The influence of resistance exercise training prescription variables on skeletal muscle mass,strength,and physical function in healthy adults:An umbrella review

Purpose:The aim of this umbrella review was to determine the impact of resistance training(RT)and individual RT prescription variables on muscle mass,strength,and physical function in healthy adults.Methods:Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses(PRISMA)guidelines,we systematically searched and screened eligible systematic reviews reporting the effects of differing RT prescription variables on muscle mass(or its proxies),strength,and/or physical function in healthy adults aged>18 years.Results:We identified 44 systematic reviews that met our inclusion criteria.The methodological quality of these reviews was assessed using A Measurement Tool to Assess Systematic Reviews;standardized effectiveness statements were generated.We found that RT was consistently a potent stimulus for increasing skeletal muscle mass(4/4 reviews provide some or sufficient evidence),strength(4/6 reviews provided some or sufficient evidence),and physical function(1/1 review provided some evidence).RT load(6/8 reviews provided some or sufficient evidence),weekly frequency(2/4 reviews provided some or sufficient evidence),volume(3/7 reviews provided some or sufficient evidence),and exercise order(1/1 review provided some evidence)impacted RT-induced increases in muscular strength.We discovered that 2/3 reviews provided some or sufficient evidence that RT volume and contraction velocity influenced skeletal muscle mass,while 4/7 reviews provided insufficient evidence in favor of RT load impacting skeletal muscle mass.There was insufficient evidence to conclude that time of day,periodization,inter-set rest,set configuration,set end point,contraction velocity/time under tension,or exercise order(only pertaining to hypertrophy)influenced skeletal muscle adaptations.A paucity of data limited insights into the impact of RT prescription variables on physical function.Conclusion:Overall,RT increased muscle mass,strength,and physical function compared to no exercise.RT intensity(load)and weekly frequency impacted RT-induced increases in muscular strength but not muscle hypertrophy.RT volume(number of sets)influenced muscular strength and hypertrophy.

Surgical intervention combined with weight-bearing walking training promotes recovery in patients with chronic spinal cord injury:a randomized controlled study

For patients with chronic spinal cord injury,the co nventional treatment is rehabilitation and treatment of spinal cord injury complications such as urinary tract infection,pressure sores,osteoporosis,and deep vein thrombosis.Surgery is rarely perfo rmed on spinal co rd injury in the chronic phase,and few treatments have been proven effective in chronic spinal cord injury patients.Development of effective therapies fo r chronic spinal co rd injury patients is needed.We conducted a randomized controlled clinical trial in patients with chronic complete thoracic spinal co rd injury to compare intensive rehabilitation(weight-bearing walking training)alone with surgical intervention plus intensive rehabilitation.This clinical trial was registered at ClinicalTrials.gov(NCT02663310).The goal of surgical intervention was spinal cord detethering,restoration of cerebrospinal fluid flow,and elimination of residual spinal cord compression.We found that surgical intervention plus weight-bearing walking training was associated with a higher incidence of American Spinal Injury Association Impairment Scale improvement,reduced spasticity,and more rapid bowel and bladder functional recovery than weight-bearing walking training alone.Overall,the surgical procedures and intensive rehabilitation were safe.American Spinal Injury Association Impairment Scale improvement was more common in T7-T11 injuries than in T2-T6 injuries.Surgery combined with rehabilitation appears to have a role in treatment of chronic spinal cord injury patients.

Numerical analysis of moving train induced vibrations on tunnel,surrounding ground and structure

This study is focused on the effect of vibration induced by moving trains in tunnels on the surrounding ground and structures.A three-dimensional finite element model is established for a one-track railway tunnel and an adjacent twelve-storey building frame by using commercial software Midas GTS-NX(2019)and Midas Gen.This study considered the moving load effect of a complete train,which varies with space as well as with time.The effect of factors such as train speed,overburden pressure on the tunnel and variation in soil properties are studied in the time domain.As a result,the variations in horizontal and vertical acceleration for two different sites,i.e.,the free ground surface(without structure)and the area containing the structure,are compared.Also,the displacement pattern of the raft foundation is plotted for different train velocities.At lower speeds,the heaving phenomenon is negligible,but as the speed increases,both the heaving and differential settlement increase in the foundation.This study demonstrates that the effect of moving train vibrations should be considered in the design of new nearby structures and proper ground improvement should be considered for existing structures.

