1. AI for scientific discovery
WEF Emerging Technology 2024
While artificial intelligence (AI) has been used in research for many years, advances in deep learning, generative AI and foundation models are revolutionizing the scientific discovery process. AI will enable researchers to make unprecedented connections and advancements in understanding diseases, proposing new materials, and enhancing knowledge of the human body and mind.
Competitive Environment in AI for scientific discovery
Companies and universities with conceptually close patents to the technology definition
Countries in technology
(number of active patents in technology in 2024 country by inventor address)
Development of patent publications
(publications per year)
EconSight comment and short analysis
The competitive environment shows a balanced picture between large companies, small specialists and universities. With regard to their technological focus, it can be seen that the technology is essentially moving in two directions: pharma and industry/material. These are also offset in time. The material applications have already been patented earlier, i.e. the application of AI for the discovery of new materials is also more visible in companies. Automotive companies such as Volkswagen and Toyota are leading the way here, as are small players such as Automat Solutions, which is heavily involved in the development of new battery materials. Others, such as Resonac and Hitachi, are developing new material design systems that can be used to develop a whole range of new materials.
Of the large pharmaceutical companies, only Roche is active in the field of patents. In the case of small companies, on the other hand, the pharmaceutical focus is clearly visible.
Overall, around 25% of all patents in this technology come from basic research at universities and research institutions. The Korean Jeju National University is particularly close to the technological centre of this technology, using reverse learning methods to predict general material properties in order to develop a material recommendation system.
Background
- Textual concepts are generated for each WEF technology.
- The concepts are applied to our full-text AI-RAG (retrieval augmented generation system) which is optimised for highest precision patent analysis to identify the semantically close patents for each technology.
- A competitive environment with the most relevant companies and research institutions is developed where the relevance is calculated as the similarity of their patents compared to the technology concept. The closer to the technology core, the higher the similarity. Large corporates, small specialists and research institutions are shown separately.
AI artificial intelligence, machine learning framework utilizing deep learning models to predict disease progression and outcomes. The framework integrates multimodal data sources, including genetic, clinical, and environmental datasets, to train robust predictive models. These models are designed to identify potential disease markers and simulate disease pathways, offering insights that are critical for early diagnosis and personalized treatment strategies. The system employs reinforcement learning techniques to adapt and optimize treatment recommendations based on real-time patient data, significantly improving the accuracy and effectiveness of medical interventions.Your Content Goes Here
Generative AI system for the discovery and design of new materials. The system uses a foundation model trained on vast datasets of material properties and synthesis processes to generate candidates for high-performance materials with desired characteristics. The AI, artificial intelligence, machine learning leverages a combination of supervised and unsupervised learning techniques to predict material behaviors and properties under various conditions. This technology is particularly useful in developing next-generation batteries, superconductors, and polymers, with applications ranging from energy storage to electronics.
Revolutionary approach to understanding human physiology and disorders using foundation models in AI, artificial intelligence, machine learning. The system integrates data from genomics, proteomics, and metabolomics with patient health records to create detailed models of human biological processes. By applying techniques from transfer learning and neural network interpretability, the models provide novel insights into complex biological mechanisms, potentially leading to breakthroughs in the understanding and treatment of diseases. Additionally, the system includes a visualization tool that helps researchers and clinicians visualize and interpret model predictions, enhancing the decision-making process in medical and research settings.
AI, artificial intelligence, machine learning for scientific discovery: While artificial intelligence (AI) has been used in research for many years, advances in deep learning, generative AI and foundation models are revolutionizing the scientific discovery process.
Further information on our analysis approach and how we identify the most exciting startups and newcomers in highly specialised technology domains and evaluate them for private equity and venture capital can be found in our Precision Investing approach.
