the Trial of Org 10172 in Acute Stroke Treatment （TOAST）, based primarily on etiology. The other is the Oxfordshire Community Stroke Project （OCSP）, based on clinical features. OBJECTIVE： To evaluate the relationship between OCSP and TOAST classifications in terms of stroke location and etiology in 126 patients with acute ischemic stroke confirmed by transcranial magnetic resonance imaging （MRI）. DESIGN, TIME AND SETTING： Retrospective case analysis. Transcranial MRI, diffusion weighted imaging, and magnetic resonance angiography were performed in 126 patients with acute stroke during the first 48 hours following admission to the Department of Neurology, Navy General Hospital of Chinese PLA. PARTICIPANTS： A total of 126 patients with acute stroke, comprised of 71 males and 55 females, admitted to the Navy General Hospital of Chinese PLA between December 2005 and April 2006 were included. METHODS： Of 126 patients with acute stroke, 13 exhibited total anterior circulation infarcts （TACI）, 51 had partial anterior circulation infarcts （PACI）, 28 suffered posterior circulation infarcts （POCI）, and 34 had lacunar infarcts （LACI） based on OCSP classification. However, according to TOAST classification, 19 cases were a result of large-artery atherosclerosis, 32 by cardioembolism, 36 by small-vessel occlusion, 1 by stroke of other determined etiology, and 38 by stroke of undetermined etiology. MAIN OUTCOME MEASURES： The corresponding relationship of the subtypes of acute stroke based on OCSP and TOAST. RESULTS： Of patients with TACI, 8 （61.5%） were caused by cardioembolism. Of patients with PACI, 16 （31.4%） were caused by large-artery atherosclerosis and 17 （33.3%） by cardioembolism. Of patients with POCI, 12 （42.8%） were a result of small-vessel occlusion. Of patients with LACI, 17 （50.0%） were caused by hypertension and arteriolar sclerosis. CONCLUSION： （1） The OCSP system is related to anatomical and pathophysiological processes, and can correctly identify the size, location, and cause of cerebral infarcts. It can also act as a reference for prognosis estimation and recurrence prevention of ischemic stroke. （2） The stroke of undetermined etiology occupies the largest percent in the subtypes of stroke, according to TOAST. （3） The classification of OCSP exhibits a close relationship to TOAST.

Applying rock mass classifications to carbonate rocks for engineering purposes with a new approach using the rock engineering system

Classical rock mass classification systems are not applicable to carbonate rocks,especially when these are affected by karst processes.Their applications to such settings could therefore result in outcomes not representative of the real stress-strain behavior.In this study,we propose a new classification of carbonate rock masses for engineering purposes,by adapting the rock engineering system(RES) method by Hudson for fractured and karstified rock masses,in order to highlight the problems of implementation of geomechanical models to carbonate rocks.This new approach allows a less rigid classification for carbonate rock masses,taking into account the local properties of the outcrops,the site conditions and the type of engineering work as well.

Traditional Medicine Diagnostic Codes in ICD-11 and Alternative Diagnostic Classifications in the Mainstream Healthcare

In 2018,the 11^(th) Edition of the International Classification of Diseases(ICD-11)defined a diagnostic code list for standard traditional medicine(TM)conditions.The codes improve patient safety by providing more comprehensive and accurate medical records for hospitals in the Western Pacific Region.In these facilities,TM is often a standard of care for those populations.In several mainstream media sources,writers are circumventing evidence-based peer-reviewed medical literature by unduly influencing public opinion and,in this case,against the new ICD-11 codes.The dangers imposed by the transgression of popular writing onto the discipline of peer-reviewed works are present since best practices in medical record-keeping will fail without the inclusion of TM in the ICD-11 codes.Such failures directly affect the health of the patients and policymakers in regions where TM and conventional medicine are combined.This article investigates the boundaries between substantial evidence and popular opinion.In this era where media is used to manipulate evidence,the reader’s use of sound judgment and critical thought are thwarted.This article also challenges three controversial themes in pop literature,including the threat to endangered species,increased patient risk,and contaminants in the TM.These themes are made without evidence and are,in fact,of flawed logic.There is no reason to assume that improved medical record-keeping and knowledge of patient cases increase risks.

