Chaoshan drawnwork handkerchief design exhibits self-similarity and fractal characteristics due to their grid-based structure,overall symmetry,and the way local motifs reflect the whole pattern.To explore the potentia...Chaoshan drawnwork handkerchief design exhibits self-similarity and fractal characteristics due to their grid-based structure,overall symmetry,and the way local motifs reflect the whole pattern.To explore the potential of fractals in traditional textile design,a fractal-based generative framework was proposed for efficiently creating drawnwork patterns suitable for practical handicraft production.The research was initiated with an analysis of the structural composition of center,skeleton,and filler motifs extracted from a pattern sample library.Based on this hierarchical classification,the box-counting method was employed to calculate their respective fractal dimensions.Building on fractal art theory,generative algorithms,and studies on the application of Ultra Fractal,a Chaoshan drawnwork fractal design model was established.Using this model,51 drawnwork fractal patterns and 153 handkerchief patterns were generated.These patterns were subsequently applied in real-world production to validate the feasibility and value of fractal techniques in textile design.展开更多
Over the past century,advancements in chemistry have significantly propelled human innovation,enhancing both industrial and consumer products.However,this rapid progression has resulted in chemical pollution increasin...Over the past century,advancements in chemistry have significantly propelled human innovation,enhancing both industrial and consumer products.However,this rapid progression has resulted in chemical pollution increasingly surpassing planetary boundaries,as production and release rates have outpaced our monitoring capabilities.To catalyze more impactful efforts,this study transitions from traditional chemical assessment to inverse chemical design,introducing a generative graph latent diffusion model aimed at discovering safer alternatives.In a case study on the design of green solvents for cyclohexane/benzene extraction distillation,we constructed a design database encompassing functional,environmental hazards,and process constraints.Virtual screening of previous design dataset revealed distinct trade-off trends between these design requirements.Based on the screening outcomes,an unconstrained generative model was developed,which covered a broader chemical space and demonstrated superior capabilities for structural interpolation and extrapolation.To further optimize molecular generation towards desired properties,a multi-objective latent diffusion method was applied,yielding 19 candidate molecules.Of these,7 were identified in PubChem as the most viable green solvent candidates,while the remaining 12 as potential novel candidates.Overall,this study effectively designed green solvent candidates for safer and more sustainable industrial production,setting a promising precedent for the development of environmentally friendly alternatives in other areas of chemical research.展开更多
Aircraft conceptual design is a critical step in the development and research of aircraft,involving complex processes and multiple disciplines.Improving the efficiency of aircraft conceptual design while ensuring qual...Aircraft conceptual design is a critical step in the development and research of aircraft,involving complex processes and multiple disciplines.Improving the efficiency of aircraft conceptual design while ensuring quality is an important challenge.Intelligent technologies such as neural networks have played significant roles in areas like aerodynamics and structural analysis.However,due to issues such as high data demands and difficulties in transfer learning,their application in the conceptual design phase has been limited.The rise of generative artificial intelligence,exemplified by Large Language Model(LLM),offers a new approach to this problem.Therefore,this study proposes a methodology for generating aircraft conceptual design solutions based on LLMs and develops a prototype system.First,four of the current best-performing general-purpose LLMs are selected for deployment as foundational models.Then,based on the general prompt framework of LLMs,schema for aircraft conceptual design solutions,and real-world design cases,task prompts for generating aircraft conceptual design solutions are crafted,resulting in three types of prompts:Full-Instruction,1-Shot,and 5-Shot.Finally,the prototype system is utilized to design conceptual solutions,and the model-generated solutions are compared with those designed by engineers from both objective and subjective perspectives.The experimental results indicate that LLMs demonstrate conceptual design capabilities comparable to those of engineers,exhibiting strong generalization ability and potential for innovative design.展开更多
Multifocal metalenses are of great concern in optical communications,optical imaging and micro-optics systems,but their design is extremely challenging.In recent years,deep learning methods have provided novel solutio...Multifocal metalenses are of great concern in optical communications,optical imaging and micro-optics systems,but their design is extremely challenging.In recent years,deep learning methods have provided novel solutions to the design of optical planar devices.Here,an approach is proposed to explore the use of generative adversarial networks(GANs)to realize the design of metalenses with different focusing positions at dual wavelengths.This approach includes a forward network and an inverse network,where the former predicts the optical response of meta-atoms and the latter generates structures that meet specific requirements.Compared to the traditional search method,the inverse network demonstrates higher precision and efficiency in designing a dual-wavelength bifocal metalens.The results will provide insights and methodologies for the design of tunable wavelength metalenses,while also highlighting the potential of deep learning in optical device design.展开更多
To ensure an uninterrupted power supply,mobile power sources(MPS)are widely deployed in power grids during emergencies.Comprising mobile emergency generators(MEGs)and mobile energy storage systems(MESS),MPS are capabl...To ensure an uninterrupted power supply,mobile power sources(MPS)are widely deployed in power grids during emergencies.Comprising mobile emergency generators(MEGs)and mobile energy storage systems(MESS),MPS are capable of supplying power to critical loads and serving as backup sources during grid contingencies,offering advantages such as flexibility and high resilience through electricity delivery via transportation networks.This paper proposes a design method for a 400 V–10 kV Dual-Winding Induction Generator(DWIG)intended for MEG applications,employing an improved particle swarmoptimization(PSO)algorithmbased on a back-propagation neural network(BPNN).A parameterized finite element(FE)model of the DWIG is established to derive constraints on its dimensional parameters,thereby simplifying the optimization space.Through sensitivity analysis between temperature rise and electromagnetic loss of the DWIG,the main factors influencing the machine’s temperature are identified,and electromagnetic loss is determined as the optimization objective.To obtain an accurate fitting function between electromagnetic loss and dimensional parameters,the BPNN is employed to predict the nonlinear relationship between the optimization objective and the parameters.The Latin hypercube sampling(LHS)method is used for random sampling in the FE model analysis for training,testing,and validation,which is then applied to compute the cost function in the PSO.Based on the relationships obtained by the BPNN,the PSO algorithm evaluates the fitness and cost functions to determine the optimal design point.The proposed optimization method is validated by comparing simulation results between the initial design and the optimized design.展开更多
