3D-SMGE: a pipeline for scaffold-based molecular generation and evaluation.

In the process of drug discovery, one of the key problems is how to improve the biological activity and ADMET properties starting from a specific structure, which is also called structural optimization. Based on a starting scaffold, the use of deep generative model to generate molecules with desired drug-like properties will provide a powerful tool to accelerate the structural optimization process. However, the existing generative models remain challenging in extracting molecular features efficiently in 3D space to generate drug-like 3D molecules. Moreover, most of the existing ADMET prediction models made predictions of different properties through a single model, which can result in reduced prediction accuracy on some datasets. To effectively generate molecules from a specific scaffold and provide basis for the structural optimization, the 3D-SMGE (3-Dimensional Scaffold-based Molecular Generation and Evaluation) work consisting of molecular generation and prediction of ADMET properties is presented. For the molecular generation, we proposed 3D-SMG, a novel deep generative model for the end-to-end design of 3D molecules. In the 3D-SMG model, we designed the cross-aggregated continuous-filter convolution (ca-cfconv), which is used to achieve efficient and low-cost 3D spatial feature extraction while ensuring the invariance of atomic space rotation. 3D-SMG was proved to generate valid, unique and novel molecules with high drug-likeness. Besides, the proposed data-adaptive multi-model ADMET prediction method outperformed or maintained the best evaluation metrics on 24 out of 27 ADMET benchmark datasets. 3D-SMGE is anticipated to emerge as a powerful tool for hit-to-lead structural optimizations and accelerate the drug discovery process.

Design, synthesis, and mechanism of action of novel µ-conotoxin KIIIA analogues for inhibition of the voltage-gated sodium channel Nav1.7.

µ-Conotoxin KIIIA, a selective blocker of sodium channels, has strong inhibitory activity against several Nav isoforms, including Nav1.7, and has potent analgesic effects, but it contains three pairs of disulfide bonds, making structural modification difficult and synthesis complex. To circumvent these difficulties, we designed and synthesized three KIIIA analogues with one disulfide bond deleted. The most active analogue, KIIIA-1, was further analyzed, and its binding pattern to hNav1.7 was determined by molecular dynamics simulations. Guided by the molecular dynamics computational model, we designed and tested 32 second-generation and 6 third-generation analogues of KIIIA-1 on hNav1.7 expressed in HEK293 cells. Several analogues showed significantly improved inhibitory activity on hNav1.7, and the most potent peptide, 37, was approximately 4-fold more potent than the KIIIA Isomer I and 8-fold more potent than the wildtype (WT) KIIIA in inhibiting hNav1.7 current. Intraperitoneally injected 37 exhibited potent in vivo analgesic activity in a formalin-induced inflammatory pain model, with activity reaching ∼350-fold of the positive control drug morphine. Overall, peptide 37 has a simplified disulfide-bond framework and exhibits potent in vivo analgesic effects and has promising potential for development as a pain therapy in the future.

Molybdenum-Doped Cobalt-Free Cathode Realizing the Electrochemical Stability by Enhanced Covalent Bonding.

The doping strategy effectively enhances the capacity and cycling stability of cobalt-free nickel-rich cathodes. Understanding the intrinsic contributions of dopants is of great importance to optimize the performances of cathodes. This study investigates the correlation between the structure modification and their performances of Mo-doped LiNi0.8Mn0.2O2 (NM82) cathode. The role of doped Mo's valence state has been proved functional in both lattice structural modification and electronic state adjustment. Although the high-valence of Mo at the cathode surface inevitably reduces Ni valence for electronic neutrality and thus causes ion mixing, the original Mo valence will influence its diffusion depth. Structural analyses reveal Mo doping leads to a mixed layer on the surface, where high-valence Mo forms a slender cation mixing layer, enhancing structural stability and Li-ion transport. In addition, it is found that the high-valence dopant of Mo6+ ions partially occupies the unfilled 4d orbitals, which may strengthen the Mo─O bond through increased covalency and therefore reduce the oxygen mobility. This results in an impressive capacity retention (90.0% after 200 cycles) for Mo-NM82 cathodes with a high Mo valence state. These findings underscore the valence effect of doping on layered oxide cathode performance, offering guidance for next-generation cathode development.

Synthesis of 2-phenylnaphthalenoid amide derivatives and their topoisomerase IIα inhibitory and antiproliferative activities.

