Experimental atmosphere and analysis course of
Appendices 1 and 2 current the experimental atmosphere utilized in this examine, detailing the {hardware} and software program configurations. The parameter settings for the VAE and RL fashions are supplied in Appendices 3 and 4.
The experiment evaluates the efficiency of the proposed VAE and RL mixture in cultural and creative product design utilizing a number of metrics, together with mannequin accuracy, technology high quality, consumer satisfaction, and computational effectivity. These outcomes are in contrast with these of conventional single fashions, which haven’t built-in a number of strategies or strategies. These single fashions embrace basic approaches resembling GANs, VAEs, and RL.
In evaluating the standard of the generated designs, each quantitative and qualitative measures are utilized. Quantitatively, the Structural Similarity Index (SSIM) is employed to evaluate the similarity between AI-generated designs and reference designs from historic information, offering an goal measure of output accuracy and high quality. SSIM evaluates picture similarity primarily based on brightness, distinction, and structural options. Moreover, Distinction Constancy, Texture Constancy, and Colour Constancy are used to evaluate visible distinction, texture element, and colour copy accuracy, respectively, by evaluating the generated designs to real-world counterparts. Texture Constancy might be analyzed by extracting texture options, resembling utilizing Gabor filters, and calculating the similarity between these extracted options. The SSIM formulation is expressed in Eq. (4):
$$SSIMleft( {x,y} proper)=frac{{left( {2{mu _x}{mu _y}+{C_1}} proper)left( {2{sigma _{xy}}+{C_2}} proper)}}{{left( {mu _{x}^{2}+mu _{y}^{2}+{C_1}} proper)left( {sigma _{x}^{2}+sigma _{y}^{2}+{C_2}} proper)}}$$
(4)
In Eq. (4), ({mu _x}) and ({mu _y}) are the means of photographs x and y, (sigma _{x}^{2}) and (sigma _{y}^{2}) are their variances, ({sigma _{xy}}) represents the covariance between the 2 photographs, and ({C_1}) and ({C_2}) are constants to stop division by zero. Distinction Constancy, which assesses how nicely the distinction is preserved in the generated picture, is outlined as Eq. (5):
$${textual content{Distinction~Constancy}}=frac{{{textual content{st}}{{textual content{d}}_{{textual content{generated}}}}}}{{{textual content{st}}{{textual content{d}}_{{textual content{reference}}}}}}$$
(5)
In Eq. (5), ({textual content{st}}{{textual content{d}}_{{textual content{generated}}}}) is the usual deviation of the generated picture, and ({textual content{st}}{{textual content{d}}_{{textual content{reference}}}}) is the usual deviation of the reference picture. A Distinction Constancy worth near 1 signifies sturdy distinction preservation.
Colour Constancy is assessed utilizing the CIEDE2000 colour distinction formulation, as proven in Eq. (6).
$${textual content{varvec{Delta}}}{E_{00}}=sqrt {{{(L_{1}^{*} – L_{2}^{*})}^2}+{{(a_{1}^{*} – a_{2}^{*})}^2}+{{(b_{1}^{*} – b_{2}^{*})}^2}}$$
(6)
In Eq. (6), L∗, a∗, and b∗ signify the colour parts in the CIE Lab colour house.
For the qualitative evaluation, a complete set of analysis standards is used to make sure the design works meet excessive requirements. These standards embrace innovation, aesthetic enchantment, cultural relevance, cultural adaptability, and design practicality. Innovation refers back to the uniqueness and creativity of the design. Consultants evaluated the extent of innovation utilizing a scoring rubric, using a scale from 1 to 10, the place 1 signifies a scarcity of innovation and 10 represents excessive creativity. The scale is outlined as follows. For instance: 1–3 factors point out a scarcity of innovation, with designs intently resembling frequent, repetitive patterns; 4–6 factors mirror some innovation however inadequate distinction; 7–8 factors display sturdy innovation with notable uniqueness; and 9–10 factors signify extremely modern designs showcasing ground-breaking concepts.
Aesthetic enchantment focuses on the visible impression of the design, together with parts resembling colour combos, shapes, and compositions. Each customers and consultants price the designs primarily based on their aesthetic high quality, guaranteeing objectivity and consistency in the analysis course of. Cultural relevance assesses how nicely the design aligns with a particular tradition. Consultants rating designs primarily based on their effectiveness in cultural communication, guaranteeing that the designs not solely have aesthetic enchantment but additionally resonate with cultural significance. Cultural adaptability evaluates how successfully the design integrates into a particular cultural context. Consultants rating primarily based on the cultural parts embodied in the design, guaranteeing its relevance to the goal tradition. Lastly, design practicality issues the feasibility of the design in sensible functions, assessing performance, usability, and market acceptance.