Inverse design has long been an efficient and powerful design tool in the aircraft industry.In this paper,a novel inverse design method for supercritical airfoils is proposed based on generative models in deep learnin...Inverse design has long been an efficient and powerful design tool in the aircraft industry.In this paper,a novel inverse design method for supercritical airfoils is proposed based on generative models in deep learning.A Conditional Variational Auto Encoder(CVAE)and an integrated generative network CVAE-GAN that combines the CVAE with the Wasserstein Generative Adversarial Networks(WGAN),are conducted as generative models.They are used to generate target wall Mach distributions for the inverse design that matches specified features,such as locations of suction peak,shock and aft loading.Qualitative and quantitative results show that both adopted generative models can generate diverse and realistic wall Mach number distributions satisfying the given features.The CVAE-GAN model outperforms the CVAE model and achieves better reconstruction accuracies for all the samples in the dataset.Furthermore,a deep neural network for nonlinear mapping is adopted to obtain the airfoil shape corresponding to the target wall Mach number distribution.The performances of the designed deep neural network are fully demonstrated and a smoothness measurement is proposed to quantify small oscillations in the airfoil surface,proving the authenticity and accuracy of the generated airfoil shapes.展开更多
Multi-principal element alloys(MPEAs),inclusive of high entropy alloys(HEAs),continue to attract significant research attention owing to their potentially desirable properties.Although MPEAs remain under extensive res...Multi-principal element alloys(MPEAs),inclusive of high entropy alloys(HEAs),continue to attract significant research attention owing to their potentially desirable properties.Although MPEAs remain under extensive research,traditional(i.e.empirical)alloy production and testing are both costly and timeconsuming,partly due to the inefficiency of the early discovery process which involves experiments on a large number of alloy compositions.It is intuitive to apply machine learning in the discovery of this novel class of materials,of which only a small number of potential alloys have been probed to date.In this work,a proof-of-concept is proposed,combining generative adversarial networks(GANs)with discriminative neural networks(NNs),to accelerate the exploration of novel MPEAs.By applying the GAN model herein,it was possible to directly generate novel compositions for MPEAs,and to predict their phases.To verify the predictability of the model,alloys designed by the model are presented and a candidate produced-as validation.This suggests that the model herein offers an approach that can significantly enhance the capacity and efficiency of development of novel MPEAs.展开更多
With the increasing demands of health care,the design of hospital buildings has become increasingly demanding and complicated.However,the traditional layout design method for hospital is labor intensive,time consuming...With the increasing demands of health care,the design of hospital buildings has become increasingly demanding and complicated.However,the traditional layout design method for hospital is labor intensive,time consuming and prone to errors.With the development of artificial intelligence(AI),the intelligent design method has become possible and is considered to be suitable for the layout design of hospital buildings.Two intelli-gent design processes based on healthcare systematic layout planning(HSLP)and generative adversarial network(GAN)are proposed in this paper,which aim to solve the generation problem of the plane functional layout of the operating departments(ODs)of general hospitals.The first design method that is more like a mathemati-cal model with traditional optimization algorithm concerns the following two steps:developing the HSLP model based on the conventional systematic layout planning(SLP)theory,identifying the relationship and flows amongst various departments/units,and arriving at the preliminary plane layout design;establishing mathematical model to optimize the building layout by using the genetic algorithm(GA)to obtain the optimized scheme.The specific process of the second intelligent design based on more than 100 sets of collected OD drawings includes:labelling the corresponding functional layouts of each OD plan;building image-to-image translation with conditional ad-versarial network(pix2pix)for training OD plane layouts,which is one of the most representative GAN models.Finally,the functions and features of the results generated by the two methods are analyzed and compared from an architectural and algorithmic perspective.Comparison of the two design methods shows that the HSLP and GAN models can autonomously generate new OD plane functional layouts.The HSLP layouts have clear functional area adjacencies and optimization goals,but the layouts are relatively rigid and not specific enough.The GAN outputs are the most innovative layouts with strong applicability,but the dataset has strict constraints.The goal of this paper is to help release the heavy load of architects in the early design stage and present the effectiveness of these intelligent design methods in the field of medical architecture.展开更多
Natural products（NPs） have long been recognized as a valuable resource for drug discovery, and bringing NP-related features to virtual libraries is believed to be an effective way to increase the coverage of druggab...Natural products（NPs） have long been recognized as a valuable resource for drug discovery, and bringing NP-related features to virtual libraries is believed to be an effective way to increase the coverage of druggable chemical space. Here, deep learning-based molecule generative model, which is a recent technique in de novo molecule design, was applied to generate virtual libraries with NP-like properties. Results demonstrated that the model was effective in generating molecules that highly resemble NPs. Moreover, the model was also found to be capable of generating NP-like molecules that were also easy to synthesize, significantly increasing the practical value of the compound library.展开更多