Topoisomerases are highly associated with cell proliferation, becoming an important target for the development of antitumor drugs. 2-Phenylnaphthalenoids (2PNs) have been identified as human DNA topoisomerase IIα (TopoIIα) inhibitors. In this study, based on the 2PN scaffold, 20 amide derivatives (J1-J10, K1-K10) were synthesized. Among them, K10 showed high TopoIIα inhibitory activity and stronger antiproliferation activity against HepG-2 and MDA-MB-231 cells (IC50 0.33 and 0.63 µM, respectively) than the positive control VP-16 (IC50 9.19 and 10.86 µM) and the lead F2 (IC50 0.64 and 1.51 µM). Meanwhile, K10 could also inhibit migration and promote apoptosis of HepG-2 and MDA-MB-231 cells. Therefore, K10 can be developed into a potent TopoIIα inhibitor as an antitumor agent. The structure-activity relationship was also discussed.

Suppression of Railway Catenary Galloping Based on Structural Parameters' Optimization.

Railway catenary galloping, induced by aerodynamic instability, poses a significant threat by disrupting the electric current connection through sliding contact with the contact wire. This disruption leads to prolonged rail service interruptions and damage to the catenary's suspension components. This paper delves into the exploration of optimizing the catenary system's structure to alleviate galloping responses, addressing crucial parameters such as span length, stagger dropper distribution, and tension levels. Employing a finite element model, the study conducts simulations to analyze the dynamic response of catenary galloping, manipulating structural parameters within specified ranges. To ensure accurate and comprehensive exploration, the Sobol sequence is utilized to generate low-discrepancy, quasi-random, and super-uniform distribution sequences for the high-dimensional parameter inputs. Subsequent to the simulation phase, a genetic algorithm based on neural networks is employed to identify optimal parameter settings for suppressing catenary galloping, taking into account various constraints. The results gleaned from this investigation affirm that adjusting structural parameters can effectively diminish the galloping amplitude of the railway catenary. The most impactful strategy involves augmenting tension and reducing span length. Moreover, even when tension and span length are fixed, adjusting other parameters demonstrates efficacy in reducing galloping amplitudes. The adjustment of messenger-wire tension, dropper distribution, and stagger can achieve a 22.69% reduction in the maximum vertical galloping amplitude. Notably, maintaining a moderate stagger value and a short steady arm-dropper distance is recommended to achieve the minimum galloping amplitude. This research contributes valuable insights into the optimization of railway catenary systems, offering practical solutions to mitigate galloping-related challenges and enhance overall system reliability.

A review of flow field characteristics in submerged hollow fiber membrane bioreactor: Micro-interface, module and reactor.

As an important part of the membrane field, hollow fiber membranes (HFM) have been widely concerned by scholars. HFM fouling in the industrial application results in a reduction in its lifespan and an increase in cost. In recent years, various explorations on the HFM fouling control strategies have been carried out. In the current work, we critically review the influence of flow field characteristics in HFM-based bioreactor on membrane fouling control. The flow field characteristics mainly refer to the spatial and temporal variation of the related physical parameters. In the HFM field, the physical parameter mainly refers to the variation characteristics of the shear force, flow velocity and turbulence caused by hydraulics. The factors affecting the flow field characteristics will be discussed from three levels: the micro-flow field near the interface of membrane (micro-interface), the flow field around the membrane module and the reactor design related to flow field, which involves surface morphology, crossflow, aeration, fiber packing density, membrane vibration, structural design and other related parameters. The study of flow field characteristics and influencing factors in the HFM separation process will help to improve the performance of HFM in full-scale water treatment plants.

Redox-active ligands as a challenge for electronic structure methods.

To improve the catalytic activity of 3d transition metal catalysts, redox-active ligands are a promising tool. These ligands influence the oxidation state of the metal center as well as the ground spin-state and make the experimental determination of both properties challenging. Therefore, first-principles calculations, in particular employing density functional theory with a proper choice of exchange-correlation (xc) functional, are crucial. Common xc functionals were tested on a simple class of metal complexes: homoleptic, octahedral tris(diimine) iron(II) complexes. The spin-state energy splittings for most of these complexes showed the expected linear dependence on the amount of exact exchange included in the xc functionals. Even though varying redox-activity affects the electronic structure of the complexes considerably, the sensitivity of the spin-state energetics to the exact exchange admixture is surprisingly small. For iron(II) complexes with highly redox-active ligands and for a broad range of ligands in the reduced tris(diimine) iron(I) complexes, self-consistent field convergence to local minima was observed, which differ from the global minimum in the redox state of the ligand. This may also result in convergence to a molecular structure that corresponds to an energetically higher-lying local minimum. One criterion to detect such behavior is a change in the sign of the slope for the dependence of the spin-state energy splittings on the amount of exact exchange. We discuss possible protocols for dealing with such artifacts in cases in which a large number of calculations makes checking by hand unfeasible.