Diagnostic performance of endoscopic classifications for neoplastic lesions in patients with ulcerative colitis:A retrospective casecontrol study

BACKGROUND It is unclear whether the Japan Narrow-Band Imaging Expert Team(JNET)classification and pit pattern classification are applicable for diagnosing neoplastic lesions in patients with ulcerative colitis(UC).AIM To clarify the diagnostic performance of these classifications for neoplastic lesions in patients with UC.METHODS This study was conducted as a single-center,retrospective case-control study.Twenty-one lesions in 19 patients with UC-associated neoplasms(UCAN)and 23 lesions in 22 UC patients with sporadic neoplasms(SN),evaluated by magnifying image-enhanced endoscopy,were retrospectively and separately assessed by six endoscopists(three experts,three non-experts),using the JNET and pit pattern classifications.The results were compared with the pathological diagnoses to evaluate the diagnostic performance.Inter-and intra-observer agreements were calculated.RESULTS In this study,JNET type 2 A and pit pattern typeⅢ/Ⅳwere used as indicators of low-grade dysplasia,JNET type 2 B and pit pattern typeⅥlow irregularity were used as indicators of highgrade dysplasia to shallow submucosal invasive carcinoma,JNET type 3 and pit pattern typeⅥhigh irregularity/VN were used as indicators of deep submucosal invasive carcinoma.In the UCAN group,JNET type 2 A and pit pattern typeⅢ/Ⅳhad a low positive predictive value(PPV;50.0%and 40.0%,respectively);however,they had a high negative predictive value(NPV;94.7%and 100%,respectively).Conversely,in the SN group,JNET type 2 A and pit pattern typeⅢ/Ⅳhad a high PPV(100%for both)but a low NPV(63.6%and 77.8%,respectively).In both groups,JNET type 3 and pit pattern typeⅥ-high irregularity/VN showed high specificity.The interobserver agreement of JNET classification and pit pattern classification for UCAN among experts were 0.401 and 0.364,in the same manner for SN,0.666 and 0.597,respectively.The intra-observer agreements of JNET classification and pit pattern classification for UCAN among experts were 0.387,0.454,for SN,0.803 and 0.567,respectively.CONCLUSION The accuracy of endoscopic diagnosis using both classifications was lower for UCAN than for SN.Endoscopic diagnosis of UCAN tended to be underestimated compared with the pathological results.

Peridotites and Chromitites from the Dingqing Ophiolite in the Eastern Segment of Bangong–Nujiang Suture Zone, Tibet: Occurrence Characteristics and Classifications

The Dingqing ophiolite is located in the eastern segment of the Bangong-Nujiang suture zone. This suture zone is W–E trending parallel with the Yarlung–Zangbo suture zone and is an strategic area for exploring chromite deposits in China. The Dingqign ophiolite is distributed in near SE-NW direction. According to the spatial distribution, the Dingqing ophiolite is sudivided into two massifs, including the East and the West massifs. The Dingqing ophiolite covers an area of nearly 600 km2. This ophiolite is composed of peridotite, pyroxenite, gabbro, diabase, basalt, plagiogranite and chert(Fig. 1). The peridotite is the main lithology of the Dingqing ophiolite. The peridotite covers about 90% of the total area of the Dingqing ophiolite. The Dingqing ophiolite is dominated by harzburgite with a small amounts of dunite. The Dingqing harzburgite displays different textures, such as massive, Taxitic, oriented and spherulitic textures(Fig. 2d–i). These four types of harzburgite occur in both the East and West massifs, especially in the Laraka area of the eastern part of the East massif. Dunites have different occurrences in the field outcrops, such as lenticular or stripshaped, thin-shell and agglomerate varieties(Fig. 2a–c). On the basis of detailed field work, we have discovered 83 chromitite bodies, including 27 in the East massif and 56 in the West massif. According to the occurrence scale and quantity of the chromitite bodies, we have identified four prospecting areas, namely Laraka, Latanguo, Langda and Nazona. Chromitites in the Dingqing ophiolite show different textures, including massive, disseminated, veined and disseminated-banded textures(Fig. 3). On the basis of the Cr#(=Cr/(Cr+Al)×100) of chromite, we have classified the Dingqing chromitite into high-Cr, medium high chromium type, medium chromium type and low chromium type chromitite(Figs. 4, 5). Among them, low chromium type chromitite Cr# is extremely low, ranging from 9.23 to 14.01, with an average of 11.89;TiO2 content is 0.00% to 0.04%, and the average value is 0.01%, which may be a new output type of chromitite. These different types of chromitites have different associations/assemblages of mineral inclusions. The inclusions in high chromium type chromitite are mainly clinopyroxene and a small amount of olivine;medium high chromium chromitite are mainly amphibole, a small amount of clinopyroxene and phlogopite;while low-chromium chromite rarely develops mineral inclusions, and micron-sized clinopyroxene inclusions are common in olivines which are gangue mineral in it. These different types of chromite ore bodies have a certain correspondence with the field output, and may also restrict their genesis. This part will be further developed in the follow-up work.