Application value research of swallowing treatment device combined with swallowing rehabilitation training in the treatment of swallowing disorders after stroke

BACKGROUND Stroke is a common disabling disease,whether it is ischemic stroke or hemorrhagic stroke,both can result in neuronal damage,leading to various manifestations of neurological dysfunction.AIM To explore of the application value of swallowing treatment device combined with swallowing rehabilitation training in the treatment of swallowing disorders after stroke.METHODS This study selected 86 patients with swallowing disorders after stroke admitted to our rehabilitation department from February 2022 to December 2023 as research subjects.They were divided into a control group(n=43)and an observation group(n=43)according to the treatment.The control group received swallowing rehabilitation training,while the observation group received swallowing treatment device in addition to the training.Both groups underwent continuous intervention for two courses of treatment.RESULTS The total effective rate in the observation group(93.02%)was higher than that in the control group(76.74%)(P=0.035).After intervention,the oral transit time,swallowing response time,pharyngeal transit time,and laryngeal closure time decreased in both groups compared to before intervention.In the observation group,the oral transit time,swallowing response time,and pharyngeal transit time were shorter than those in the control group after intervention.However,the laryngeal closure time after intervention in the observation group was compared with that in the control group(P=0.142).After intervention,average amplitude value and duration of the genioglossus muscle group during empty swallowing and swallowing 5 mL of water are reduced compared to before intervention in both groups.After intervention,the scores of the chin-tuck swallowing exercise and the Standardized Swallowing Assessment are both reduced compared to pre-intervention levels in both groups.However,the observation group scores lower than the control group after intervention.Additionally,the Functional Oral Intake Scale scores of both groups are increased after intervention compared to pre-intervention levels,with the observation group scoring higher than the control group after intervention(P<0.001).The cumulative incidence of complications in the observation group is 9.30%,which is lower than the 27.91%in the control group(P=0.027).CONCLUSION The combination of swallowing therapy equipment with swallowing rehabilitation training can improve the muscle movement level of the genioglossus muscle group,enhance swallowing function,and prevent the occurrence of swallowing-related complications after stroke.

NeTrainSim:a network-level simulator for modeling freight train longitudinal motion and energy consumption

Although train modeling research is vast, most available simulation tools are confined to city-or trip-scale analysis, primarily offering micro-level simulations of network segments. This paper addresses this void by developing the Ne Train Sim simulator for heavy long-haul freight trains on a network of multiple intersecting tracks. The main objective of this simulator is to enable a comprehensive analysis of energy consumption and the associated carbon footprint for the entire train system. Four case studies were conducted to demonstrate the simulator's performance. The first case study validates the model by comparing Ne Train Sim output to empirical trajectory data. The results demonstrate that the simulated trajectory is precise enough to estimate the train energy consumption and carbon dioxide emissions. The second application demonstrates the train-following model considering six trains following each other. The results showcase the model ability to maintain safefollowing distances between successive trains. The next study highlights the simulator's ability to resolve train conflicts for different scenarios. Finally, the suitability of the Ne Train Sim for modeling realistic railroad networks is verified through the modeling of the entire US network and comparing alternative powertrains on the fleet energy consumption.

Research and Practice of Cardiopulmonary Rehabilitation Training Programs

In today’s society, the incidence of cardiopulmonary diseases is increasing annually, seriously affecting patients’ quality of life. Therefore, developing a scientific and effective rehabilitation training program is of great significance. This study first analyzes the theoretical basis of cardiopulmonary rehabilitation training, including the effects of aerobic exercise, interval training, and strength training on cardiopulmonary function. Based on this, a comprehensive rehabilitation training program is designed, which includes personalized training plans, comprehensive interventions, multidisciplinary collaboration, patient education, and regular follow-up visits. The cardiopulmonary rehabilitation training plan developed in this study has certain scientific practicability, which provides a theoretical basis for cardiopulmonary rehabilitation training, and also provides a reference for medical institutions, rehabilitation centers and communities, which is helpful for promotion and application to a wider range of patients with cardiopulmonary diseases.

