Mixed neuroendocrine non-neuroendocrine neoplasms in gastroenteropancreatic tract

Mixed neuroendocrine non-neuroendocrine neoplasms(MiNENs)are a hetero-geneous group of malignant neoplasms that can settle in the gastroenteropan-creatic tract.They are composed of a neuroendocrine(NE)and a non-NE compo-nent in at least 30%of each tumour.The non-NE component can include different histological combinations of glandular,squamous,mucinous and sarcomatoid phenotypes,and one or both of the components can be low-or high grade malignant.Recent changes in the nomenclature of these neoplasms might lead to great deal of confusion,and the lack of specific clinical trials is the main reason why their management is difficult.The review aims to clarify the definition of MiNEN and analyze available evidence about their diagnosis and treatment options according to their location and extension through careful analysis of the available data.It would be important to reach a general consensus on their diagnosis in order to construct a classification that remains stable over time and facilitates the design of clinical trials that,due to their low incidence,will require long recruitment periods.

Mixed neuroendocrine and adenocarcinoma of gastrointestinal tract:A complex diagnosis and therapeutic challenge

In this editorial we comment on the manuscript describing a case of adenocarcinoma mixed with a neuroendocrine carcinoma of the gastroesophageal junction.Mixed neuroendocrine and non-neuroendocrine neoplasms of the gastrointestinal system are rare heterogeneous group of tumors characterized by a high malignant potential,rapid growth,and poor prognosis.Due to the rarity of these cancers,the standard therapy is poorly defined.The diagnosis of these tumors is based on combination of morphological features,immunohistochemical and neuroendocrine and epithelial cell markers.Both endocrine and epithelial cell components can act independently of each other and thus,careful grading of each component separately is required.These cancers are aggressive in nature and the potential of each component has paramount importance in the choice of treatment and response.Regardless of the organ of origin,these tumors portend poor prognosis with increased proportion of neuroendocrine component.Multidisciplinary services and strategies are required for the management of these mixed malignancies to provide the best oncological outcomes.The etiopathogenesis of these mixed tumors remains obscure but poses interesting question.We briefly discuss a few salient points in this editorial.

Dendroclimatological study of Sabina saltuaria and Abies faxoniana in the mixed forests of the Qionglai Mountains,eastern Tibetan Plateau

Tree-ring chronologies were developed for Sabina saltuaria and Abies faxoniana in mixed forests in the Qionglai Mountains of the eastern Tibetan Plateau.Climate-growth relationship analysis indicated that the two co-exist-ing species reponded similarly to climate factors,although S.saltuaria was more sensitive than A.faxoniana.The strong-est correlation was between S.saltuaria chronology and regional mean temperatures from June to November.Based on this relationship,a regional mean temperature from June to November for the period 1605-2016 was constructed.Reconstruction explained 37.3%of the temperature variance during th period 1961-2016.Six major warm periods and five major cold periods were identified.Spectral analysis detected significant interannual and multi-decadal cycles.Reconstruction also revealed the influence of the Atlantic Multi-decadal Oscillation,confirming its importance on climate change on the eastern Tibetan Plateau.

Amino-functionalized UiO-66-doped mixed matrix membranes with high permeation performance and fouling resistance

For the reduction of bovine serum proteins from wastewater,a novel mixed matrix membrane was prepared by functionalizing the substrate material polyaryletherketone(PAEK),followed by carboxyl groups(C-SPAEKS),and then adding amino-functionalized UiO-66-NH_(2)(Am-UiO-66-NH_(2)).Aminofunctionalization of UiO-66 was accomplished by melamine,followed by an amidation reaction to immobilize Am-UiO-66-NH_(2),which was immobilized on the surface of the membrane as well as in the pore channels,which enhanced the hydrophilicity of the membrane surface while increasing the negative potential of the membrane surface.This nanoparticle-loaded ultrafiltration membrane has good permeation performance,with a pure water flux of up to 482.3 L·m^(-2)·h^(-1) for C-SPAEKS/AmUiO-66-NH_(2) and a retention rate of up to 98.7%for bovine serum albumin(BSA)-contaminated solutions.Meanwhile,after several hydrophilic modifications,the flux recovery of BSA contaminants by this series of membranes increased from 56.2%to 80.55%of pure membranes.The results of ultra-filtration flux time tests performed at room temperature showed that the series of ultrafiltration membranes remained relatively stable over a test time of 300 min.Thus,the newly developed mixed matrix membrane showed potential for high efficiency and stability in wastewater treatment containing bovine serum proteins.

