The field of molecular design and analysis is witnessing a significant shift towards leveraging machine learning and deep learning techniques to accelerate the discovery of tailored molecules. Researchers are exploring novel vector embeddings, transformer-based architectures, and modified Generative Adversarial Networks (GANs) to generate molecules with desired properties. Additionally, visual fingerprinting of chemical structures and natural language processing tools are being developed to extract valuable information from scientific literature and patents. These advances are paving the way for improved molecular odor prediction, drug discovery, and materials science applications. Noteworthy papers include:

• Improved Molecular Generation through Attribute-Driven Integrative Embeddings and GAN Selectivity, which introduces a transformer-based vector embedding generator combined with a modified GAN to generate molecules with desired properties.
• SubGrapher: Visual Fingerprinting of Chemical Structures, which proposes a method for visual fingerprinting of chemical structure images using learning-based instance segmentation.
• Multi-Hierarchical Fine-Grained Feature Mapping Driven by Feature Contribution for Molecular Odor Prediction, which proposes a Feature Contribution-driven Hierarchical Multi-Feature Mapping Network (HMFNet) for molecular odor prediction.