The distribution of material phases is crucial to determine the composite’s mechanical property.While the full structure-mechanics relationship of highly ordered material distributions can be studied with finite numb...The distribution of material phases is crucial to determine the composite’s mechanical property.While the full structure-mechanics relationship of highly ordered material distributions can be studied with finite number of cases,this relationship is difficult to be revealed for complex irregular distributions,preventing design of such material structures to meet certain mechanical requirements.The noticeable developments of artificial intelligence(AI)algorithms in material design enables to detect the hidden structure-mechanics correlations which is essential for designing composite of complex structures.It is intriguing how these tools can assist composite design.Here,we focus on the rapid generation of bicontinuous composite structures together with the stress distribution in loading.We find that generative AI,enabled through fine-tuned Low Rank Adaptation models,can be trained with a few inputs to generate both synthetic composite structures and the corresponding von Mises stress distribution.The results show that this technique is convenient in generating massive composites designs with useful mechanical information that dictate stiffness,fracture and robustness of the material with one model,and such has to be done by several different experimental or simulation tests.This research offers valuable insights for the improvement of composite design with the goal of expanding the design space and automatic screening of composite designs for improved mechanical functions.展开更多
Three-factor orthogonal design（OD） of Er3＋/Gd3＋/T（calcination temperature） is used to optimize the luminescent intensity of Na Y（Gd）（MoO4）2：Er3＋phosphor.Firstly,the uniform design（UD） is introduced to ex...Three-factor orthogonal design（OD） of Er3＋/Gd3＋/T（calcination temperature） is used to optimize the luminescent intensity of Na Y（Gd）（MoO4）2：Er3＋phosphor.Firstly,the uniform design（UD） is introduced to explore the doping concentration range of Er3＋/Gd3＋.Then OD and range analysis are performed based on the results of UD to obtain the primary and secondary sequence and the best combination of Er3＋,Gd3＋,and T within the experimental range.The optimum sample is prepared by the high temperature solid state method.Photoluminescence excitation and emission spectra of the optimum sample are detected.The intense green emissions（530 nm and 550 nm） are observed which originate from Er3＋2H11/2→4I15/2and4S3/2→4I15/2,respectively.Thermal effect is investigated in the optimum NaY（Gd3＋）（MoO4）2：Er3＋phosphors,and the green emission intensity decreases as temperature increases.展开更多
With the digital transformation of global education and China's emphasis on education digital,generative AI technology has been widely used in the field of higher education.In this paper,the development of generat...With the digital transformation of global education and China's emphasis on education digital,generative AI technology has been widely used in the field of higher education.In this paper,the development of generative AI technology and its potential in personalized learning,interactive content creation and adaptive assessment in education were introduced firstly.Then,the application case of generative AI tools in teaching content creation,scenario-based teaching content development,visual teaching content development,complex concept deconstruction and analogy,student-led application practice and other aspects in the teaching of Building Decoration Materials was discussed.Through the teaching experiment and effect evaluation,the positive influence of generative AI technology on the improvement of students'learning effect and teaching efficiency was verified.Finally,some thoughts and inspirations on the combination of educational theory and generative AI technology,the integration of teaching design and generative AI technology,and the practice cases and effect evaluation were put forward,and the importance of teacher role transformation and personalized learning path design was emphasized to provide theoretical and practical support for the innovative development of higher education.展开更多
With the continuous scaling of integrated circuit technologies,design for manufacturability(DFM)is becoming more critical,yet more challenging.Alongside,recent advances in machine learning have provided a new computin...With the continuous scaling of integrated circuit technologies,design for manufacturability(DFM)is becoming more critical,yet more challenging.Alongside,recent advances in machine learning have provided a new computing paradigm with promising applications in VLSI manufacturability.In particular,generative learning-regarded among the most interesting ideas in present-day machine learning-has demonstrated impressive capabilities in a wide range of applications.This paper surveys recent results of using generative learning in VLSI manufacturing modeling and optimization.Specifically,we examine the unique features of generative learning that have been leveraged to improve DFM efficiency in an unprecedented way;hence,paving the way to a new data-driven DFM approach.The state-of-the-art methods are presented,and challenges/opportunities are discussed.展开更多
Severe acute respiratory syndrome coronavirus 2(SARS-CoV-2)mutations are influenced by random and uncontrollable factors,and the risk of the next widespread epidemic remains.Dual-target drugs that synergistically act ...Severe acute respiratory syndrome coronavirus 2(SARS-CoV-2)mutations are influenced by random and uncontrollable factors,and the risk of the next widespread epidemic remains.Dual-target drugs that synergistically act on two targets exhibit strong therapeutic effects and advantages against mutations.In this study,a novel computational workflow was developed to design dual-target SARS-CoV-2 candidate inhibitors with the Envelope protein and Main protease selected as the two target proteins.The drug-like molecules of our self-constructed 3D scaffold database were used as high-throughput molecular docking probes for feature extraction of two target protein pockets.A multi-layer perceptron(MLP)was employed to embed the binding affinities into a latent space as conditional vectors to control conditional distribution.Utilizing a conditional generative neural network,cG-SchNet,with 3D Euclidean group(E3)symmetries,the conditional probability distributions of molecular 3D structures were acquired and a set of novel SARS-CoV-2 dual-target candidate inhibitors were generated.The 1D probability,2D joint probability,and 2D cumulative probability distribution results indicate that the generated sets are significantly enhanced compared to the training set in the high binding affinity area.Among the 201 generated molecules,42 molecules exhibited a sum binding affinity exceeding 17.0 kcal/mol while 9 of them having a sum binding affinity exceeding 19.0 kcal/mol,demonstrating structure diversity along with strong dual-target affinities,good absorption,distribution,metabolism,excretion,and toxicity(ADMET)properties,and ease of synthesis.Dual-target drugs are rare and difficult to find,and our“high-throughput docking-multi-conditional generation”workflow offers a wide range of options for designing or optimizing potent dual-target SARS-CoV-2 inhibitors.展开更多