Synergistic anionic/zwitterionic mixed surfactant system with high emulsification efficiency for enhanced oil recovery in low permeability reservoirs

Emulsification is one of the important mechanisms of surfactant flooding. To improve oil recovery for low permeability reservoirs, a highly efficient emulsification oil flooding system consisting of anionic surfactant sodium alkyl glucosyl hydroxypropyl sulfonate(APGSHS) and zwitterionic surfactant octadecyl betaine(BS-18) is proposed. The performance of APGSHS/BS-18 mixed surfactant system was evaluated in terms of interfacial tension, emulsification capability, emulsion size and distribution, wettability alteration, temperature-resistance and salt-resistance. The emulsification speed was used to evaluate the emulsification ability of surfactant systems, and the results show that mixed surfactant systems can completely emulsify the crude oil into emulsions droplets even under low energy conditions. Meanwhile,the system exhibits good temperature and salt resistance. Finally, the best oil recovery of 25.45% is achieved for low permeability core by the mixed surfactant system with a total concentration of 0.3 wt%while the molar ratio of APGSHS:BS-18 is 4:6. The current study indicates that the anionic/zwitterionic mixed surfactant system can improve the oil flooding efficiency and is potential candidate for application in low permeability reservoirs.

A mixed-coordination electron trapping-enabled high-precision touch-sensitive screen for wearable devices

Touch-sensitive screens are crucial components of wearable devices.Materials such as reduced graphene oxide(rGO),carbon nanotubes(CNTs),and graphene offer promising solutions for flexible touch-sensitive screens.However,when stacked with flexible substrates to form multilayered capacitive touching sensors,these materials often suffer from substrate delamination in response to deformation;this is due to the materials having different Young’s modulus values.Delamination results in failure to offer accurate touch screen recognition.In this work,we demonstrate an induced charge-based mutual capacitive touching sensor capable of high-precision touch sensing.This is enabled by electron trapping and polarization effects related to mixed-coordinated bonding between copper nanoparticles and vertically grown graphene nanosheets.Here,we used an electron cyclotron resonance system to directly fabricate graphene-metal nanofilms(GMNFs)using carbon and copper,which are firmly adhered to flexible substrates.After being subjected to 3000 bending actions,we observed almost no change in touch sensitivity.The screen interaction system,which has a signal-to-noise ratio of 41.16 dB and resolution of 650 dpi,was tested using a handwritten Chinese character recognition trial and achieved an accuracy of 94.82%.Taken together,these results show the promise of touch-sensitive screens that use directly fabricated GMNFs for wearable devices.

Analysis of factors impacting postoperative pain and quality of life in patients with mixed hemorrhoids:A retrospective study