GeoNER:Geological Named Entity Recognition with Enriched Domain Pre-Training Model and Adversarial Training

As important geological data,a geological report contains rich expert and geological knowledge,but the challenge facing current research into geological knowledge extraction and mining is how to render accurate understanding of geological reports guided by domain knowledge.While generic named entity recognition models/tools can be utilized for the processing of geoscience reports/documents,their effectiveness is hampered by a dearth of domain-specific knowledge,which in turn leads to a pronounced decline in recognition accuracy.This study summarizes six types of typical geological entities,with reference to the ontological system of geological domains and builds a high quality corpus for the task of geological named entity recognition(GNER).In addition,Geo Wo BERT-adv BGP(Geological Word-base BERTadversarial training Bi-directional Long Short-Term Memory Global Pointer)is proposed to address the issues of ambiguity,diversity and nested entities for the geological entities.The model first uses the fine-tuned word granularitybased pre-training model Geo Wo BERT(Geological Word-base BERT)and combines the text features that are extracted using the Bi LSTM(Bi-directional Long Short-Term Memory),followed by an adversarial training algorithm to improve the robustness of the model and enhance its resistance to interference,the decoding finally being performed using a global association pointer algorithm.The experimental results show that the proposed model for the constructed dataset achieves high performance and is capable of mining the rich geological information.

Attack-Resilient Distributed Cooperative Control of Virtually Coupled High-Speed Trains via Topology Reconfiguration

Dear Editor,This letter addresses the resilient distributed cooperative control problem of a virtually coupled train convoy under stochastic disturbances and cyber attacks.The main purpose is to achieve distributed coordination of virtually coupled high-speed trains with the prescribed inter-train distance and same cruise velocity.

Impact of task-oriented training based on acupuncture rehabilitation on upper extremity function and quality of life of patients with early stroke

BACKGROUND Eighty percent of stroke patients develop upper limb dysfunction,especially hand dysfunction,which has a very slow recovery,resulting in economic burden to families and society.AIM To investigate the impact of task-oriented training based on acupuncture therapy on upper extremity function in patients with early stroke.METHODS Patients with early stroke hemiplegia who visited our hospital between January 2021 and October 2022 were divided into a control group and an observation group,each with 50 cases.The control group underwent head acupuncture plus routine upper limb rehabilitation training(acupuncture therapy).In addition to acupuncture and rehabilitation,the observation group underwent upper limb task-oriented training(30 min).Each group underwent treatment 5 d/wk for 4 wk.Upper extremity function was assessed in both groups using the Fugl-Meyer Assessment-Upper Extremity(FMA-UE),Wolf Motor Function Rating Scale(WMFT),modified Barthel Index(MBI),and Canadian Occupational Performance Measure(COPM).Quality of life was evaluated using the Short-Form 36-Item Health Survey(SF-36).Clinical efficacy of the interventions was also evaluated.RESULTS Before intervention,no significant differences were observed in the FMA-UE,MBI,and WMFT scores between the two groups(P>0.05).After intervention,the FMA-UE,WMFT,MBI,COPM-Functional Mobility and Satisfaction,and SF-36 scores increased in both groups(P<0.05),with even higher scores in the observation group(P<0.05).The observation group also obtained a higher total effective rate than the control group(P<0.05).CONCLUSION Task-oriented training based on acupuncture rehabilitation significantly enhanced upper extremity mobility,quality of life,and clinical efficacy in patients with early stroke.

Evaluating the feasibility and preliminary effects of an online compassion training program for nursing students:A pilot randomized controlled trial

Objectives:This study aimed to assess the feasibility of an online compassion training program for nursing students and preliminarily investigate its effects on mindfulness,self-compassion,and stress reduction.Methods:This study employed a randomized controlled trial design.Second-year students from a nursing college in Guangzhou,China,were recruited as research participants in August 2023.The intervention group participated in an 8-week online compassion training program via the WeChat platform,comprising three stages:mindfulness(weeks 1e2),self-compassion(weeks 3e5),and compassion for others(weeks 6 e8).Each stage included four activities:psychoeducation,mindfulness practice,weekly diary,and emotional support.Program feasibility was assessed through recruitment and retention rates,program engagement,and participant acceptability.Program effectiveness was measured with the Mindful Attention Awareness Scale,Self-Compassion Scale-Short Form,and Perceived Stress Scale.Results:A total of 28 students completed the study(13 in the intervention group,15 in the control group).The recruitment rate was 36.46%,with a high retention rate of 93.3%.Participants demonstrated high engagement:69.2%accessed learning materials every 1e2 days,93.3%practiced mindfulness at least weekly,with an average of 4.69 diary entries submitted per person and 23.30 WeChat interactions with instructors.Regarding acceptability,all participants expressed satisfaction with the program,with 92.4%finding it“very helpful”or“extremely helpful.”In terms of intervention effects,the intervention group showed a significant increase in mindfulness levels from pre-intervention(51.54±10.93)to postintervention(62.46±13.58)(P<0.05),while no significant change was observed in the control group.Although there were no statistically significant differences between the two groups in post-intervention self-compassion and perceived stress levels,the intervention group showed positive trends:selfcompassion levels increased(35.85±8.60 vs.40.85±5.54),and perceived stress levels slightly decreased(44.77±8.65 vs.42.00±5.77).Conclusions:This pilot study demonstrated the feasibility of an online compassion training program for nursing students and suggested its potential effectiveness in enhancing mindfulness,self-compassion,and stress reduction.Despite limitations such as small sample size and lack of long-term follow-up,preliminary evidence indicates promising prospects for integrating such training into nursing education.Further research is warranted to confirm thesefindings and assess the sustained impact of this approach on nursing education and practice.