To meet the extreme precision requirements of nanometer-scale semiconductor manufacturing and micrometer-level aerospace component processing,the complexity of precision manufacturing equipment design has exceeded the...To meet the extreme precision requirements of nanometer-scale semiconductor manufacturing and micrometer-level aerospace component processing,the complexity of precision manufacturing equipment design has exceeded the capabilities of traditional design methodologies.Conventional experience-driven design approaches exhibit fundamental limitations when confronting high-dimensional parameter spaces,complex multidisciplinary coupling effects,and dynamic performance prediction requirements,rendering trial-and-error iterative optimization processes inefficient and incapable of achieving optimal solutions.Intelligent design offers new pathways to overcome these limitations through the integration of artificial intelligence(AI)with traditional engineering workflows.However,the transition from theoretical concepts to manufacturing practice encounters three critical technical bottlenecks:the sparsity and heterogeneity of design data constrain the development of domain-specific large models,hallucination phenomena in generative design compromise solution trustworthiness,and numerical simulation methods face fundamental trade-offs between computational accuracy and efficiency.This paper conducts comprehensive analysis of the underlying causes of these challenges and proposes a knowledge-generation-simulation integrated intelligent design ecosystem as a development pathway.This approach achieves deep integration of large models with manufacturing domain knowledge,seamless fusion of AI with Computer-Aided Design/Computer-Aided Engineering(CAD/CAE)systems,and comprehensive synthesis of physics-based mechanisms with data-driven methods,driving the evolution of intelligent design from human-dominated iterative processes toward autonomous collaborative innovation systems,thereby providing robust support for technological breakthroughs in precision and extreme manufacturing equipment while facilitating the intelligent transformation of the manufacturing industry.展开更多
In this study,we introduce a deep generative model,named Multi-Species Generative Adversarial Network(MS-GAN),which is developed to extract the low-dimensional manifold of three-dimensional multi-species surfaces.In t...In this study,we introduce a deep generative model,named Multi-Species Generative Adversarial Network(MS-GAN),which is developed to extract the low-dimensional manifold of three-dimensional multi-species surfaces.In the development of MS-GAN,we extend the freeform deformation by incorporating principal component analysis to increase the non-linear deformation ability while maintaining geometric smoothness.The implicit information of multiple baselines is embedded in the feature extraction layers,to enhance the diversity and parameterization of multi-species dataset.Furthermore,Wasserstein GAN with a gradient penalty is used to ensure the stability and convergence of the training networks.Two experiments,ruled surfaces and propeller blade surfaces,are performed to demonstrate the advantages and superiorities of MS-GAN.展开更多
Recent advances in contrastive language-image pretraining(CLIP)models and generative AI have demonstrated significant capabilities in cross-modal understanding and content generation.Based on these developments,this s...Recent advances in contrastive language-image pretraining(CLIP)models and generative AI have demonstrated significant capabilities in cross-modal understanding and content generation.Based on these developments,this study introduces a novel framework for airfoil design via natural language interfaces.To the authors’knowledge,this study establishes the first end-to-end,bidirectional mapping between textual descriptions(e.g.,“low-drag supercritical wing for transonic conditions”)and parametric airfoil geometries represented by class-shape transformation parameters.The proposed approach integrates a CLIP-inspired architecture that aligns text embeddings with airfoil parameter spaces through contrastive learning,along with a semantically conditioned decoder that produces physically plausible airfoil geometries from latent representations.The experimental results validate the framework’s ability to generate aerodynamically plausible airfoils from natural language specifications and to classify airfoils accurately based on given textual labels.This research reduces the expertise threshold for preliminary airfoil design and highlights the potential for human-AI collaboration in aerospace engineering.展开更多
Generative artificial intelligence(Generative AI)is reshaping both learning and teaching paradigms in medical education.With the advancement of Large Language Models(LLMs)-based tools such as ChatGPT,Gemini,and other ...Generative artificial intelligence(Generative AI)is reshaping both learning and teaching paradigms in medical education.With the advancement of Large Language Models(LLMs)-based tools such as ChatGPT,Gemini,and other medical-domain-specific models,Generative AI shows strong potential to address persistent challenges in medical education,including rigid curricula,unequal access to educational resources,and the diverse learning needs of medical students.This review summarizes the applications of Generative AI across key domains:(1)personalized learning through real-time analysis of student performance;(2)clinical skills training via immersive simulations and virtual patients;(3)automated generation of teaching materials such as clinical cases and assessments;and(4)support for student research and academic writing.Empirical evidence indicates that Generative AI-enhanced instruction can improve knowledge acquisition,clinical reasoning,and overall educational efficiency.However,challenges remain,including the generation of inaccurate or fabricated content,risks to academic integrity,algorithmic bias,data privacy concerns,and unresolved ethical issues regarding AI’s role in teaching.Without proper oversight,these risks may compromise educational quality and equity.To ensure responsible adoption,this review advocates for the establishment of institutional policies,enhancement of educators’AI literacy,transparent model validation,and a human-centered design framework that positions Generative AI as a collaborative teaching assistant.When responsibly integrated,Generative AI holds the transformative potential to cultivate future medical professionals equipped with clinical competence,responsibility,and innovative thinking.展开更多
文摘Chaoshan drawnwork handkerchief design exhibits self-similarity and fractal characteristics due to their grid-based structure,overall symmetry,and the way local motifs reflect the whole pattern.To explore the potential of fractals in traditional textile design,a fractal-based generative framework was proposed for efficiently creating drawnwork patterns suitable for practical handicraft production.The research was initiated with an analysis of the structural composition of center,skeleton,and filler motifs extracted from a pattern sample library.Based on this hierarchical classification,the box-counting method was employed to calculate their respective fractal dimensions.Building on fractal art theory,generative algorithms,and studies on the application of Ultra Fractal,a Chaoshan drawnwork fractal design model was established.Using this model,51 drawnwork fractal patterns and 153 handkerchief patterns were generated.These patterns were subsequently applied in real-world production to validate the feasibility and value of fractal techniques in textile design.