BACKGROUND Hemorrhoids are among the most common and frequently encountered chronic anorectal diseases in anorectal surgery.They are venous clusters formed by con-gestion,expansion,and flexion of the venous plexus in the lower part of the rec-tum.Mixed hemorrhoids bleed easily and recurrently,and this can result in severe anemia.Hence,they may have a negative effect on the health of the patient and surgical treatment is required.Milligan-Morgan hemorrhoidectomy has been widely used since 1937 for the treatment of grade III and IV hemorrhoids.How-ever,most patients experience different degrees of postoperative pain that may cause anxiety.with mixed hemorrhoids post-surgery.METHODS The clinical data of patients with mixed hemorrhoids who underwent Milligan-Morgan hemorrhoidectomy were collected retrospectively.The basic character-istics of the enrolled patients with mixed hemorrhoids were recorded,and based on the Goligher clinical grading system,the hemorrhoids were classified as grades III or IV.The endpoint of this study was the disappearance of pain in all patients.Quantitative data were presented as mean±SD,such as age,pain score,and QoL score.Student’s t-test was used to compare the groups.RESULTS A total of 164 patients were enrolled.The distribution of the visual analog scale pain scores of all patients at 3,7,14 and 28 d after surgery showed that post-surgery pain was significantly reduced with the passage of time.Fourteen days after the operation,the pain had completely disappeared in some patients.Twenty-eight days after the surgery,none of the patients experienced any pain.Comparing the World Health Or-ganization Quality of Life-BREF self-reporting questionnaire scores of patients between 14 and 28 d after surgery,we observed that the quality-of-life scores of the patients post-surgery had significantly improved.There were six items that were compared at 14-and 28-d post-surgery.The mean QoL score 28 d after surgery(4.79±0.46)was higher than that at 14 d post-surgery(3.79±0.57).The mean health condition score 28 d after surgery(4.80±0.41)was also higher than that at 14 d post-surgery(4.01±0.62).The mean physical health score 28 d after surgery(32.10±2.96)was significantly higher than that at 14 d post-surgery(23.41±2.85).The mean psychological health score 28 d after surgery(27.22±1.62)was significantly higher than that at 14 d post-surgery(21.37±1.70).The mean social relations score 28 d after surgery(12.21±1.59)was significantly higher than that at 14 d post-surgery(6.32±1.66).The mean surrounding environment score 28 d after surgery(37.13±2.88)was significantly higher than that at 14 d post-surgery(28.42±2.86).The differences in quality-of-life scores at day 14 and day 28 post-surgery were ob-served to be statistically significant(P<0.001).CONCLUSION Milligan-Morgan hemorrhoidectomy can significantly improve the postoperative QoL of patients.Age,sex,and the number of surgical resections were important factors influencing Milligan-Morgan hemorrhoidectomy.

Soil properties and not plant factors affect both abundant and rare microbial taxa after thinning in a mixed stand of Cunninghamia lanceolata and Sassafras tzumu

Thinning is an effective management step for sustainable forest development,yet less attention is paid to the restoration of soil microbiota after thinning.In this study,both abundant and rare soil microbial communities(i.e.,bacterial,fungal),were evaluated under various thinning treatments in a mixed stand of Cunninghamia lanceolata and Sassafras tzumu using Mi Seq sequencing.Thinning did not significantly change either abundant or rare bacterial and fungal community composition,but affected their alpha diversity.The Shannon–Wiener indexes of rare fungal taxa under medium thinning were significantly lower than in the light thinning(P<0.05 level).Xanthobacteraceae dominated the abundant bacterial taxa,and Saitozyma and Mortierlla the abundant fungal taxa.The most common rare bacterial taxa varied;there was no prevalent rare fungal taxa under different thinnings.In addition,soil available nitrogen,total phosphorus,and p H had significant effects on rare bacterial taxa.Nutrients,especially available phosphorus,but not nitrogen,affected abundant and rare soil fungi.The results indicate that soil properties rather than plant factors affect abundant and rare microbial communities in soils of mixed stands.Thinning,through mediating soil properties,influences both abundant and rare bacterial and fungal communities in the mixed C.lanceolata and S.tzumu stand.

Highly dispersed MgInCe-mixed metal oxides catalyzed direct carbonylation of glycerol and CO_(2)into glycerol carbonate

Glycerol carbonate,an important glycerol value-added product,has been widely used as an active intermediate and inert solvent in the synthesis of cosmetics,detergents,chemical intermediates,polymers,and so on.The direct carbonylation from glycerol with CO_(2)is considered a promising route,but still tough work due to the thermodynamic stability and the kinetic inertness of CO_(2).In this work,highlyselective direct carbonylation of glycerol and CO_(2)into glycerol carbonate has been achieved over highly dispersed MgInCe-mixed metal oxides(MgInCe-MMO),which were prepared through the topological transformation derived from the MgInCe-layered double hydroxides(MgInCe-LDHs).By precisely modulating the surface basic-acidic properties and the oxygen vacancies,an efficient carbonylation of glycerol with CO_(2)has been achieved with a selectivity of up to>99%to glycerol carbonate.Deep investigation into the synergistic catalysis of base-acid sites and oxygen vacancies has been clarified.