To assemble professional suggestions, professionals from varied fields throughout the cultural and creative industries—resembling designers, cultural researchers, and market analysts—are invited to take part. Previous to the analysis, consultants bear coaching on the analysis standards and scoring system, guaranteeing they totally perceive the necessities and that means behind every indicator. Consultants use standardized overview varieties to attain the designs, guaranteeing a scientific and constant analysis course of. Concurrently, consumer suggestions is collected by means of structured questionnaires, protecting points resembling visible enchantment, cultural communication effectiveness, and design practicality. Customers price every facet utilizing a Likert scale, supplemented with open-ended questions for deeper insights (e.g., 1 indicating sturdy disagreement and 5 indicating sturdy settlement).
Through the evaluation part, the collected scoring information are subjected to statistical evaluation to compute the imply and commonplace deviation for every criterion. Evaluation of variance (ANOVA) is used to evaluate vital variations between the assorted designs. Moreover, thematic evaluation is carried out on customers’ open-ended suggestions to establish key patterns and themes, offering worthwhile insights. This analysis framework helps to raised perceive design efficiency throughout the cultural and creative industries, providing theoretical foundations and sensible suggestions for future design practices.
Person interplay information, resembling actions like viewing, deciding on, and modifying designs, offers insights into consumer engagement with the designs. For computational effectivity, the main focus is on assessing the runtime and useful resource consumption of the mannequin through the design technology course of. Time measurements for every mannequin’s design technology are recorded underneath equivalent {hardware} circumstances, permitting for a comparative evaluation of computational effectivity throughout completely different fashions. This information is then in contrast with the efficiency of conventional single fashions, offering an intensive analysis of the benefits of combining VAE and RL fashions in the creation of cultural and creative merchandise.
Efficiency analysis
Mannequin technology outcomes
The outcomes of a number of design instances generated by the proposed mannequin are introduced in Fig. 5.
These examples display that the generated designs efficiently retain the unique design parts whereas integrating modern options to boost their creativity and imaginative enchantment. As an illustration, in Fig. 5a, the Jingdezhen ceramics incorporate extra creative patterns and flowing parts impressed by the unique museum assortment, ensuing in a extra dynamic and visually partaking interpretation of the porcelain. In Fig. 5b, the textiles reinterpret conventional Qing Dynasty clothes by introducing distinctive colours and intricate embroidery, mixing conventional craftsmanship with trendy stylistic parts to align with up to date aesthetic preferences. Equally, Fig. 5c reimagines an historical chair by incorporating trendy design parts, making a visually modern piece that harmonizes the ingenuity of historical Chinese language design with trendy aesthetic requirements, rendering it extremely interesting to up to date audiences.
Quantitative analysis outcomes of generated merchandise
The analysis outcomes for mannequin accuracy are proven in Fig. 6.
In Fig. 6, the VAE + RL mannequin demonstrates distinctive efficiency throughout all metrics, notably excelling in accuracy and F1 rating in comparison with different fashions. The accuracy of the VAE + RL mannequin reaches 94.5%, considerably surpassing different single fashions resembling VAE (92.3%), RL (88.7%), and GAN (87.0%). These findings underscore the effectiveness of combining VAE with RL in capturing each international and native options throughout advanced design technology duties, ensuing in enhanced general efficiency. Compared, GPT and Llama-3 fashions additionally exhibit commendable efficiency, with accuracies of 90.4% and 91.2%, respectively. Whereas these fashions excel in textual content technology duties, they face slight limitations in dealing with cross-modal duties, notably these requiring seamless integration of visible and textual information, when in comparison with the VAE + RL mannequin. Notably, GPT and Llama-3 fashions obtain recall scores of 91.7% and 90.8%, respectively, reflecting their sturdy capabilities in recognizing design parts. Nevertheless, the VAE + RL mannequin outperforms in design variety and innovation, attaining an F1 rating of 93.4%, additional reinforcing its superiority in phrases of design high quality and consumer satisfaction.
