Dynamic feature fusion for lightweight material recommendation in CAD assemblies

Learning ordinality-aware multimodal representations for composite materials design

This work builds ORDER, a multimodal framework linking microstructures and descriptors with preserved property trends, aiding prediction, retrieval, and microstructure generation and consistently outperforms alignment-oriented and property-aware baselines across property prediction, cross-modal retrieval, and microstructure generation tasks.

User Demand Prediction and Solution Generation in Textile Product Development and Design Based on Machine Learning

Accurate translation of heterogeneous user requirements into manufacturable design solutions remains a challenging task in intelligent product engineering systems. This study proposes a process-constrained multimodal conditional variational autoencoder (PC-MCVAE) framework for demand representation, solution generation, and engineering feasibility optimization. The proposed architecture integrates multimodal information from textual descriptions and visual inputs through contrastive representation learning to construct a unified latent demand space. To incorporate domain knowledge into the generation process, a process knowledge graph is transformed into differentiable constraint functions that characterize feasible manufacturing regions and are embedded directly into the optimization objective. A conditional variational decoder is then employed to generate hybrid design solutions consisting of discrete structural representations and continuous engineering parameters. Furthermore, an end-to-end training strategy is developed to jointly optimize semantic consistency and process compliance. Experimental results demonstrate that the proposed framework achieves a semantic alignment score of 0.782, a process compliance rate of 91.4%, and a Top-3 user preference prediction accuracy of 83.7%. The model exhibits strong robustness across different product categories and application scenarios while maintaining high generation quality and manufacturing feasibility. The proposed framework provides an effective methodology for multimodal information fusion, knowledge-guided generative modeling, and intelligent decision support in engineering-oriented design systems.

Lightweight Multimodal-Guided Diffusion Transformer for Agricultural Product Packaging Visual Concept Generation

To address the problems of high costs during the conceptual design stage of agricultural product packaging and weak correlation between visual styles and market feedback, this study proposes a Lightweight Multimodal-Guided Diffusion Transformer (LMG-DiT). Driven by the Amazon Product Dataset (APD), the model first adopts a Multimodal Large Language Model (MLLM) to extract design-oriented market instructions from massive commercial data. Subsequently, it relies on the Diffusion Transformer (DiT) architecture to maintain the rational spatial layout of packaging pages. On this basis, two lightweight strategies, namely Latent Consistency Distillation (LCD) and 4-bit Integer quantization (INT4), are integrated to enable efficient operation on terminal devices with limited computing resources. Experimental results demonstrate that the visual concepts of agricultural product packaging generated by LMG-DiT achieve outstanding performance across multiple evaluation metrics, with a Fréchet Inception Distance (FID) of 12.51, an Inception Score (IS) of 16.85, and an average inference time of merely 1.5 seconds per solution. Compared with open-source baselines including Stable Diffusion v1.5, Vanilla DiT-XL/2, and Layout-Transformer, the proposed model presents distinct advantages in visual authenticity, commercial semantic matching, and operational efficiency. Overall, this method provides an intelligent solution that balances efficiency, cost control and market adaptability for the visual branding construction of agricultural products in resource-constrained scenarios.

Machine learning for predictive modeling and process optimization in fused deposition modeling of polymer parts: a systematic review

This article presents how ML approaches may speed up material discovery, reduce trial-and-error, and enable individualized design solutions for 3D-printed polymers across a variety of contexts.

MBHS-DETR: a lightweight collaborative network for bearing surface defect detection

Multimodal Machine Learning for Data-Driven Materials Selection: A Review of Foundations, Advances, and Future Directions

This review aims to offer a conceptual framework and a solid reference for building intelligent multimodal material selection systems and presents a forward-looking research agenda covering self-supervised learning, knowledge-enhanced models, and interpretable human–AI collaboration.