Erratum:“Gamma-ray bursts with extended emission:classifications,energy correlations and radiation properties”(2020,RAA,20,201)

As a result of an error by the authors,in the paper,one of the authors’names,“Xu-Juan Li”,was misspelled.The correct spelling of the name should be“Xiu-Juan Li”.

Enhancing Cloud Performance Using File Format Classifications

Metaheuristic approaches in cloud computing have shown significant results due to theirmulti-objective advantages.These approaches are now considering hybridmetaheuristics combining the relative optimized benefits of two or more algorithms resulting in the least tradeoffs among several factors.The critical factors such as execution time,throughput time,response time,energy consumption,SLA violations,communication overhead,makespan,and migration time need careful attention while designing such dynamic algorithms.To improve such factors,an optimizedmulti-objective hybrid algorithm is being proposed that combines the relative advantages of Cat Swarm Optimization(CSO)with machine learning classifiers such as Support Vector Machine(SVM).The adopted approach is based on SVMone to many classification models of machine learning that performs the classifications of various data format types in the cloud with best accuracy.In CSO,grouping phase is used to divide the data files as audio,video,image,and text which is further extended by polynomial Kernel function based on various input features and used for optimized load balancing.Overall,proposed approach works well and achieved performance efficiency in evaluated QoS metrics such as average energy consumption by 12%,migration time by 9%,and optimization time by 10%,in the presence of competitor baselines.

Optimal Hybrid Feature Extraction with Deep Learning for COVID-19 Classifications

Novel coronavirus 2019(COVID-19)has affected the people’s health,their lifestyle and economical status across the globe.The application of advanced Artificial Intelligence(AI)methods in combination with radiological imaging is useful in accurate detection of the disease.It also assists the physicians to take care of remote villages too.The current research paper proposes a novel automatedCOVID-19 analysismethod with the help ofOptimal Hybrid Feature Extraction(OHFE)and Optimal Deep Neural Network(ODNN)called OHFE-ODNN from chest x-ray images.The objective of the presented technique is for performing binary and multi-class classification of COVID-19 analysis from chest X-ray image.The presented OHFE-ODNN method includes a sequence of procedures such as Median Filtering(MF)-based pre-processed,feature extraction and finally,binary(COVID/Non-COVID)and multiclass(Normal,COVID,SARS)classification.Besides,in OHFE-based feature extraction,Gray Level Co-occurrence Matrix(GLCM)and Histogram of Gradients(HOG)are integrated together.The presented OHFE-ODNN model includes Squirrel Search Algorithm(SSA)for finetuning the parameters of DNN.The performance of the presented OHFEODNN technique is conducted using chest x-rays dataset.The presented OHFE-ODNN method classified the binary classes effectively with a maximumprecision of 95.82%,accuracy of 94.01%and F-score of 96.61%.Besides,multiple classes were classified proficiently by OHFE-ODNN model with a precision of 95.63%,accuracy of 95.60%and an F-score of 95.73%.

