Search for: All records

Training with an emphasis on “hard-to-learn” components of the data has been proven as an effective method to improve the generalization of machine learning models, especially in the settings where robustness (e.g., generalization across distributions) is valued. Existing literature discussing this “hard-to-learn” concept are mainly expanded either along the dimension of the samples or the dimension of the features. In this paper, we aim to introduce a simple view merging these two dimensions, leading to a new, simple yet effective, heuristic to train machine learning models by emphasizing the worst-cases on both the sample and the feature dimensions. We name our method W2D following the concept of “Worst-case along Two Dimensions”. We validate the idea and demonstrate its empirical strength over standard benchmarks. 

Multiferroic materials are an interesting functional material family combining two ferroic orderings, e.g. , ferroelectric and ferromagnetic orderings, or ferroelectric and antiferromagnetic orderings, and find various device applications, such as spintronics, multiferroic tunnel junctions, etc. Coupling multiferroic materials with plasmonic nanostructures offers great potential for optical-based switching in these devices. Here, we report a novel nanocomposite system consisting of layered Bi 1.25 AlMnO 3.25 (BAMO) as a multiferroic matrix and well dispersed plasmonic Au nanoparticles (NPs) and demonstrate that the Au nanoparticle morphology and the nanocomposite properties can be effectively tuned. Specifically, the Au particle size can be tuned from 6.82 nm to 31.59 nm and the 6.82 nm one presents the optimum ferroelectric and ferromagnetic properties and plasmonic properties. Besides the room temperature multiferroic properties, the BAMO-Au nanocomposite system presents other unique functionalities including localized surface plasmon resonance (LSPR), hyperbolicity in the visible region, and magneto-optical coupling, which can all be effectively tailored through morphology tuning. This study demonstrates the feasibility of coupling single phase multiferroic oxides with plasmonic metals for complex nanocomposite designs towards optically switchable spintronics and other memory devices. 

Machine learning has demonstrated remarkable prediction accuracy over i.i.d data, but the accuracy often drops when tested with data from another distribution. In this paper, we aim to offer another view of this problem in a perspective assuming the reason behind this accuracy drop is the reliance of models on the features that are not aligned well with how a data annotator considers similar across these two datasets. We refer to these features as misaligned features. We extend the conventional generalization error bound to a new one for this setup with the knowledge of how the misaligned features are associated with the label. Our analysis offers a set of techniques for this problem, and these techniques are naturally linked to many previous methods in robust machine learning literature. We also compared the empirical strength of these methods demonstrated the performance when these previous techniques are combined, with implementation available here