Research Progress of Organic Field-Effect Transistor Based Chemical Sensors.

Due to their high sensitivity and selectivity, chemical sensors have gained significant attention in various fields, including drug security, environmental testing, food safety, and biological medicine. Among them, organic field-effect transistor (OFET) based chemical sensors have emerged as a promising alternative to traditional sensors, exhibiting several advantages such as multi-parameter detection, room temperature operation, miniaturization, flexibility, and portability. This review paper presents recent research progress on OFET-based chemical sensors, highlighting the enhancement of sensor performance, including sensitivity, selectivity, stability, etc. The main improvement programs are improving the internal and external structures of the device, as well as the organic semiconductor layer and dielectric structure. Finally, an outlook on the prospects and challenges of OFET-based chemical sensors is presented.

Discovery of 4-oxo-4,5-dihydropyrazolo[1,5-a]quinoxaline-7-carboxamide derivatives as PI3Kα inhibitors via virtual screening and docking-based structure optimization.

Compound 1 with pyrazolo[1,5-a]quinoxalin-4(5H)-one scaffold was identified as a PI3Kα inhibitor hit via virtual screening strategy. Additional similarity search and molecular docking based structural modification yielded a novel series of pyrazolo[1,5-a]quinoxalin-4(5H)-one derivatives. The most potent compound 49b exhibited remarkably improved PI3Kα inhibitory activity with IC50 value of 0.24 µM and moderate to good isoform selectivity over other class I PI3K isoforms. In addition, 49b significantly inhibited the proliferation of Kasumi-1 and T47D cells with IC50 value of 1.64 and 1.82 µM, respectively. Further PK study demonstrated that it has favorable pharmacokinetic profiles (AUC0-t = 3294.05 ng·h/mL at 5.0 mg/kg PO, F = 91.8%). All these data indicated that compound 49b was a promising PI3Kα inhibitor with beneficial drug-like properties and merited further development.

Structure optimization of an F-indole-chalcone (FC116) on 4-methoxyphenyl group and therapeutic potential against colorectal cancers with low cytotoxicity.

Advanced metastatic colorectal cancers (CRCs) are regarded as a challenge in clinical cancer therapy. Our previous studies have demonstrated that a representative fluoro-substituted indole-chalcone (FC116), was obtained to display highly potent activity against CRC using multiple in vitro and in vivo mouse models by targeting microtubules. However, several problems, such as low dose tolerance and highly toxic to the brain and colon, low solubility unsuitable for intravenous (i.v.) administration, are still existed and limit further development. Herein, we developed two series of FC116 derivatives on the 4-methoxyphenyl group by a structure-based design strategy. Among them, FC11619 with an amino terminus maintained the in vitro cytotoxicity against HCT-116 CRC in a low nanomolar range. This compound could induce G2/M phase arrest via regulating cyclin B1 expression, produce excess reactive oxygen species (ROS), and target tubulin in CRC cells. In vivo, FC11619 significantly suppressed tumor growth, achieving 65.3 and 73.4 % at doses of 5 and 10 mg/kg/d (i.v., 21 d), which were much better than 54.1% of Taxol at 7 mg/kg. In addition, this compound showed better in vivo tolerance compared to that of FC116 (only 3 mg/kg tolerance, intraperitoneal, i.p.), and no major organ-related toxicity, especially no apparent degenerated neurons, intestinal obstruction in clinical Taxol standard therapy. Taken together, the 4-amino-substitutedphenyl indole-chalcones represent lead compounds as chemotherapy of CRC for further drug development in this field.

Structure optimization, synthesis, and biological evaluation of 6-(2-amino-1H-benzo[d]imidazole-6-yl)-quinazolin-4(3H)-one derivatives as potential multi-targeted anticancer agents via Aurora A/ PI3K/BRD4 inhibition.

