Investigation of out-of-distribution detection across various models and training methodologies.

Machine learning-based algorithms demonstrate impressive performance across numerous fields; however, they continue to suffer from certain limitations. Even sophisticated and precise algorithms often make erroneous predictions when implemented with datasets having different distributions compared to the training set. Out-of-distribution (OOD) detection, which distinguishes data with different distributions from that of the training set, is a critical research area necessary to overcome these limitations and create more reliable algorithms. The OOD issue, particularly concerning image data, has been extensively studied. However, recently developed OOD methods do not fulfill the expectation that OOD performance will increase as the accuracy of in-distribution classification improves. Our research presents a comprehensive study on OOD detection performance across multiple models and training methodologies to verify this phenomenon. Specifically, we explore various pre-trained models popular in the computer vision field with both old and new OOD detection methods. The experimental results highlight the performance disparity in existing OOD methods. Based on these observations, we introduce Trimmed Rank with Inverse softMax probability (TRIM), a remarkably simple yet effective method for model weights with newly developed training methods. The proposed method could serve as a potential tool for enhancing OOD detection performance owing to its promising results. The OOD performance of TRIM is highly compatible with the in-distribution accuracy model and may bridge the efforts on improving in-distribution accuracy to the ability to distinguish OOD data.

Adjusting microwave sensing frequency through aspect ratio variation and bending repetitions in Permalloy ellipses.

The Ferromagnetic Resonance (FMR) phenomenon, marked by the selective absorption of microwave radiation by magnetic materials in the presence of a magnetic field, plays a pivotal role in the development of radar absorbing materials, high speed magnetic storage, and magnetic sensors. This process is integral for technologies requiring precise control over microwave absorption frequencies. We explored how variations in resonance fields can be effectively modulated by adjusting both the shape and stress anisotropies of magnetic materials on a flexible substrate. Utilizing polyethylene-naphthalate (PEN) as the substrate and Permalloy (Ni79Fe21, noted for its positive magnetostriction coefficient) as the magnetic component, we demonstrated that modifications in the aspect ratio and bending repetitions can significantly alter the resonance field. The results, consistent with Kittel's equation and the predictions of a uniaxial magnetic anisotropy model, underscore the potential for flexible substrates in enhancing the sensitivity and versatility of RF-based magnetic devices.

Tunable asymmetric spin wave excitation and propagation in a magnetic system with two rectangular blocks.

Asymmetric spin wave excitation and propagation are key properties to develop spin-based electronics, such as magnetic memory, spin information and logic devices. To date, such nonreciprocal effects cannot be manipulated in a system because of the geometrical magnetic configuration, while large values of asymmetry ratio are achieved. In this study, we suggest a new magnetic system with two blocks, in which the asymmetric intensity ratio can be changed between 0.276 and 1.43 by adjusting the excitation frequency between 7.8 GHz and 9.4 GHz. Because the two blocks have different widths, they have their own spin wave excitation frequency ranges. Indeed, the spin wave intensities in the two blocks, detected by the Brillouin light scattering spectrum, were observed to be frequency-dependent, yielding tuneable asymmetry ratio. Thus, this study provides a new path to enhance the application of spin waves in spin-based electronics.