Dual-functional poly(2-methyl-2-oxazoline)/poly(2-(dimethylamine)ethyl methacrylate)mixed brushes with switchable protein adsorption and antibacterial properties

Herein,binary mixed brushes consisting of poly(2-methyl-2-oxazoline)(PMOXA)and poly(2-(dimethylamine)ethyl methacrylate)(PDMAEMA)with different chain lengths were fabricated by successive grafting of NH_(2)-terminated PMOXA and SH-terminated PDMAEMA onto polydopamine-anchored substrates.The mixed-brush coating was characterized by variable-angle spectroscopic ellipsometry,X-ray photoelectron spectroscopy,Fourier transform infrared spectroscopy,zeta potential measurements,water contact angle,and atomic force microscopy.The mixed brushes showed tunable surface charge,wettability,and surface roughness,depending on the degree of PDMAEMA swelling under varying pH and ionic strength(Ⅰ).Then the adsorption behaviors of pepsin,bovine serum albumin(BSA),γ-globulin,and lysozyme,four very different proteins with regard to isoelectric point,on the mixed brushes coating were studied by using fluorescence microscopy and surface plasmon resonance.When the chain length of PDMAEMA was about twice as long as PMOXA,the mixed brushes not only had high adsorption capacity for pepsin,BSA,and y-globulin but also had a desorption efficiency of 86.9%,87.1%,and 93.5%,respectively.It is explained that electrostatic attraction between the protonated PDMAEMA and positively charged acidic proteins(pepsin and BSA,whose isoelectric points were below the pK_(a) of PDMAEMA)would drive the intensive adsorption(at pH 3,I=10^(-3)mol·L^(-1)for pepsin,and pH 5,I=10^(-5)mol·L^(-1)for BSA),while desorption was dominated by the hydrophilic PMOXA when PDMAEMA was shrinking(at pH 7,I=10^(-1)mol·L^(-1)for pepsin,and pH 9,I=10^(-1)mol·L^(-1)for BSA).Furthermore,the isoelectric precipitation led to the adsorption of neutral protein(γ-globulin,whose isoelectric point was near the pK_a of PDMAEMA)at pH 7,I=10^(-5)mol·L^(-1),while electrostatic repulsion and antifouling PMOXA triggered the desorption of y-globulin at pH 3,I-10^(-1)mol·L^(-1).However,alkaline protein(lysozyme,whose isoelectric point was higher than the pK_(a) of PDMAEMA)exhibited slight adsorption on PMOXA/PDMAEMA mixed brushes under test conditions,regardless of whether PMOXA or PDMAEMA occupied the outermost layer.The antibacterial property of the mixed brushes against Escherichia coli was investigated.PMOXA/PDMAEMA mixed brushes showed significant bactericidal activity at pH 3,I=10^(-3)mol·L^(-1),while the rinse of pH 9,I=10^(-1)mol·L^(-1)solution could remove most of the residual bacteria.This work not only enables controlled adsorption of proteins with different isoelectric points but also ensures that the surface of the coating is minimized from bacterial contamination.