Aurora A (Aurora kinase A), a critical regulator of cell mitosis, is frequently overexpressed in many malignant cancers, and has been considered as a promising drug target for cancer therapy. Likewise, Phosphatidylinositol 3-kinase alpha (PI3Kα) is also regarded as one of the most important targets in cancer therapy by mediating the cell growth and angiogenesis of various human cancers. In addition, Bromodomain-containing protein 4 (BRD4) modulates oncogene expressions of Myc, Aurora kinase and various RTKs. Recently, accumulating evidences indicated that hyperactivated or abnormally expressed Aurora A, PI3Kα or BRD4 are closely associated with drug resistance and poor prognosis of non-small cell lung cancer (NSCLC). Hence, simultaneous inhibition of Aurora A, PI3Kα, and BRD4 is expected to be a new strategy for NSCLC therapy. In this study, we performed further structure optimization of 6-(2-amino-1H-benzo[d]imidazole-6-yl)-quinazolin-4(3H) -one based on previous study to obtain a series of derivatives for discovering potential Aurora A, PI3Kα and BRD4 multi-targeted inhibitors. MTT assay showed that most of the newly synthesized compounds exhibited an evident anticancer activity against the NSCLC cells. Among them, the IC50 values of the most potent compound 9a were 0.83, 0.26 and 1.02 µM against A549, HCC827 and H1975 cells, respectively. In addition, 9a markedly inhibited the Aurora A and PI3Kα kinase activities with IC50 values of 10.19 nM and 13.12 nM. Compound 9a induced G2/M phase arrests and apoptosis of HCC827 cells by simultaneous inhibition of Aurora A/PI3K/ BRD4 signaling pathways. Collectively, our studies suggested that 9a might be a potential multi-targeted inhibitor for NSCLC therapy.

Discovery of 5-alkyl-2-pyrazol-oxazolidin-4-one derivatives as novel hepatitis B virus capsid assembly modulators.

The "magic methyl effect" strategy was used to design a series of 5-alkyl-2-pyrazol-oxazolidin-4-one derivatives as novel hepatitis B virus (HBV) capsid assembly modulators. Most of these compounds exhibited potent HBV inhibitory activities with low cytotoxicities in HepG2.2.15 cells. The most promising compounds 9d and 10b had single-digit nanomolar IC50 values with a high selectivity index. Compared with the lead compound (3.0%), they caused 15% and 18% decreases in HBe antigen secretion at 1.0 µM, respectively. In addition, compounds 9d and 10b possessed good pharmacokinetic profiles with oral bioavailability values of 56.1% and 48.9%, respectively. These results indicated that the two compounds were potential therapeutic agents for HBV infection.

Design of OMC-Sagnac Loop Using PDMS and Different Package Structures to Improve Sensing Performance and Optimize the Ill-Conditioned Matrix.

In the process of ocean exploration, highly accurate and sensitive measurements of seawater temperature and pressure significantly impact the study of seawater's physical, chemical, and biological processes. In this paper, three different package structures, V-shape, square-shape, and semicircle-shape, are designed and fabricated, and an optical microfiber coupler combined Sagnac loop (OMCSL) is encapsulated in these structures with polydimethylsiloxane (PDMS). Then, the temperature and pressure response characteristics of the OMCSL, under different package structures, are analyzed by simulation and experiment. The experimental results show that structural change hardly affects temperature sensitivity, and square-shape has the highest pressure sensitivity. In addition, with an input error of 1% F.S., temperature and pressure errors were calculated, which shows that a semicircle-shape structure can increase the angle between lines in the sensitivity matrix method (SMM), and reduce the effect of the input error, thus optimizing the ill-conditioned matrix. Finally, this paper shows that using the machine learning method (MLM) effectively improves demodulation accuracy. In conclusion, this paper proposes to optimize the ill-conditioned matrix problem in SMM demodulation by improving sensitivity with structural optimization, which essentially explains the cause of the large errors for multiparameter cross-sensitivity. In addition, this paper proposes to use the MLM to solve the problem of large errors in the SMM, which provides a new method to solve the problem of the ill-conditioned matrix in SMM demodulation. These have practical implications for engineering an all-optical sensor that can be used for detection in the ocean environment.