Increasing realism in modelling energy losses in railway vehicles and their impact to energy-efficient train control

The reduction of energy consumption is an increasingly important topic of the railway system.Energy-efficient train control(EETC)is one solution,which refers to mathematically computing when to accelerate,which cruising speed to hold,how long one should coast over a suitable space,and when to brake.Most approaches in literature and industry greatly simplify a lot of nonlinear effects,such that they ignore mostly the losses due to energy conversion in traction components and auxiliaries.To fill this research gap,a series of increasingly detailed nonlinear losses is described and modelled.We categorize an increasing detail in this representation as four levels.We study the impact of those levels of detail on the energy optimal speed trajectory.To do this,a standard approach based on dynamic programming is used,given constraints on total travel time.This evaluation of multiple test cases highlights the influence of the dynamic losses and the power consumption of auxiliary components on railway trajectories,also compared to multiple benchmarks.The results show how the losses can make up 50%of the total energy consumption for an exemplary trip.Ignoring them would though result in consistent but limited errors in the optimal trajectory.Overall,more complex trajectories can result in less energy consumption when including the complexity of nonlinear losses than when a simpler model is considered.Those effects are stronger when the trajectory includes many acceleration and braking phases.

Transurethral resection of bladder tumor: A systematic review of simulator-based training courses and curricula

Objective:Transurethral resection of bladder tumor is one of the most common everyday urological procedures.This kind of surgery demands a set of skills that need training and experience.In this review,we aimed to investigate the current literature to find out if simulators,phantoms,and other training models could be used as a tool for teaching urologists.Methods:A systematic review was performed according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses statement and the recommendations of the European Association of Urology guidelines for conducting systematic reviews.Fifteen out of 932 studies met our inclusion criteria and are presented in the current review.Results:The UroTrainer(Karl Storz GmbH,Tuttlingen,Germany),a virtual reality training simulator,achieved positive feedback and an excellent face and construct validity by the participants.The inspection of bladder mucosa,blood loss,tumor resection,and procedural time was improved after the training,especially for inexperienced urologists and medical students.The construct validity of UroSim®(VirtaMed,Zurich,Switzerland)was established.SIMBLA simulator(Samed GmbH,Dresden,Germany)was found to be a realistic and useful tool by experts and urologists with intermediate experience.The test objective competency model based on SIMBLA simulator could be used for evaluating urologists.The porcine model of the Asian Urological Surgery Training and Education Group also received positive feedback by the participants that tried it.The Simulation and Technology Enhanced Learning Initiative Project had an extraordinary face and content validity,and 60%of participants would like to use the simulators in the future.The 5-day multimodal training curriculum“Boot Camp”in the United Kingdom achieved an increase of the level of confidence of the participants that lasted months after the project.Conclusion:Simulators and courses or curricula based on a simulator training could be a valuable learning tool for any surgeon,and there is no doubt that they should be a part of every urologist's technical education.

Aerodynamic Analysis and Optimization of Pantograph Streamline Fairing for High-Speed Trains

A pantograph serves as a vital device for the collection of electricity in trains.However,its aerodynamic resistance can limit the train’s running speed.As installing fairings around the pantograph is known to effectively reduce the resistance,in this study,different fairing lengths are considered and the related aerodynamic performances of pantograph are assessed.In particular,this is accomplished through numerical simulations based on the k-ωShear Stress Transport(SST)two-equation turbulence model.The results indicate that the fairing diminishes the direct impact of high-speed airflow on the pantograph,thereby reducing its aerodynamic resistance.However,it also induces interferences in the flow field around the train,leading to variations in the aerodynamic resistance and lift of train components.It is shown that a maximum reduction of 56.52%in pantograph aerodynamic resistance and a peak decrease of 3.38%in total train aerodynamic resistance can be achieved.

