Recent development of soft materials for medical, pharmaceutical, and bio-inspired devices has produced and accumulated enormous substances concerning spectra and images of target materials. An effective design and development on these fields would require the utilization of scientific assets. Data-driven approach would play the important role to give the novel design strategy of soft interfacial materials.

Data composed of images and spectra generally contain rich and complex information, and therefore the interpretation of the data is crucial to extract essential information of samples. Multivariate analysis such as principal component analysis has been applied to such data to interpret complicated sample data. Recently, machine learning and deep learning methods have been employed in a variety of fields. In this article, the features of machine learning and deep learning methods in terms of analysis of image and spectrum data of soft materials obtained by TOF-SIMS are described by comparing those of multivariate analysis. And, the future development of machine learning and deep learning application to surface analysis data are discussed.

In this review, we summarize the current situation of materials informatics in the field of biomaterials. Different from other fields working on solid state materials, the functions of biomaterials are difficult to predict by using theoretical or computational approaches. Therefore, we need to collect data experimentally for the construction of dataset. We introduce our recent achievements on the prediction of protein adsorption onto self-assembled monolayers (SAMs) and polymer blush films. In addition, we describe the procedure of the data acquisition, selection of descriptors and algorithms.

Data-driven analytics can provide useful information derived from the images of scanning probe microscopy for biological systems including large variety of species and complicated interactions. Bayesian inference is applied to correlation analysis for height distribution of atomic force microscopy images on SasA-KaiC complex formation involved in information transfer from circadian clock oscillator to the output pathway in cyanobacteria. Principal component analysis is applied to the ion images constituted from mass spectra obtained by tapping-mode scanning probe electrospray ionization.

Understandings of protein adsorption and cell adhesion on the surface of materials are significant for the development of soft interface on biomaterials. However, it is difficult to understand completely the behavior of protein adsorption and cell adhesion on the surface, because of lots of interactions, such as chemical, topographical and mechanical. This review explains how to understand the interaction between proteins and/or cells and the material's surface based on the well-defined polymer brush structures. I believe that the method is the better way to develop advanced biomaterials by use of data from the experiment.

This article introduces two topics on surface mediated photoinduced effects : photoalignment control of liquid crystals (LCs) and mass transport process in azobenzene-containing polymers. Due to a large cooperative behavior of LCs, the photoisomerization of azobenzene on a substrate surface (command layer) can control the alignment of a large number of LC molecules. Recent work has revealed that the command layer can be also set at the free surface for LC polymer films. On the other hand, the photoinduced mass transport is also a subject of extensive study in azobenzene films. We have recently demonstrated the crucial role of the free surface in the mass transport process. It is expected that optimization of the systems and elucidation of unresolved issues for such soft-interfaces can be made through the data-driven approaches.