The role of digitalization on green economic growth: Does industrial structure optimization and green innovation matter?

The digital economy has demonstrated strong resilience and great potential, under the interwoven influence of the global pandemic and severe environmental concerns across the world. Therefore, there is a need to focus on the value of green economic growth in the digital economy. This paper constructs an evaluation index system and adopts the SEEA (System of Environmental and Economic Accounting) method to measure the digitalization level (Digi) and green economy growth level (GEG) of China. The internal mechanism and linear relationship between digitalization and green economy growth are examined based on the panel data from 2013 to 2019. Moreover, this study explores the spatial spillover effect. The major study findings are as follows: (1) Digitalization and green economy growth represent a steady growth trend, and the former as a whole significantly promotes the latter, with a marginal effect of 1.648. (2) The mechanism analysis indicates the intermediary effects' size of three crucial intermediaries: green technology innovation > advanced industrial structure > the rationalization of industrial structure. (3) Both the "local effect" (0.556; 0.574) and "neighboring effect" (1.382; 1.415) of digitalization on green economy growth are positive under the two weight matrices and display "simultaneous resonance" characteristics based on the spatial perspective. (4) There exists obvious regional spatial heterogeneity and resource endowment heterogeneity. Finally, this study put forward corresponding policy implications, such as construction of new digital infrastructures and guiding green-energy consumption.

Can the marketization of urban land transfer improve energy efficiency?

Local government intervention in land resource allocation can lead to the misallocation of land resources and serious pollutant emissions. As an important market-oriented economic reform in China, the marketization of urban land transfer (MULT) might have the potential to contribute to improving resource allocation efficiency by curbing local government intervention. Therefore, this study aims to provide empirical evidence on the impact of MULT on energy efficiency. We improve the MULT evaluation method to test the mechanism through which MULT affects energy efficiency. The results show that, first, the proportion of land sold by allocation and listing methods, which is characterized by a low degree of marketization, has rapidly increased in recent years, lowering the overall level of MULT. Second, MULT has a direct and significant positive impact on improving energy efficiency. Third, the mechanism analysis indicates that MULT helps enhance energy efficiency by advancing industrial structure optimization and technological progress. Moreover, the heterogeneity analysis demonstrates that the impact of MULT on improving energy efficiency differs significantly in different reform stages and between central and peripheral cities. This study sheds light on the importance of land resource allocation in improving energy efficiency and thus has practical policy implications for promoting low-carbon energy transition in emerging countries.

Accurate multi-objective prediction of CO2 emission performance indexes and industrial structure optimization using multihead attention-based convolutional neural network.

The establishment of specific targets for the global carbon peaking and neutrality raises urgent requirements for prediction of CO2 emission performance indexes (CEPIs) and industrial structure optimization. However, accurate multi-objective prediction of CEPIs is still a knotty problem. In the present study, multihead attention-based convolutional neural network (MHA-CNN) model was proposed for accurate prediction of 4 CEPIs and further provided the rational suggestions for further industrial structure optimization. The proposed MHA-CNN model introduces deep learning mechanism with efficient resolution strategies for training model overfitting, feature extraction, and self-supervised learning to acquire the adaptability for CEPIs. Multihead attention (MHA) mechanism plays important roles in influence weight interpretation of variables to facilitate the prediction performance of CNN on CEPIs. The MHA-CNN model presented its overwhelmingly superior performance to CNN model and long short-term memory (LSTM) model, two frequently-used models, in multi-objective prediction of CEPIs using 8 influence variables, which highlighted advantages of MHA module in multi-dimensional feature extraction. Additionally, contributions of influence variables to CEPIs based on MHA analyses presented relatively high consistency with the geographical distribution analyses, indicating the excellent capacity of the MHA module in variable weights identification and contribution dissection. Based on the more accurate prediction results by MHA-CNN than those by CNN and LSTM model, the increase in the tertiary industry and the decreases in the first and secondary industries are conducive to improvement of total-factor carbon emission efficiency and further enhancement of effective energy utilization in regions with inefficient carbon emissions. This study provides insights towards the critical roles of the proposed MHA-CNN model in accurate multi-objective prediction of CEPIs and further industrial structure optimization for improvement of total-factor carbon emission efficiency.

Sand Discharge Simulation and Flow Path Optimization of a Particle Separator.