Beam Training and Tracking in mmWave Communication:A Survey

Communicating on millimeter wave(mmWave)bands is ushering in a new epoch of mobile communication which provides the availability of 10 Gbps high data rate transmission.However,mmWave links are easily prone to short transmission range communication because of the serious free space path loss and the blockage by obstacles.To overcome these challenges,highly directional beams are exploited to achieve robust links by hybrid beamforming.Accurately aligning the transmitter and receiver beams,i.e.beam training,is vitally important to high data rate transmission.However,it may cause huge overhead which has negative effects on initial access,handover,and tracking.Besides,the mobility patterns of users are complicated and dynamic,which may cause tracking error and large tracking latency.An efficient beam tracking method has a positive effect on sustaining robust links.This article provides an overview of the beam training and tracking technologies on mmWave bands and reveals the insights for future research in the 6th Generation(6G)mobile network.Especially,some open research problems are proposed to realize fast,accurate,and robust beam training and tracking.We hope that this survey provides guidelines for the researchers in the area of mmWave communications.

Dynamic analysis of axle box bearings on the high-speed train caused by wheel-rail excitation

To explore the impact of wheel-rail excitation on the dynamic performance of axle box bearings,a dynamic model of the high-speed train including axle box bearings is developed.Subsequently,the dynamic response characteristics of the axle box bearing are examined.The investigation focuses on the acceleration characteristics of bearing vibration under excitation of track irregularities and wheel flats.In addition,experiments on both normal and faulty bearings are conducted separately,and the correctness of the model and some conclusions are verified.According to the research,track irregularity is unfavorable for bearing fault detection based on resonance demodulation.Under the same speed conditions,the acceleration peak of bearing is inversely proportional to the length of the wheel flat and directly proportional to its depth.The paper will contribute to a deeper understanding of the dynamic performance of axle box bearings.

Enhancing clinical proficiency through laparoscopic training using Japanese origami

Objective This study aimed to investigate the clinical efficacy of laparoscopic training using origami,a traditional Japanese papercraft,using laparoscopic forceps to create origami cranes.Methods In this retrospective study,4 surgeons were randomly divided into 2 groups:The training group,consisting of surgeons 1 and 2,and the non-training group,consisting of surgeons 3 and 4.Over the course of a one-year study period,the training group regularly underwent laparoscopic surgery training with a dry box,wherein they folded a total of 1000 origami cranes using laparoscopic instruments.The non-training group periodically underwent common laparoscopic surgery training of techniques such as suturing and ligation.Each surgeon regularly performed the transabdominal preperitoneal approach for inguinal hernias.Each training was conducted concurrently with the surgeries.The procedure time(peritoneum detachment,mesh placement,and closure of the peritoneum),total operation time(time from peritoneum detachment to closure of the peritoneum),and surgical outcomes were examined.Results The training group showed greater improvement in the total operation time and more stable performance than the non-training group.Additionally,the time taken for peritoneum detachment was significantly shorter in the training group.Conclusion Laparoscopic training using origami has the potential to enhance laparoscopic surgical skills and improve surgical outcomes.

Influences of the Fresh Air Volume on the Removal of Cough-Released Droplets in a Passenger Car of a High-Speed Train Using CFD

The spread and removal of pollution sources,namely,cough-released droplets in three different areas(front,middle,and rear areas)of a fully-loaded passenger car in a high-speed train under different fresh air flow volume were studied using computational fluid dynamics(CFD)method.In addition,the structure of indoor flow fields was also analysed.The results show that the large eddies are more stable and flow faster in the air supply under Mode 2(fresh air volume:2200m3/h)compared to Mode 1(fresh air volume:1100m3/h).By analysing the spreading process of droplets sprayed at different locations in the passenger car and with different particle sizes,the removal trends for droplets are found to be similar under the two air supply modes.However,when increasing the fresh air flow volume,the droplets in the middle and front areas of the passenger car are removed faster.When the droplets had dispersed for 60s,Mode 2 exhibited a removal rate approximately 1%–3%higher than Mode 1 for small and medium-sized droplets with diameters of 10 and 50μm.While those in the rear area,the situation is reversed,with Mode 1 slightly surpassing Mode 2 by 1%–3%.For large droplets with a diameter of 100μm,both modes achieved a removal rate of over 96%in all three regions at the 60 s.The results can provide guidance for air supply modes of passenger cars of high-speed trains,thus suppressing the spread of virus-carrying droplets and reducing the risk of viral infection among passengers.