Regardless of the widespread use of the GAN mannequin in varied technology duties, its efficiency in this examine is comparatively suboptimal, attaining an F1 rating of solely 87.5%, which is decrease than that of different fashions. This end result underscores the GAN mannequin’s limitations in managing design complexity and innovation, because it struggles to steadiness technology high quality with design creativity as successfully because the VAE + RL mannequin. These findings spotlight that the VAE + RL mannequin excels not solely in accuracy and recall but additionally outperforms different fashions in general efficiency, notably in producing high-quality, modern design samples. Whereas GPT and Llama-3 fashions showcase distinctive skills in textual content technology, their effectiveness diminishes when addressing duties that demand design innovation and the mixing of multimodal information. This distinction underscores the substantial benefit of combining VAE and RL strategies, which reinforces each the technology and optimization of designs in the cultural and creative product design area.
Determine 7 offers a visible comparability of the generated high quality amongst completely different fashions, illustrating the superior outcomes achieved by the VAE + RL strategy.
Determine 7 offers a comparative evaluation of generative high quality throughout completely different fashions, emphasizing the prevalence of the VAE + RL mannequin throughout all evaluated metrics. The VAE + RL mannequin achieves a SSIM of 0.92, a Distinction Constancy of 0.95, a Texture Constancy of 0.94, and a Digital Colour Constancy of 0.93. These outcomes display that designs generated by the VAE + RL mannequin not solely intently align with real-world structural traits but additionally exhibit distinctive constancy in distinction, texture, and colour. This superior efficiency is attributed to the synergy between the VAE and RL. The VAE successfully captures various latent representations from current designs, whereas RL optimizes these designs iteratively by means of interplay with suggestions mechanisms, producing high-quality outputs.
Compared, GPT and Llama-3 fashions additionally ship sturdy generative efficiency, notably in structural similarity and distinction constancy. GPT achieves an SSIM of 0.89 and a Distinction Constancy of 0.91, whereas Llama-3 information an SSIM of 0.88 and a Distinction Constancy of 0.90. These fashions, though primarily developed for pure language technology, show sturdy generative capabilities that reach to design duties. Nevertheless, their efficiency in Texture Constancy and Digital Colour Constancy is barely decrease, reflecting their restricted specialization in image-based technology in comparison with the VAE + RL strategy.
When used independently, VAE and RL obtain average generative high quality. The VAE information an SSIM of 0.87, a Distinction Constancy of 0.90, a Texture Constancy of 0.88, and a Digital Colour Constancy of 0.85. Equally, RL achieves an SSIM of 0.85, a Distinction Constancy of 0.88, a Texture Constancy of 0.85, and a Digital Colour Constancy of 0.84. Whereas each fashions produce coherent designs individually, their outputs lack the polished high quality achieved by means of their mixed use. In distinction, the GAN mannequin underperforms throughout all metrics, with an SSIM of 0.83, a Distinction Constancy of 0.86, a Texture Constancy of 0.83, and a Digital Colour Constancy of 0.81. Whereas GANs are identified for his or her capacity to generate various designs, they typically face challenges resembling mode collapse and coaching instability, which negatively impression their consistency and general high quality. These outcomes underscore the substantial advantages of integrating VAE and RL, notably for duties demanding high-fidelity outputs and modern design capabilities. The mixture proves to be a sturdy strategy in the area of cultural and creative product design, providing superior efficiency and adaptability in comparison with conventional single-model strategies.
Qualitative analysis outcomes of generated merchandise
Determine 8 illustrates the qualitative analysis scores assigned by consultants and customers for varied design works primarily based on the outlined metrics, with scores starting from 1 (very poor high quality) to 10 (glorious high quality).
The scoring outcomes introduced in Fig. 8 present that:
-
1.
Design A, which incorporates three ceramic merchandise, acquired excessive scores for innovativeness and aesthetic enchantment, with specific emphasis on design practicality, ensuing in a consumer satisfaction price of 95.2%. This design was acknowledged for its capacity to mix conventional ceramic strategies with trendy creative parts, guaranteeing each visible enchantment and performance.
-
2.
Design B, that includes three improved variations of Qing Dynasty clothes, excelled in cultural relevance and cultural adaptability, scoring 9. This rating displays the design’s success in integrating conventional cultural parts with up to date points. Though its aesthetic enchantment was barely decrease than that of Design A, its general efficiency remained sturdy, with a consumer satisfaction price of 90.3%, indicating its effectiveness in mixing cultural heritage with trendy aesthetics.