Inter-and intra-rater reliability of vertebral fracture classifications in the Swedish fracture register

AIM To investigate the inter-and intra-rater reliability of the vertebral fracture classifications used in the Swedish fracture register.METHODS Radiological images of consecutive patients with cervical spine fractures(n = 50)were classified by 5 raters with different experience levels at two occasions. An identical process was performed with thoracolumbar fractures(n = 50). Cohen's kappa was used to calculate the inter-and intra-rater reliability.RESULTS The mean kappa coefficient for inter-rater reliability ranged between 0.54 and0.79 for the cervical fracture classifications, between 0.51 and 0.72 for the thoracolumbar classifications(overall and for different sub classifications), and between 0.65 and 0.77 for the presence or absence of signs of ankylosing disorder in the fracture area. The mean kappa coefficient for intra-rater reliability ranged between 0.58 and 0.80 for the cervical fracture classifications, between 0.46 and0.68 for the thoracolumbar fracture classifications(overall and for different sub classifications) and between 0.79 and 0.81 for the presence or absence of signs of ankylosing disorder in the fracture area.CONCLUSION The classifications used in the Swedish fracture register for vertebral fractures have an acceptable inter-and intra-rater reliability with a moderate strength of agreement.

Classifications of Satellite Imagery for Identifying Urban Area Structures

This study compares three types of classifications of satellite data to identify the most suitable for making city maps in a semi-arid region. The source of our data was GeoEye 1 satellite. To classify this data, two pro-grammes were used: an Object-Based Classification and a Pixel-Based Classification. The second classification programme was further subdi-vided into two groups. The first group included classes (buildings, streets, vacant land, vegetations) which were treated simultaneously and on a single image basis. The second, however, was where each class was identified individually, and the results of each class produced a single image and were later enhanced. The classification results were then as-sessed and compared before and after enhancement using visual then automatic assessment. The results of the evaluation showed that the pix-el-based individual classification of each class was rated the highest after enhancement, increasing the Overall Classification Accuracy by 2%, from 89% to 91.00%. The results of this classification type were adopted for mapping Jeddah’s buildings, roads, and vegetations.

Climate paleogeography knowledge graph and deep time paleoclimate classifications

The climate paleogeography,especially the climate classifications,helps to interpret the global and regional climate changes and intuitively compare the climate conditions in different regions.However,the application of climate classification in deep time(i.e.,climate paleogeography)is prohibited due to the usually qualitatively constrained paleoclimate and the inconsistent descriptions and semantic heterogeneity of the climate types.In this study,a climate paleogeography knowledge graph is established under the framework of the Deep-Time Digital Earth program(DDE).The hierarchical knowledge graph consists of five paleoclimate classifications based on various strategies.The classifications are described and their strengths and weaknesses are fully evaluated in four aspects:“simplicity,applicability,quantifiability,and comparability”.We also reconstruct the global climate distributions in the Late Cretaceous according to these classifications.The results are compared and the relationships among these climate types in different classifications are evaluated.Our study unifies scientific concepts from different paleoclimate classifications,which provides an important theoretical basis for the application of paleoclimate classifications in deep time.

AI Fairness-From Machine Learning to Federated Learning

This article reviews the theory of fairness in AI-frommachine learning to federated learning,where the constraints on precision AI fairness and perspective solutions are also discussed.For a reliable and quantitative evaluation of AI fairness,many associated concepts have been proposed,formulated and classified.However,the inexplicability of machine learning systems makes it almost impossible to include all necessary details in the modelling stage to ensure fairness.The privacy worries induce the data unfairness and hence,the biases in the datasets for evaluating AI fairness are unavoidable.The imbalance between algorithms’utility and humanization has further reinforced suchworries.Even for federated learning systems,these constraints on precision AI fairness still exist.Aperspective solution is to reconcile the federated learning processes and reduce biases and imbalances accordingly.