A numerical simulation method is used to optimize the removal of sand from a helicopter engine particle separator. First, the classic configuration of a particle separator based on the literature is simulated using two boundary conditions. The results show that the boundary conditions for the total pressure inlet and mass flow outlet are much more closely aligned with the experimental environment. By modifying the material at the front of the shroud, the separation efficiencies of coarse Arizona road dust (AC-Coarse) and MIL-E-5007C (C-Spec) can be improved to 93.3% and 97.6%, respectively. Configuration modifications of the particle separator with dual protection can increase the separation efficiencies of AC-Coarse and C-Spec to 91.7% and 97.7%.

Discovery, synthesis, and optimization of teixobactin, a novel antibiotic without detectable bacterial resistance.

Discovering new antibiotics with novel chemical scaffolds and antibacterial mechanisms presents a challenge for medicinal scientists worldwide as the ever-increasing bacterial resistance poses a serious threat to human health. A new cyclic peptide-based antibiotic termed teixobactin was discovered from a screen of uncultured soil bacteria through iChip technology in 2015. Teixobactin exhibits excellent antibacterial activity against all the tested gram-positive pathogens and Mycobacterium tuberculosis, including drug-resistant strains. Given that teixobactin targets the highly conserved lipid II and lipid III, which induces the simultaneous inhibition of both peptidoglycan and teichoic acid synthesis, the emergence of resistance is considered to be rather difficult. The novel structure, potent antibacterial activity, and highly conservative targets make teixobactin a promising lead compound for further antibiotic development. This review provides a comprehensive treatise on the advances of teixobactin in the areas of discovery processes, antibacterial activity, mechanisms of action, chemical synthesis, and structural optimizations. The synthetic methods for the key building block l-allo-End, natural teixobactin, representative teixobactin analogs, as well as the structure-activity relationship studies will be highlighted and discussed in details. Finally, some insights into new trends for the generation of novel teixobactin analogs and tips for future work and directions will be commented.

Discovery of Hyrtinadine A and Its Derivatives as Novel Antiviral and Anti-Phytopathogenic-Fungus Agents.

Plant diseases caused by viruses and fungi have a serious impact on the quality and yield of crops, endangering food security. The use of new, green, and efficient pesticides is an important strategy to increase crop output and deal with the food crisis. Ideally, the best pesticide innovation strategy is to find and use active compounds from natural products. Here, we took the marine natural product hyrtinadine A as the lead compound, and designed, synthesized, and systematically investigated a series of its derivatives for their antiviral and antifungal activities. Compound 8a was found to have excellent antiviral activity against the tobacco mosaic virus (TMV) (inactivation inhibitory effect of 55%/500 µg/mL and 19%/100 µg/mL, curative inhibitory effect of 52%/500 µg/mL and 22%/100 µg/mL, and protection inhibitory effect of 57%/500 µg/mL and 26%/100 µg/mL) and emerged as a novel antiviral candidate. These compound derivatives displayed broad-spectrum fungicidal activities against 14 kinds of phytopathogenic fungi at 50 µg/mL and the antifungal activities of compounds 5c, 5g, 6a, and 6e against Rhizoctonia cerealis are higher than that of the commercial fungicide chlorothalonil. Therefore, this study could lay a foundation for the application of hyrtinadine A derivatives in plant protection.

Neural Network Structure Optimization by Simulated Annealing.

A critical problem in large neural networks is over parameterization with a large number of weight parameters, which limits their use on edge devices due to prohibitive computational power and memory/storage requirements. To make neural networks more practical on edge devices and real-time industrial applications, they need to be compressed in advance. Since edge devices cannot train or access trained networks when internet resources are scarce, the preloading of smaller networks is essential. Various works in the literature have shown that the redundant branches can be pruned strategically in a fully connected network without sacrificing the performance significantly. However, majority of these methodologies need high computational resources to integrate weight training via the back-propagation algorithm during the process of network compression. In this work, we draw attention to the optimization of the network structure for preserving performance despite compression by pruning aggressively. The structure optimization is performed using the simulated annealing algorithm only, without utilizing back-propagation for branch weight training. Being a heuristic-based, non-convex optimization method, simulated annealing provides a globally near-optimal solution to this NP-hard problem for a given percentage of branch pruning. Our simulation results have shown that simulated annealing can significantly reduce the complexity of a fully connected network while maintaining the performance without the help of back-propagation.