.jpg)
The discovery of new materials is one of the fields where artificial intelligence is generating the most profound impact. Traditionally, finding a compound with optimal properties involves a costly cycle of proposal, synthesis, and experimental characterization. However, approaches such as Bayesian design of materials with surrogate gate and embeddings are revolutionizing this process by optimizing the use of computational resources. Instead of evaluating each candidate with costly methods, a probabilistic surrogate model —for example, a Gaussian process— is introduced, acting as an intelligent filter between the structure generator and the property oracle. This strategy allows prioritizing the most promising samples, drastically reducing the number of calls to the final evaluator without sacrificing accuracy.
Recent research shows that, with only a fixed budget of four evaluations per cycle, a ranking system based on the surrogate matches or exceeds the performance of a full fine-tuning of the generative model. Furthermore, by using pre-trained embeddings such as ORB combined with Gaussian processes, very high reliability in structure ranking is achieved, reaching Spearman correlations above 0.94. This opens the door to autonomous discovery pipelines that can run on cloud infrastructures, scaling without exorbitant costs.
For companies looking to implement this type of solution, having AI for businesses adapted to their needs is key. At Q2BSTUDIO we develop custom applications that integrate AI agents capable of managing complex workflows, from candidate generation to automated validation. Our AWS and Azure cloud services provide the computing power needed to train and deploy surrogate models without worrying about infrastructure management. Additionally, we combine these capabilities with business intelligence services such as Power BI to visualize pipeline results and make informed decisions.
Cybersecurity also plays a fundamental role when handling sensitive research data or intellectual property. That is why we offer robust protection solutions at every layer of the system. Bayesian design of materials is just one example of how artificial intelligence, when implemented with AWS and Azure cloud services and custom software, can accelerate innovation in sectors such as energy, electronics, or pharmaceuticals. At Q2BSTUDIO we help companies build these ecosystems from consulting to production deployment, making the most of the potential of generative models and Bayesian surrogates.