3D printing of poly(ethyleneimine)-functionalized Mg-Al mixed metal oxide monoliths for direct air capture of CO_(2)

Direct air capture(DAC)of CO_(2)plays an indispensable role in achieving carbon-neutral goals as one of the key negative emission technologies.Since large air flows are required to capture the ultradilute CO_(2)from the air,lab-synthesized adsorbents in powder form may cause unacceptable gas pressure drops and poor heat and mass transfer efficiencies.A structured adsorbent is essential for the implementation of gas-solid contactors for cost-and energy-efficient DAC systems.In this study,efficient adsorbent poly(ethyleneimine)(PEI)-functionalized Mg-Al-CO_(3)layered double hydroxide(LDH)-derived mixed metal oxides(MMOs)are three-dimensional(3D)printed into monoliths for the first time with more than 90%adsorbent loadings.The printing process has been optimized by initially printing the LDH powder into monoliths followed by calcination into MMO monoliths.This structure exhibits a 32.7%higher specific surface area and a 46.1%higher pore volume,as compared to the direct printing of the MMO powder into a monolith.After impregnation of PEI,the monolith demonstrates a large adsorption capacity(1.82 mmol/g)and fast kinetics(0.7 mmol/g/h)using a CO_(2)feed gas at 400 ppm at 25℃,one of the highest values among the shaped DAC adsorbents.Smearing of the amino-polymers during the post-printing process affects the diffusion of CO_(2),resulting in slower adsorption kinetics of pre-impregnation monoliths compared to post-impregnation monoliths.The optimal PEI/MeOH ratio for the post-impregnation solution prevents pores clogging that would affect both adsorption capacity and kinetics.

Investigation on photonic crystal nanobeam cavity based on mixed diamond–circular holes

A photonic crystal nanobeam cavity(M-PCNC)with a structure incorporating a mixture of diamond-shaped and circular air holes is pro-posed.The performance of the cavity is simulated and studied theoretically.Using thefinite-difference time-domain method,the parameters of the M-PCNC,including cavity thickness and width,lattice constant,and radii and numbers of holes,are optimized,with the quality factor Q and mode volume Vm as performance indicators.Mutual modulation of the lattice constant and hole radius enable the proposed M-PCNC to realize outstanding performance.The optimized cavity possesses a high quality factor Q 1.45105 and an ultra-small mode=×volume Vm 0.01(λ/n)[Zeng et al.,Opt Lett 2023:48;3981–3984]in the telecommunications wavelength range.Light can be progres-=sively squeezed in both the propagation direction and the perpendicular in-plane direction by a series of interlocked anti-slots and slots in the diamond-shaped hole structure.Thereby,the energy can be confined within a small mode volume to achieve an ultra-high Q/Vm ratio.

On the Application of Mixed Models of Probability and Convex Set for Time-Variant Reliability Analysis

In time-variant reliability problems,there are a lot of uncertain variables from different sources.Therefore,it is important to consider these uncertainties in engineering.In addition,time-variant reliability problems typically involve a complexmultilevel nested optimization problem,which can result in an enormous amount of computation.To this end,this paper studies the time-variant reliability evaluation of structures with stochastic and bounded uncertainties using a mixed probability and convex set model.In this method,the stochastic process of a limit-state function with mixed uncertain parameters is first discretized and then converted into a timeindependent reliability problem.Further,to solve the double nested optimization problem in hybrid reliability calculation,an efficient iterative scheme is designed in standard uncertainty space to determine the most probable point(MPP).The limit state function is linearized at these points,and an innovative random variable is defined to solve the equivalent static reliability analysis model.The effectiveness of the proposed method is verified by two benchmark numerical examples and a practical engineering problem.

Long-term changes in radial growth of seven tree species in the mixed broadleaf-Korean pine forest in Northeast China:Are deciduous trees favored by climate change?

