Abstract: Devices and techniques are generally described for cross-domain machine learning. A first machine learning model may be trained using first data of a first domain. Predictions may be generated by inputting a plurality of domain data from other domains apart from the first domain into the first machine learning model. For each of the predictions, a prediction error may be determined. A grouping of similar domains from among the other domains may be determined based on the prediction errors. A second machine learning model may be trained for the grouping of similar domains.

Abstract: A method and apparatus for using deep learning in label-free cell classification and machine vision extraction of particles. A time stretch quantitative phase imaging (TS-QPI) system is described which provides high-throughput quantitative imaging, and utilizing photonic time stretching. In at least one embodiment, TS-QPI is integrated with deep learning to achieve record high accuracies in label-free cell classification. The system captures quantitative optical phase and intensity images and extracts multiple biophysical features of individual cells. These biophysical measurements form a hyperdimensional feature space in which supervised learning is performed for cell classification. The system is particularly well suited for data-driven phenotypic diagnosis and improved understanding of heterogeneous gene expression in cells.

Abstract: A method and apparatus for using deep learning in label-free cell classification and machine vision extraction of particles. A time stretch quantitative phase imaging (TS-QPI) system is described which provides high-throughput quantitative imaging, and utilizing photonic time stretching. In at least one embodiment, TS-QPI is integrated with deep learning to achieve record high accuracies in label-free cell classification. The system captures quantitative optical phase and intensity images and extracts multiple biophysical features of individual cells. These biophysical measurements form a hyperdimensional feature space in which supervised learning is performed for cell classification. The system is particularly well suited for data-driven phenotypic diagnosis and improved understanding of heterogeneous gene expression in cells.

Abstract: Aspects of the present disclosure provide a magnetic reader and methods for fabricating the same. The magnetic reader has a capping layer structure that can reduce or impede the corrosion and/or recession of a shield layer of the magnetic reader. In a particular embodiment, the capping layer structure includes a ruthenium (Ru) layer that is configured to impede oxygen interdiffusion between an IrMn antiferromagnetic layer and a Ta cap layer.

Abstract: A method for characterizing a carbon overcoat is provided. The method includes performing electron energy loss spectroscopy (EELS) spectrum imaging for an area of a preselected carbon-based material and an area of the carbon overcoat to generate a reference EELS dataset and a carbon overcoat EELS dataset, respectively, and determining a carbon bonding content of the carbon overcoat based on the reference EELS dataset and the carbon overcoat EELS dataset.

Abstract: Methods for characterizing relative film density using spectroscopic analysis at the device level are provided. One such method includes obtaining a composition of materials at preselected areas of a workpiece using energy dispersive X-ray spectroscopy, obtaining an electron energy loss spectrum-imaging data at each of the preselected areas using electron energy loss spectroscopy, removing, for each of the preselected areas, a preselected noise component of the electron energy spectrum-imaging data to form a plasmon energy spectrum-imaging data, generating, for each of the preselected areas, a plasmon energy map based on the respective plasmon energy spectrum-imaging data, determining, for each of the preselected areas, an average plasmon energy value from the respective plasmon energy map, and calculating a relative mass density of the preselected areas based on the average plasmon energy value, a number of valence electrons per molecule, and a molecular weight for each of the preselected areas.

Abstract: Magnetoresistive (MR) sensors are disclosed that have leads that overlap a MR structure and distribute current to and from the MR structure so that the current is not concentrated in small portions of the leads, alleviating the problems mentioned above. For example, the leads can be formed of a body-centered cubic (bcc) form of tantalum, combined with gold or other highly conductive materials. For the situation in which a thicker bcc tantalum layer covers a highly conductive gold layer, the tantalum layer protects the gold layer during MR structure etching, so that the leads can have broad layers of electrically conductive material for connection to MR structures. The broad leads also conduct heat better than the read gap material that they replace, further reducing the temperature at the connection between the leads and the MR structure.

Abstract: A method and system for providing a magnetoresistive sensor and a read head that includes the magnetoresistive sensor is disclosed. The method and system include providing a pinned layer, a nonmagnetic spacer layer and a composite sensor layer. The pinned layer has a first magnetization that is pinned in a particular direction. The nonmagnetic spacer layer resides between the composite sensor layer and the pinned layer. The composite sensor layer includes a CoFe layer and a composite layer adjacent to the CoFe layer. The composite layer includes CoFe and at least one of Ta, Hf, Ti, Nb, Zr, Au, Ag, Cu, B, C, O2, H2 and N2.