Effect of resistance training volume on body adiposity,metabolic risk,and inflammation in postmenopausal and older females:Systematic review and meta-analysis of randomized controlled trials

Purpose:This meta-analytical study aimed to explore the effects of resistance training(RT) volume on body adiposity,metabolic risk,and inflammation in postmenopausal and older females.Methods:A systematic search was performed for randomized controlled trials in PubMed,Scopus,Web of Science,and SciELO.Randomized controlled trials with postmenopausal and older females that compared RT effects on body adiposity,metabolic risk,and inflammation with a control group(CG) were included.Independent reviewers selected the studies,extracted the data,and performed the risk of bias and certainty of the evidence(Grading of Recommendations,Assessment,Development,and Evaluation(GRADE)) evaluations.Total body and abdominal adiposity,blood lipids,glucose,and C-reactive protein were included for meta-analysis.A random-effects model,standardized mean difference(Hedges’ g),and 95% confidence interval(95%CI) were used for meta-analysis.Results:Twenty randomized controlled trials(overall risk of bias:some concerns;GRADE:low to very low) with overweight/obese postmenopausal and older females were included.RT groups were divided into low-volume RT(LVRT,~44 sets/week) and high-volume RT(HVRT,~77 sets/week).Both RT groups presented improved body adiposity,metabolic risk,and inflammation when compared to CG.However,HVRT demonstrated higher effect sizes than LVRT for glucose(HVRT=-1.19;95%CI:-1.63 to-0.74;LVRT=-0.78;95%CI:-1.15 to-0.41) and C-reactive protein(HVRT=-1.00;95%CI:-1.32 to-0.67;LVRT=-0.34;95%CI,-0.63 to-0.04)) when compared to CG.Conclusion:Compared to CG,HVRT protocols elicit greater improvements in metabolic risk and inflammation outcomes than LVRT in overweight/obese postmenopausal and older females.

Optimized operation scheme of flash-memory-based neural network online training with ultra-high endurance

With the rapid development of machine learning,the demand for high-efficient computing becomes more and more urgent.To break the bottleneck of the traditional Von Neumann architecture,computing-in-memory(CIM)has attracted increasing attention in recent years.In this work,to provide a feasible CIM solution for the large-scale neural networks(NN)requiring continuous weight updating in online training,a flash-based computing-in-memory with high endurance(10^(9) cycles)and ultrafast programming speed is investigated.On the one hand,the proposed programming scheme of channel hot electron injection(CHEI)and hot hole injection(HHI)demonstrate high linearity,symmetric potentiation,and a depression process,which help to improve the training speed and accuracy.On the other hand,the low-damage programming scheme and memory window(MW)optimizations can suppress cell degradation effectively with improved computing accuracy.Even after 109 cycles,the leakage current(I_(off))of cells remains sub-10pA,ensuring the large-scale computing ability of memory.Further characterizations are done on read disturb to demonstrate its robust reliabilities.By processing CIFAR-10 tasks,it is evident that~90%accuracy can be achieved after 109 cycles in both ResNet50 and VGG16 NN.Our results suggest that flash-based CIM has great potential to overcome the limitations of traditional Von Neumann architectures and enable high-performance NN online training,which pave the way for further development of artificial intelligence(AI)accelerators.

Boosting Adversarial Training with Learnable Distribution

In recent years,various adversarial defense methods have been proposed to improve the robustness of deep neural networks.Adversarial training is one of the most potent methods to defend against adversarial attacks.However,the difference in the feature space between natural and adversarial examples hinders the accuracy and robustness of the model in adversarial training.This paper proposes a learnable distribution adversarial training method,aiming to construct the same distribution for training data utilizing the Gaussian mixture model.The distribution centroid is built to classify samples and constrain the distribution of the sample features.The natural and adversarial examples are pushed to the same distribution centroid to improve the accuracy and robustness of the model.The proposed method generates adversarial examples to close the distribution gap between the natural and adversarial examples through an attack algorithm explicitly designed for adversarial training.This algorithm gradually increases the accuracy and robustness of the model by scaling perturbation.Finally,the proposed method outputs the predicted labels and the distance between the sample and the distribution centroid.The distribution characteristics of the samples can be utilized to detect adversarial cases that can potentially evade the model defense.The effectiveness of the proposed method is demonstrated through comprehensive experiments.