The role of the temperate mixed broadleaf-Korean pine forest(BKF)in global biogeochemical cycles will depend on how the tree species community responds to climate;however,species-specific responses and vulner-abilities of common trees in BKF to extreme climates are poorly understood.Here we used dendrochronological meth-ods to assess radial growth of seven main tree species(Pinus koraiensis,Picea jezoensis,Abies nephrolepis,Fraxinus mandshurica,Phellodendron amurense,Quercus mongolica,and Ulmus davidiana)in an old-growth BKF in response to climate changes in the Xiaoxing’an Mountains and to improve predictions of changes in the tree species compo-sition.Temperature in most months and winter precipita-tion significantly negatively affected growth of P.jezoensis and A.nephrolepis,but positively impacted growth of P.koraiensis and the broadleaf species,especially F.mandshu-rica and U.davidiana.Precipitation and relative humidity in June significantly positively impacted the growth of most tree species.The positive effect of the temperature during the previous non-growing season(PNG)on growth of F.mandshurica and Q.mongolica strengthened significantly with rapid warming around 1981,while the impact of PNG temperature on the growth of P.jezoensis and A.nephrolepis changed from significantly negative to weakly negative or positive at this time.The negative response of radial growth of P.jezoensis and A.nephrolepis to precipitation during the growing season gradually weakened,and the negative response to PNG precipitation was enhanced.Among the studied species,P.koraiensis was the most resistant to drought,and U.davidiana recovered the best after extreme drought.Ulmus davidiana,P.jezoensis and A.nephrolepis were more resistant to extreme cold than the other species.Climate warming generally exacerbated the opposite growth patterns of conifer(decline)and broadleaf(increase)spe-cies.Deciduous broadleaf tree species in the old-growth BKF probably will gradually become dominant as warming continues.Species-specific growth-climate relationships should be considered in future models of biogeochemical cycles and in forestry management practices.

Mixed D-vine copula-based conditional quantile model for stochastic monthly streamflow simulation

Copula functions have been widely used in stochastic simulation and prediction of streamflow.However,existing models are usually limited to single two-dimensional or three-dimensional copulas with the same bivariate block for all months.To address this limitation,this study developed a mixed D-vine copula-based conditional quantile model that can capture temporal correlations.This model can generate streamflow by selecting different historical streamflow variables as the conditions for different months and by exploiting the conditional quantile functions of streamflows in different months with mixed D-vine copulas.The up-to-down sequential method,which couples the maximum weight approach with the Akaike information criteria and the maximum likelihood approach,was used to determine the structures of multivariate Dvine copulas.The developed model was used in a case study to synthesize the monthly streamflow at the Tangnaihai hydrological station,the inflow control station of the Longyangxia Reservoir in the Yellow River Basin.The results showed that the developed model outperformed the commonly used bivariate copula model in terms of the performance in simulating the seasonality and interannual variability of streamflow.This model provides useful information for water-related natural hazard risk assessment and integrated water resources management and utilization.

Recent advances in flat sheet mixed matrix membrane modified by Mg-based layered double hydroxides(LDHs)for salt and organic compound separations

Magnesium(Mg)is a widely used and attractive metal,known for its unique physical and chemical properties,and it has been employed in the manufacture of many practical materials.Layered Double Hydroxides(LDHs),particularly Mg-based LDHs,rank among the most prevalent two-dimensional materials utilized in separation processes,which include adsorption,extraction,and membrane technology.The high popularity of Mg-based LDHs in separation applications can be attributed to their properties,such as excellent hydrophilicity,high surface area,ion exchangeability,and adjustable interlayer space.Currently,polymer membranes play a pivotal role in semi-industrial and industrial separation processes.Consequently,the development of polymer membranes and the mitigation of their limitations have emerged as compelling topics for researchers.Several methods exist to enhance the separation performance and anti-fouling properties of polymer membranes.Among these,incorporating additives into the membrane polymer matrix stands out as a cost-effective,straightforward,readily available,and efficient approach.The use of Mg-based LDHs,either in combination with other materials or as a standalone additive in the polymer membrane matrix,represents a promising strategy to bolster the separation and anti-fouling efficacy of flat sheet mixed matrix polymer membranes.This review highlights Mg-based LDHs as high-potential additives designed to refine flat sheet mixed matrix polymer membranes for applications in wastewater treatment and brackish water desalination.