基金supported by Shanghai Science and Technology Commission Project(No.21DZ1201502)Shanghai Municipal Bureau of Ecology and Environment(Shanghai Environ-mental Science[2023]No.40)+1 种基金the Interdisciplinary Joint Research Project of Tongji University(No.2022-4-YB-12)Shanghai Science and Technology Commission Project(No.22DZ2200200).
文摘Over the past century,advancements in chemistry have significantly propelled human innovation,enhancing both industrial and consumer products.However,this rapid progression has resulted in chemical pollution increasingly surpassing planetary boundaries,as production and release rates have outpaced our monitoring capabilities.To catalyze more impactful efforts,this study transitions from traditional chemical assessment to inverse chemical design,introducing a generative graph latent diffusion model aimed at discovering safer alternatives.In a case study on the design of green solvents for cyclohexane/benzene extraction distillation,we constructed a design database encompassing functional,environmental hazards,and process constraints.Virtual screening of previous design dataset revealed distinct trade-off trends between these design requirements.Based on the screening outcomes,an unconstrained generative model was developed,which covered a broader chemical space and demonstrated superior capabilities for structural interpolation and extrapolation.To further optimize molecular generation towards desired properties,a multi-objective latent diffusion method was applied,yielding 19 candidate molecules.Of these,7 were identified in PubChem as the most viable green solvent candidates,while the remaining 12 as potential novel candidates.Overall,this study effectively designed green solvent candidates for safer and more sustainable industrial production,setting a promising precedent for the development of environmentally friendly alternatives in other areas of chemical research.
基金funded by Henan Key Laboratory of General Aviation Technology,China(No.ZHKF-240202).
文摘Aircraft conceptual design is a critical step in the development and research of aircraft,involving complex processes and multiple disciplines.Improving the efficiency of aircraft conceptual design while ensuring quality is an important challenge.Intelligent technologies such as neural networks have played significant roles in areas like aerodynamics and structural analysis.However,due to issues such as high data demands and difficulties in transfer learning,their application in the conceptual design phase has been limited.The rise of generative artificial intelligence,exemplified by Large Language Model(LLM),offers a new approach to this problem.Therefore,this study proposes a methodology for generating aircraft conceptual design solutions based on LLMs and develops a prototype system.First,four of the current best-performing general-purpose LLMs are selected for deployment as foundational models.Then,based on the general prompt framework of LLMs,schema for aircraft conceptual design solutions,and real-world design cases,task prompts for generating aircraft conceptual design solutions are crafted,resulting in three types of prompts:Full-Instruction,1-Shot,and 5-Shot.Finally,the prototype system is utilized to design conceptual solutions,and the model-generated solutions are compared with those designed by engineers from both objective and subjective perspectives.The experimental results indicate that LLMs demonstrate conceptual design capabilities comparable to those of engineers,exhibiting strong generalization ability and potential for innovative design.
基金National Natural Science Foundation of China(No.61975029)。
文摘Multifocal metalenses are of great concern in optical communications,optical imaging and micro-optics systems,but their design is extremely challenging.In recent years,deep learning methods have provided novel solutions to the design of optical planar devices.Here,an approach is proposed to explore the use of generative adversarial networks(GANs)to realize the design of metalenses with different focusing positions at dual wavelengths.This approach includes a forward network and an inverse network,where the former predicts the optical response of meta-atoms and the latter generates structures that meet specific requirements.Compared to the traditional search method,the inverse network demonstrates higher precision and efficiency in designing a dual-wavelength bifocal metalens.The results will provide insights and methodologies for the design of tunable wavelength metalenses,while also highlighting the potential of deep learning in optical device design.
基金funded by the Science and Technology Projects of State Grid Corporation of China(Project No.J2024136).
