Currently, the cosmetics industry is actively embracing AI technology and gradually penetrating it into all aspects of cosmetics research and development. This not only brings about an improvement in research and development efficiency, but also hopes to leverage the creative capabilities of generative AI to achieve disruptive innovation in research and development methods and processes.
In the field of active ingredient discovery, AI has significantly shortened the discovery cycle of innovative ingredients. AI constructs a high-dimensional feature space, integrates multiple sources of data (such as peptide sequences, raw material libraries, omics data, literature, etc.), uses machine learning algorithms to screen potential active substances, and predicts relevant probabilities, thereby obtaining the ranking of candidate innovative raw materials. Compared with traditional high-throughput "wet experiment" screening, AI "dry experiment" not only avoids the long cycle and high cost of building physical raw material libraries, but also compresses the validation scale to one thousandth, greatly reducing costs and shortening the research and development cycle. For example, Nuritas' AI technology platform in the United States utilizes convolutional neural networks to analyze millions of plant derived peptides, generate a database containing millions of peptides, and extract activity data from literature through NLP, discovering two new components, PeptiYouth and PeptiStrong; The KEPLER 90i platform of Tsinghua Yangtze River Delta Research Institute has discovered a new anti-aging peptide from cells with self repairing function in the human body, and successfully obtained the innovative raw material of stereopeptide EQ9.
In the field of efficacy target discovery, AI can not only explore new targets through biological networks and signaling pathways, but also analyze and lock in the mechanism of action of efficacy raw materials based on existing target libraries. In terms of discovering new targets, AI can use graph neural networks (GNNs) to analyze protein-protein interaction networks, key signaling pathways, and multi omics data related to regulation (transcriptome, proteome) to identify potential targets related to skin efficacy. Researchers at Nanchang University found a high correlation between genes related to fatty acid metabolism, energy generation, and inflammation regulation through GO annotation and KEGG pathway enrichment studies related to skin aging; By combining the billion level biological peptide database with Transformer models and NLP technology, it is possible to quickly query the gene codes of active substances, mine targets related to photoaging pathways, and improve the efficiency of target discovery.
In the field of formula design, cosmetics companies are trying to use AI technology to optimize product experience, shorten product formula and prototype development cycles. AI can vectorize formula data and predict synergistic effects of ingredients through deep learning algorithms; Meanwhile, AI can quickly analyze consumer sentiment data on social media and predict the performance of innovative formulas in terms of stability, safety, efficacy, and user experience. For example, Potion AI in the United States can not only search and manage raw materials and suppliers, but also analyze the formula ingredients of efficacy cosmetics through AI and predict their content, generate innovative formulas, and combine them with regulatory standards around the world to improve formula compliance.