Stability Analysis and Performance Evaluation of Additive Mixed-Precision Runge-Kutta Methods

Additive Runge-Kutta methods designed for preserving highly accurate solutions in mixed-precision computation were previously proposed and analyzed.These specially designed methods use reduced precision for the implicit computations and full precision for the explicit computations.In this work,we analyze the stability properties of these methods and their sensitivity to the low-precision rounding errors,and demonstrate their performance in terms of accuracy and efficiency.We develop codes in FORTRAN and Julia to solve nonlinear systems of ODEs and PDEs using the mixed-precision additive Runge-Kutta(MP-ARK)methods.The convergence,accuracy,and runtime of these methods are explored.We show that for a given level of accuracy,suitably chosen MP-ARK methods may provide significant reductions in runtime.

Influences of oscillation on the physical stability and explosion characteristics of solid-liquid mixed fuel

The stratification phenomenon resulting from differences in the physical properties of solid-liquid components seriously affect the final combustion and explosion characteristics of mixed fuel under the action of oscillation.The effects of oscillation on the physical stability of mixed fuel with two solid-liquid ratios and three liquid component distribution ratios have been investigated using a self-designed experimental system at oscillation frequencies of 60-300 r/min.The explosion characteristics of mixed fuel before and after oscillation are gained from a 20 L spherical explosion container system.When the mass ratio of liquid components is controlled at 66.9%,64.7%,62.6%the final explosion characteristics are stable,with a maximum difference of only 0.71%.The volume of liquid fuel precipitation increases with increasing oscillation frequency when the mass ratio of liquid components reaches 71.7%,69.6%,67.7%.The fuel explosion overpressure after oscillation decreases with increasing liquid precipitation volume,and the repeatability is poor,with a maximum standard deviation of 82.736,which is much higher than the ratio without stratification.Properly controlling the mass ratio of liquid components of the mixed fuel can effectively combat the impact of oscillation on the physical state and maintain the stability of the final explosion characteristics.

What induced the trend shift of mixed-layer depths in the Antarctic Circumpolar Current region in the mid-1980s?

An obvious trend shift in the annual mean and winter mixed layer depth(MLD)in the Antarctic Circumpolar Current(ACC)region was detected during the 1960–2021 period.Shallowing trends stopped in mid-1980s,followed by a period of weak trends.The MLD deepening trend difference between the two periods were mainly distributed in the western areas in the Drake Passage,the areas north to Victoria Land and Wilkes Land,and the central parts of the South Indian sector.The newly formed ocean current shear due to the meridional shift of the ACC flow axis between the two periods is the dominant driver for the MLD trends shift distributed in the western areas in the Drake Passage and the central parts of the South Indian sector.The saltier trends in the regions north to Victoria Land and Wilkes Land could be responsible for the strengthening mixing processes in this region.

Reinforcement Learning Based Quantization Strategy Optimal Assignment Algorithm for Mixed Precision

The quantization algorithm compresses the original network by reducing the numerical bit width of the model,which improves the computation speed. Because different layers have different redundancy and sensitivity to databit width. Reducing the data bit width will result in a loss of accuracy. Therefore, it is difficult to determinethe optimal bit width for different parts of the network with guaranteed accuracy. Mixed precision quantizationcan effectively reduce the amount of computation while keeping the model accuracy basically unchanged. In thispaper, a hardware-aware mixed precision quantization strategy optimal assignment algorithm adapted to low bitwidth is proposed, and reinforcement learning is used to automatically predict the mixed precision that meets theconstraints of hardware resources. In the state-space design, the standard deviation of weights is used to measurethe distribution difference of data, the execution speed feedback of simulated neural network accelerator inferenceis used as the environment to limit the action space of the agent, and the accuracy of the quantization model afterretraining is used as the reward function to guide the agent to carry out deep reinforcement learning training. Theexperimental results show that the proposed method obtains a suitable model layer-by-layer quantization strategyunder the condition that the computational resources are satisfied, and themodel accuracy is effectively improved.The proposed method has strong intelligence and certain universality and has strong application potential in thefield of mixed precision quantization and embedded neural network model deployment.