文摘To ensure an uninterrupted power supply,mobile power sources(MPS)are widely deployed in power grids during emergencies.Comprising mobile emergency generators(MEGs)and mobile energy storage systems(MESS),MPS are capable of supplying power to critical loads and serving as backup sources during grid contingencies,offering advantages such as flexibility and high resilience through electricity delivery via transportation networks.This paper proposes a design method for a 400 V–10 kV Dual-Winding Induction Generator(DWIG)intended for MEG applications,employing an improved particle swarmoptimization(PSO)algorithmbased on a back-propagation neural network(BPNN).A parameterized finite element(FE)model of the DWIG is established to derive constraints on its dimensional parameters,thereby simplifying the optimization space.Through sensitivity analysis between temperature rise and electromagnetic loss of the DWIG,the main factors influencing the machine’s temperature are identified,and electromagnetic loss is determined as the optimization objective.To obtain an accurate fitting function between electromagnetic loss and dimensional parameters,the BPNN is employed to predict the nonlinear relationship between the optimization objective and the parameters.The Latin hypercube sampling(LHS)method is used for random sampling in the FE model analysis for training,testing,and validation,which is then applied to compute the cost function in the PSO.Based on the relationships obtained by the BPNN,the PSO algorithm evaluates the fitness and cost functions to determine the optimal design point.The proposed optimization method is validated by comparing simulation results between the initial design and the optimized design.
基金co-supported by the National Key Project of China(No.GJXM92579)the National Natural Science Foundation of China(Nos.92052203,61903178 and61906081)。
文摘Inverse design has long been an efficient and powerful design tool in the aircraft industry.In this paper,a novel inverse design method for supercritical airfoils is proposed based on generative models in deep learning.A Conditional Variational Auto Encoder(CVAE)and an integrated generative network CVAE-GAN that combines the CVAE with the Wasserstein Generative Adversarial Networks(WGAN),are conducted as generative models.They are used to generate target wall Mach distributions for the inverse design that matches specified features,such as locations of suction peak,shock and aft loading.Qualitative and quantitative results show that both adopted generative models can generate diverse and realistic wall Mach number distributions satisfying the given features.The CVAE-GAN model outperforms the CVAE model and achieves better reconstruction accuracies for all the samples in the dataset.Furthermore,a deep neural network for nonlinear mapping is adopted to obtain the airfoil shape corresponding to the target wall Mach number distribution.The performances of the designed deep neural network are fully demonstrated and a smoothness measurement is proposed to quantify small oscillations in the airfoil surface,proving the authenticity and accuracy of the generated airfoil shapes.
文摘Multi-principal element alloys(MPEAs),inclusive of high entropy alloys(HEAs),continue to attract significant research attention owing to their potentially desirable properties.Although MPEAs remain under extensive research,traditional(i.e.empirical)alloy production and testing are both costly and timeconsuming,partly due to the inefficiency of the early discovery process which involves experiments on a large number of alloy compositions.It is intuitive to apply machine learning in the discovery of this novel class of materials,of which only a small number of potential alloys have been probed to date.In this work,a proof-of-concept is proposed,combining generative adversarial networks(GANs)with discriminative neural networks(NNs),to accelerate the exploration of novel MPEAs.By applying the GAN model herein,it was possible to directly generate novel compositions for MPEAs,and to predict their phases.To verify the predictability of the model,alloys designed by the model are presented and a candidate produced-as validation.This suggests that the model herein offers an approach that can significantly enhance the capacity and efficiency of development of novel MPEAs.
基金the Scientific Research Project of Shanghai Science and Technology Commission(No.18DZ1205603)the Science Research Plan of Shanghai Municipal Science and Technology Committee(No.20DZ1201300)the National Key Research and Development Program of China(No.2017YFC0806100)。
文摘With the increasing demands of health care,the design of hospital buildings has become increasingly demanding and complicated.However,the traditional layout design method for hospital is labor intensive,time consuming and prone to errors.With the development of artificial intelligence(AI),the intelligent design method has become possible and is considered to be suitable for the layout design of hospital buildings.Two intelli-gent design processes based on healthcare systematic layout planning(HSLP)and generative adversarial network(GAN)are proposed in this paper,which aim to solve the generation problem of the plane functional layout of the operating departments(ODs)of general hospitals.The first design method that is more like a mathemati-cal model with traditional optimization algorithm concerns the following two steps:developing the HSLP model based on the conventional systematic layout planning(SLP)theory,identifying the relationship and flows amongst various departments/units,and arriving at the preliminary plane layout design;establishing mathematical model to optimize the building layout by using the genetic algorithm(GA)to obtain the optimized scheme.The specific process of the second intelligent design based on more than 100 sets of collected OD drawings includes:labelling the corresponding functional layouts of each OD plan;building image-to-image translation with conditional ad-versarial network(pix2pix)for training OD plane layouts,which is one of the most representative GAN models.Finally,the functions and features of the results generated by the two methods are analyzed and compared from an architectural and algorithmic perspective.Comparison of the two design methods shows that the HSLP and GAN models can autonomously generate new OD plane functional layouts.The HSLP layouts have clear functional area adjacencies and optimization goals,but the layouts are relatively rigid and not specific enough.The GAN outputs are the most innovative layouts with strong applicability,but the dataset has strict constraints.The goal of this paper is to help release the heavy load of architects in the early design stage and present the effectiveness of these intelligent design methods in the field of medical architecture.
基金The National Natural Science Foundation of China(Grant No.81573273,81673279,21572010 and 21772005)National Major Scientific and Technological Special Project for"Significant New Drugs Development"(Grant No.2018ZX09735001-003)
文摘Natural products（NPs） have long been recognized as a valuable resource for drug discovery, and bringing NP-related features to virtual libraries is believed to be an effective way to increase the coverage of druggable chemical space. Here, deep learning-based molecule generative model, which is a recent technique in de novo molecule design, was applied to generate virtual libraries with NP-like properties. Results demonstrated that the model was effective in generating molecules that highly resemble NPs. Moreover, the model was also found to be capable of generating NP-like molecules that were also easy to synthesize, significantly increasing the practical value of the compound library.