-
3.
Design C, which incorporates three enhanced furnishings items (chairs), achieved a rating of 9 for each innovativeness and cultural relevance, reflecting the design’s profitable fusion of conventional cultural parts with trendy design innovation. The consumer satisfaction price for this design was 90.7%, underscoring the enchantment of the furnishings in merging design innovation with cultural expression.
The outcomes point out variations in efficiency throughout the completely different design varieties regarding innovativeness, cultural adaptability, and consumer satisfaction. Design A stands out for its sturdy efficiency in practicality and aesthetic enchantment, whereas Design B demonstrates its worth in merging cultural heritage with trendy enhancements. Design C additional highlights the potential of integrating cultural relevance with modern design, providing worthwhile insights into the sector of furnishings design.
Mannequin effectivity analysis
Determine 9 illustrates the useful resource consumption and runtime of completely different fashions.
Determine 9 compares the useful resource consumption and runtime of varied fashions, providing insights into their computational effectivity and suitability for real-time design technology:
-
1.
The VAE + RL mannequin displays distinctive efficiency with a mean coaching time of 5 h and a mean inference time of 0.2 s. This excessive effectivity makes it ultimate for sensible functions the place fast technology of high-quality designs is important, offering a aggressive benefit in environments requiring quick turnaround occasions.
-
2.
The GPT and Llama-3 fashions display sturdy generative capabilities however are comparatively much less environment friendly in phrases of useful resource consumption and runtime. GPT requires a mean of 8 h for coaching and 0.4 s for inference, whereas Llama-3 wants 9 h for coaching and 0.35 s for inference. These longer coaching occasions mirror the upper computational calls for of these fashions, which can restrict their effectiveness in real-time design functions the place velocity and effectivity are essential.
-
3.
The VAE and GAN fashions present average useful resource consumption, with coaching occasions of 6 h and 7 h, respectively, and inference occasions of 0.25 s and 0.3 s. These fashions provide a steadiness between generative high quality and useful resource effectivity, making them appropriate for functions requiring each high-quality outputs and optimized useful resource use.
-
4.
The RL mannequin, whereas environment friendly in inference, has the longest coaching time, requiring 8 h. Its inference time of 0.35 s is corresponding to different fashions, however the prolonged coaching interval is required to optimize design selections successfully.
In conclusion, the VAE + RL mannequin emerges as essentially the most environment friendly and efficient alternative for cultural and creative product design, providing each superior generative high quality and useful resource effectivity. In distinction, the GPT and Llama-3 fashions are extra computationally demanding and could also be higher fitted to situations with ample computational sources or extra stringent high quality necessities. This evaluation highlights the strengths and limitations of every mannequin, serving to information their optimum application in completely different contexts.
Statistical evaluation outcomes
To additional validate these findings, an ANOVA take a look at is carried out, and Desk 1 offers the adjusted efficiency statistics for every mannequin.
The information in Desk 1 clearly demonstrates the superior efficiency of the VAE + RL mannequin in cultural and creative product design. The conventional design strategies, with a imply rating of 72.24, are considerably outperformed by extra trendy approaches, displaying their restricted functionality in addressing advanced and modern design duties. These conventional strategies fall brief in each effectivity and design innovation when in comparison with the newer, data-driven fashions. The VAE mannequin exhibits a big enchancment over conventional strategies, with a imply rating of 86.84, reflecting its capacity to generate modern designs on account of its sturdy generative capabilities. Nevertheless, whereas VAE excels in design creativity, the RL mannequin—scoring 84.33—is especially sturdy in optimizing decision-making processes. Whereas the RL mannequin exhibits promising outcomes, its design innovation falls barely behind the VAE mannequin, as anticipated given its give attention to optimizing relatively than producing designs. The VAE + RL mannequin achieves the very best imply rating of 90.50, underscoring the synergy between the generative energy of VAE and the decision-making capabilities of RL. This mixture not solely fosters increased ranges of design innovation but additionally enhances consumer satisfaction. Moreover, the usual deviation of the VAE + RL mannequin is 2.00, indicating secure efficiency throughout varied design duties, guaranteeing constant high-quality outcomes. The extensive rating vary (from 87.00 to 94.00) additional emphasizes the mannequin’s reliability in delivering superior outcomes throughout differing kinds of cultural and creative designs.
