The study of complex oxides through combinatorial efforts would require a prohibitive amount of material resources and time if experiments were conducted using bulk techniques and at macroscopic length scales. The work in this thesis details the application of leading-edge materials processing in conjunction with advanced micron-scale synchrotron characterization to enable high-throughput studies in the search for metastable materials. Lateral-gradient laser spike annealing is used to rapidly heat and quench millimeter scale samples of thin films of metal oxides. This technique generates a multitude of locations with specific processing histories in a single experiment that are assayed by optical microscopy, micron-scale reflectance spectroscopy, and micron-scale x-ray diffraction. These analytical techniques offer a hierarchy of information on the transformed thin film with a corresponding increase in complexity and accessibility (cost). It is shown how the spatially resolved techniques can be used to construct processing phase maps denoting important phase fields and transformation boundaries as a function of the processing time and temperature, analogous to the canonical time-temperature-transformation diagrams. Two optical methods are demonstrated as quantifiably relevant proxies for identifying crystallographic transformations as validated by x-ray diffraction. This is demonstrated by cross correlating and comparing observed transformations identified using the three analytical techniques as applied to amorphous thin films of Ga2O3 and a pseudo-binary composition spread of amorphous La2O3-Mn3O4, in both cases spanning a range of processing times and temperatures. Exploration of these material spaces is improved upon by incorporating artificial intelligence and machine learning methods with the development of SARA, the Scientific Autonomous Reasoning Agent. The first iteration of SARA aims to autonomously and rapidly construct these diagrams from readily available optical methods. This is achieved through the use of experimentally informed and intelligently designed Gaussian process regression models. The performance of this learning method is validated by comparison with an exhaustively explored phase map of the Bi2O3 system. Finally, a systematic study of the laser spike annealed transformations in amorphous SnO2 and amorphous Sn2O3 is conducted. Each film comprises regions capped with and without an Al2O3 oxygen passivation layer in an attempt to disentangle the competing non-equilibrium effects of laser spike annealing under atmospheric conditions and initial oxygen composition. Rapid phase identification and quantification algorithms applied to the spatially resolved x-ray diffraction data are validated in preparation for the next generation of autonomous experimentation, in which exploration and learning are derived from the underlying crystallographic information. There are transformations in both of these chemical systems that reveal the ability to produce metastable materials even in extensively explored and highly engineered systems. Notable are the retention of the litharge-type SnO compound, even when processing temperature exceeds the known decomposition temperature of the amorphous Sn2O3 film, and the observation of a previously unreported nucleating precursor phase from the amorphous SnO2 film.