基金supported by the National Science Foundation CA-REER Grant(Grant No.2145392)the startup funding at Syracuse Uni-versity for supporting the research work.
文摘The distribution of material phases is crucial to determine the composite’s mechanical property.While the full structure-mechanics relationship of highly ordered material distributions can be studied with finite number of cases,this relationship is difficult to be revealed for complex irregular distributions,preventing design of such material structures to meet certain mechanical requirements.The noticeable developments of artificial intelligence(AI)algorithms in material design enables to detect the hidden structure-mechanics correlations which is essential for designing composite of complex structures.It is intriguing how these tools can assist composite design.Here,we focus on the rapid generation of bicontinuous composite structures together with the stress distribution in loading.We find that generative AI,enabled through fine-tuned Low Rank Adaptation models,can be trained with a few inputs to generate both synthetic composite structures and the corresponding von Mises stress distribution.The results show that this technique is convenient in generating massive composites designs with useful mechanical information that dictate stiffness,fracture and robustness of the material with one model,and such has to be done by several different experimental or simulation tests.This research offers valuable insights for the improvement of composite design with the goal of expanding the design space and automatic screening of composite designs for improved mechanical functions.
基金Project supported by Education Reform Fund of Dalian Maritime University,China(Grant No.2015Y37)the Natural Science Foundation of Liaoning Province,China(Grant Nos.2015020190 and 2014025010)+1 种基金the Open Fund of the State Key Laboratory on Integrated Optoelectronics,China(Grant No.IOSKL2015KF27)the Fundamental Research Funds for the Central Universities,China(Grant No.3132016121)
文摘Three-factor orthogonal design（OD） of Er3＋/Gd3＋/T（calcination temperature） is used to optimize the luminescent intensity of Na Y（Gd）（MoO4）2：Er3＋phosphor.Firstly,the uniform design（UD） is introduced to explore the doping concentration range of Er3＋/Gd3＋.Then OD and range analysis are performed based on the results of UD to obtain the primary and secondary sequence and the best combination of Er3＋,Gd3＋,and T within the experimental range.The optimum sample is prepared by the high temperature solid state method.Photoluminescence excitation and emission spectra of the optimum sample are detected.The intense green emissions（530 nm and 550 nm） are observed which originate from Er3＋2H11/2→4I15/2and4S3/2→4I15/2,respectively.Thermal effect is investigated in the optimum NaY（Gd3＋）（MoO4）2：Er3＋phosphors,and the green emission intensity decreases as temperature increases.
文摘With the digital transformation of global education and China's emphasis on education digital,generative AI technology has been widely used in the field of higher education.In this paper,the development of generative AI technology and its potential in personalized learning,interactive content creation and adaptive assessment in education were introduced firstly.Then,the application case of generative AI tools in teaching content creation,scenario-based teaching content development,visual teaching content development,complex concept deconstruction and analogy,student-led application practice and other aspects in the teaching of Building Decoration Materials was discussed.Through the teaching experiment and effect evaluation,the positive influence of generative AI technology on the improvement of students'learning effect and teaching efficiency was verified.Finally,some thoughts and inspirations on the combination of educational theory and generative AI technology,the integration of teaching design and generative AI technology,and the practice cases and effect evaluation were put forward,and the importance of teacher role transformation and personalized learning path design was emphasized to provide theoretical and practical support for the innovative development of higher education.
文摘With the continuous scaling of integrated circuit technologies,design for manufacturability(DFM)is becoming more critical,yet more challenging.Alongside,recent advances in machine learning have provided a new computing paradigm with promising applications in VLSI manufacturability.In particular,generative learning-regarded among the most interesting ideas in present-day machine learning-has demonstrated impressive capabilities in a wide range of applications.This paper surveys recent results of using generative learning in VLSI manufacturing modeling and optimization.Specifically,we examine the unique features of generative learning that have been leveraged to improve DFM efficiency in an unprecedented way;hence,paving the way to a new data-driven DFM approach.The state-of-the-art methods are presented,and challenges/opportunities are discussed.
基金supported by Interdisciplinary Innova-tion Project of“Bioarchaeology Laboratory”of Jilin University,China,and“MedicineþX”Interdisciplinary Innovation Team of Norman Bethune Health Science Center of Jilin University,China(Grant No.:2022JBGS05).
文摘Severe acute respiratory syndrome coronavirus 2(SARS-CoV-2)mutations are influenced by random and uncontrollable factors,and the risk of the next widespread epidemic remains.Dual-target drugs that synergistically act on two targets exhibit strong therapeutic effects and advantages against mutations.In this study,a novel computational workflow was developed to design dual-target SARS-CoV-2 candidate inhibitors with the Envelope protein and Main protease selected as the two target proteins.The drug-like molecules of our self-constructed 3D scaffold database were used as high-throughput molecular docking probes for feature extraction of two target protein pockets.A multi-layer perceptron(MLP)was employed to embed the binding affinities into a latent space as conditional vectors to control conditional distribution.Utilizing a conditional generative neural network,cG-SchNet,with 3D Euclidean group(E3)symmetries,the conditional probability distributions of molecular 3D structures were acquired and a set of novel SARS-CoV-2 dual-target candidate inhibitors were generated.The 1D probability,2D joint probability,and 2D cumulative probability distribution results indicate that the generated sets are significantly enhanced compared to the training set in the high binding affinity area.Among the 201 generated molecules,42 molecules exhibited a sum binding affinity exceeding 17.0 kcal/mol while 9 of them having a sum binding affinity exceeding 19.0 kcal/mol,demonstrating structure diversity along with strong dual-target affinities,good absorption,distribution,metabolism,excretion,and toxicity(ADMET)properties,and ease of synthesis.Dual-target drugs are rare and difficult to find,and our“high-throughput docking-multi-conditional generation”workflow offers a wide range of options for designing or optimizing potent dual-target SARS-CoV-2 inhibitors.