Though the GPT mannequin (imply rating: 89.00) and the Llama-3 mannequin (imply rating: 87.00) carry out nicely in innovation, their strengths are primarily in textual content technology and dealing with advanced design points. With regards to determination optimization and general design processing, each fashions fall behind the VAE + RL mannequin. GPT and Llama-3, although excelling in textual content technology, lack the built-in design optimization capabilities that the VAE + RL mannequin presents. In conclusion, the VAE + RL mannequin stands out as the best strategy for cultural and creative product design, not solely enhancing design innovation but additionally bettering the effectivity and stability of the design course of. This mixture leads in phrases of design high quality, consumer satisfaction, and design optimization, providing worthwhile insights and sturdy technical help for future developments in the sector. In comparison with the person VAE, RL, GPT, and Llama-3 fashions, VAE + RL delivers a extra complete and efficient resolution for cultural and creative product design.
Turning take a look at outcomes
The Turing take a look at carried out to evaluate the intelligence of the VAE + RL mannequin includes figuring out whether or not the designs generated by the mannequin might be perceived as indistinguishable from these created by human designers. This take a look at offers direct perception into the mannequin’s generative capabilities and its capacity to imitate human creativity.
The VAE + RL mannequin is initially employed to generate a collection of design schemes, guaranteeing a broad vary of variety and creativity. These designs integrated varied kinds and themes to simulate real-world cultural and creative design duties. The generated designs are then combined with these created by human designers to type a complete analysis set. To take care of equity in the analysis course of, all designs are anonymized, eradicating any identifiable markers that would point out whether or not the design was generated by the VAE + RL mannequin or by a human designer. A double-blind evaluation is carried out with a panel consisting of 20 design consultants and 20 unusual customers. Every evaluator is supplied with a set of designs, which included each VAE + RL model-generated and human-created designs. Evaluators are tasked with figuring out whether or not every design is created by a human, with their judgments primarily based on components resembling innovation, practicality, and inventive worth. After amassing the evaluations, the proportion of model-generated designs which can be incorrectly recognized as human-created was calculated. The classification outcomes from every evaluator are then aggregated, and a confusion matrix is used, together with accuracy metrics, to evaluate the intelligence degree of the mannequin.
Desk 2 presents the suggestions primarily based on differ lease design varieties and evaluator teams. Every take a look at consists of 20 design samples from each the VAE + RL mannequin and human designers.
Desk 2 illustrates the efficiency variations of the VAE + RL mannequin throughout completely different design varieties. The mannequin’s designs in trendy artwork and digital illustration are sometimes perceived as human-created, demonstrating sturdy efficiency in these areas. In distinction, the VAE + RL mannequin exhibits weaker efficiency in conventional craft and product design, with notably decrease accuracy charges in product design, as evaluated by each consultants and basic customers.
Within the fields of trendy artwork and digital illustration, the VAE + RL mannequin performs notably nicely. For contemporary artwork, consultants acknowledge 12 of the model-generated designs as resembling human creations, ensuing in an accuracy price of 60.0%. Basic customers assess 17 designs with the next accuracy price of 85.0%. Within the digital illustration class, consultants consider 17 designs with an accuracy price of 85.0%, whereas basic customers assess 19 designs, attaining a 95.0% accuracy price. Conversely, the mannequin is much less efficient in conventional craft and product design. For conventional craft, consultants establish 13 designs with an accuracy price of 65.0%, whereas basic customers assess 17 designs, yielding an 85.0% accuracy price. In product design, consultants consider 15 designs with a 75.0% accuracy price, whereas basic customers assess 16 designs with an 80.0% accuracy price. General, design consultants usually obtain increased accuracy charges in comparison with basic customers, reflecting their extra exact analysis of the designs. Basic customers carry out higher in trendy artwork and digital illustration however display decrease accuracy charges in conventional craft and product design. This discrepancy may very well be attributed to variations in design expertise and sensitivity to particulars.
Dialogue
The excellent efficiency of the VAE + RL mannequin in design high quality and consumer satisfaction underscores its effectiveness in producing high-quality and engaging product designs. Jang et al.29 spotlight that combining generative fashions with determination optimization strategies can considerably enhance the standard and variety of design options. This discovering aligns with that viewpoint, as each SSIM and consumer satisfaction indicators had been increased than these of different fashions. Moreover, the excessive score of design innovation by customers emphasizes the essential function of innovation in the cultural and creative industries50, reflecting Chen51 who identifies innovation as a key driver of success in cultural merchandise. Regardless of the VAE + RL mannequin’s spectacular efficiency, its useful resource consumption and runtime limitations are necessary concerns. Zhan et al.52 be aware that extremely advanced mannequin optimization may result in vital will increase in computational prices. This implies that whereas the mannequin excels in design high quality and consumer satisfaction, its computational calls for might hinder its broader adoption, notably in resource-constrained environments.