基金supported by the National Key Research and Development Program of China(Grant No.2024YFB3309500)the National Natural Science Foundation of China(Grant Nos.U24B6005,U22A6001)。
文摘To meet the extreme precision requirements of nanometer-scale semiconductor manufacturing and micrometer-level aerospace component processing,the complexity of precision manufacturing equipment design has exceeded the capabilities of traditional design methodologies.Conventional experience-driven design approaches exhibit fundamental limitations when confronting high-dimensional parameter spaces,complex multidisciplinary coupling effects,and dynamic performance prediction requirements,rendering trial-and-error iterative optimization processes inefficient and incapable of achieving optimal solutions.Intelligent design offers new pathways to overcome these limitations through the integration of artificial intelligence(AI)with traditional engineering workflows.However,the transition from theoretical concepts to manufacturing practice encounters three critical technical bottlenecks:the sparsity and heterogeneity of design data constrain the development of domain-specific large models,hallucination phenomena in generative design compromise solution trustworthiness,and numerical simulation methods face fundamental trade-offs between computational accuracy and efficiency.This paper conducts comprehensive analysis of the underlying causes of these challenges and proposes a knowledge-generation-simulation integrated intelligent design ecosystem as a development pathway.This approach achieves deep integration of large models with manufacturing domain knowledge,seamless fusion of AI with Computer-Aided Design/Computer-Aided Engineering(CAD/CAE)systems,and comprehensive synthesis of physics-based mechanisms with data-driven methods,driving the evolution of intelligent design from human-dominated iterative processes toward autonomous collaborative innovation systems,thereby providing robust support for technological breakthroughs in precision and extreme manufacturing equipment while facilitating the intelligent transformation of the manufacturing industry.
基金support of the National Natural Science Foundation of China(No.12372221)is acknowledged.
文摘In this study,we introduce a deep generative model,named Multi-Species Generative Adversarial Network(MS-GAN),which is developed to extract the low-dimensional manifold of three-dimensional multi-species surfaces.In the development of MS-GAN,we extend the freeform deformation by incorporating principal component analysis to increase the non-linear deformation ability while maintaining geometric smoothness.The implicit information of multiple baselines is embedded in the feature extraction layers,to enhance the diversity and parameterization of multi-species dataset.Furthermore,Wasserstein GAN with a gradient penalty is used to ensure the stability and convergence of the training networks.Two experiments,ruled surfaces and propeller blade surfaces,are performed to demonstrate the advantages and superiorities of MS-GAN.
基金supported by the National Natural Science Foundation of China(Grant Nos.U23A2069,12372288,12388101,and 92152301)Jilin Province Science and Technology Development Program,China(Grant No.20220301013GX)Aeronautical Science Foundation of China(Grant No.2020Z006058002)。
文摘Recent advances in contrastive language-image pretraining(CLIP)models and generative AI have demonstrated significant capabilities in cross-modal understanding and content generation.Based on these developments,this study introduces a novel framework for airfoil design via natural language interfaces.To the authors’knowledge,this study establishes the first end-to-end,bidirectional mapping between textual descriptions(e.g.,“low-drag supercritical wing for transonic conditions”)and parametric airfoil geometries represented by class-shape transformation parameters.The proposed approach integrates a CLIP-inspired architecture that aligns text embeddings with airfoil parameter spaces through contrastive learning,along with a semantically conditioned decoder that produces physically plausible airfoil geometries from latent representations.The experimental results validate the framework’s ability to generate aerodynamically plausible airfoils from natural language specifications and to classify airfoils accurately based on given textual labels.This research reduces the expertise threshold for preliminary airfoil design and highlights the potential for human-AI collaboration in aerospace engineering.
文摘Generative artificial intelligence(Generative AI)is reshaping both learning and teaching paradigms in medical education.With the advancement of Large Language Models(LLMs)-based tools such as ChatGPT,Gemini,and other medical-domain-specific models,Generative AI shows strong potential to address persistent challenges in medical education,including rigid curricula,unequal access to educational resources,and the diverse learning needs of medical students.This review summarizes the applications of Generative AI across key domains:(1)personalized learning through real-time analysis of student performance;(2)clinical skills training via immersive simulations and virtual patients;(3)automated generation of teaching materials such as clinical cases and assessments;and(4)support for student research and academic writing.Empirical evidence indicates that Generative AI-enhanced instruction can improve knowledge acquisition,clinical reasoning,and overall educational efficiency.However,challenges remain,including the generation of inaccurate or fabricated content,risks to academic integrity,algorithmic bias,data privacy concerns,and unresolved ethical issues regarding AI’s role in teaching.Without proper oversight,these risks may compromise educational quality and equity.To ensure responsible adoption,this review advocates for the establishment of institutional policies,enhancement of educators’AI literacy,transparent model validation,and a human-centered design framework that positions Generative AI as a collaborative teaching assistant.When responsibly integrated,Generative AI holds the transformative potential to cultivate future medical professionals equipped with clinical competence,responsibility,and innovative thinking.