The benefits of the VAE + RL mannequin transcend conventional analysis metrics, demonstrating adaptability to market calls for. As Vuong and Mai53 assert, integrating generative fashions with determination optimization extra successfully meets market wants for innovation and personalization. This integration allows the VAE + RL mannequin to provide design options that higher align with consumer expectations, thereby strengthening its aggressive place in the market. Nevertheless, given the mannequin’s excessive computational calls for, additional analysis ought to give attention to optimizing computational effectivity. Methods resembling mannequin pruning and quantization might assist cut back useful resource consumption whereas sustaining core efficiency. Choudhary et al.54 recommend that mannequin compression strategies are efficient in decreasing computational prices and enhancing sensible application effectivity. Moreover, methods like distributed computing and parallel processing might cut back mannequin coaching and inference occasions, bettering computational effectivity. Past the cultural and creative design sector, the design optimization capabilities of the VAE + RL mannequin maintain potential for varied different domains. Functions might prolong to fields like architectural design and product growth. Future analysis ought to discover these cross-domain functions, assess the mannequin’s efficiency in completely different design duties, and develop focused optimization methods to boost its utility.
A comparability with related research underscores the benefits of the proposed analysis methodology and its outcomes. For instance, Liu et al.55 explored the application of AI in the labor market, emphasizing data-driven determination help primarily based on statistical evaluation. In distinction, the present analysis not solely offers data-driven design determination help but additionally generates various design options by means of VAEs and enhances design effectivity and consumer satisfaction by means of RL. Moreover, this strategy is particularly tailor-made to the cultural and creative design area, with the specificity of the application situation and the creativity of the generated options representing key benefits of the mannequin.
Equally, Li et al.56 targeted on AI-supported industrial notion, addressing sensor and information processing challenges in clever manufacturing. Whereas their analysis emphasizes {hardware} integration and industrial optimization, the current examine focuses on optimizing design processes in the cultural and creative sector, considerably bettering design high quality and consumer expertise. As an illustration, consumer satisfaction in this examine reached 95%, whereas current fashions in industrial notion typically overlook consumer suggestions on design options. Moreover, the mannequin addresses gaps in current analysis by incorporating cultural adaptability and variety technology assessments. Moreover, Zhu57 proposed an adaptive agent decision-making mannequin primarily based on deep RL, utilized to determination optimization in the logistics sector. Whereas each research make the most of RL frameworks, the present analysis distinguishes itself by integrating RL with VAEs. This not solely optimizes design selections but additionally leverages generative fashions to boost the range of design options for cultural and creative merchandise. Moreover, the mannequin emphasizes the technology of optimized options primarily based on consumer suggestions, marking a big departure from the logistics give attention to effectivity and path optimization. As such, this strategy is especially fitted to the design innovation area, demonstrating larger adaptability and sensible significance.
The comparative evaluation above highlights the specificity and innovativeness of the proposed analysis methodology in the context of cultural and creative design, providing worthwhile insights for future research in associated fields. This analysis introduces a generative optimization mannequin that mixes VAE and RL, attaining each theoretical and sensible developments in the area of cultural and creative product design. The experimental outcomes and comparative evaluation clearly display the strategy’s superiority in design high quality, variety, and consumer satisfaction. Extra importantly, this examine presents a brand new paradigm for AI-assisted design throughout the cultural and creative industries. Not like conventional design help methods, the proposed strategy not solely facilitates decision-making but additionally generates design options. The integration of generative fashions shifts the design course of from “choice optimization” to “resolution creation,” establishing a basis for tackling extra advanced and diversified design challenges in the longer term. In phrases of application prospects, the mannequin developed in this examine extends past the cultural and creative sector and holds vital cross-domain potential. As an illustration, its capabilities in technology and optimization might be utilized to industrial design, academic content material creation, and different areas, thus increasing the probabilities for AI-driven clever design. This cross-domain adaptability underscores the mannequin’s versatility and offers substantial alternatives for future analysis and growth.