Search

Deep Learning based models are currently dominating most state-of-the-art solutions for disease prediction. Existing works employ RNNs along with multiple levels of attention mechanisms to provide interpretability. These deep learning models, with trainable parameters running into millions, require huge amounts of compute and data to train and deploy. These requirements are sometimes so huge that they render usage of such models as unfeasible. We address these challenges by developing a simpler yet interpretable non-deep learning based model for application to EHR data. We model and showcase our work's results on the task of predicting first occurrence of a diagnosis, often overlooked in existing works. We push the capabilities of a tree based model and come up with a strong baseline for more sophisticated models. Its performance shows an improvement over deep learning based solutions (both, with and without the first-occurrence constraint) all the while maintaining interpretability., Comment: 7 pages, 10 figures

Combining multiple emergent correlated properties such as superconductivity and magnetism within the topological matrix can have exceptional consequences in garnering new and exotic physics. Here, we study the topological surface states from a noncentrosymmetric $\alpha$-BiPd superconductor by employing angle-resolved photoemission spectroscopy (ARPES) and first principle calculations. We observe that the Dirac surface states of this system have several interesting and unusual properties, compared to other topological surface states. The surface state is strongly anisotropic and the in-plane Fermi velocity varies rigorously on rotating the crystal about the $y$-axis. Moreover, it acquires an unusual band gap as a function of $k_y$, possibly due to hybridization with bulk bands, detected upon varying the excitation energy. Coexistence of all the functional properties, in addition to the unusual surface state characteristics make this an interesting material., Comment: submitted on 26th April, 2016; to appear in PRL

Using angle-resolved photoemission spectroscopy (ARPES), we studied the effect of the impurity potential on the electronic structure of FeTe$_{0.5}$Se$_{0.5}$ superconductor by substituting 10\% of Ni for Fe which leads to an electron doping of the system. We could resolve three hole pockets near the zone center and an electron pocket near the zone corner in the case of FeTe$_{0.5}$Se$_{0.5}$, whereas only two hole pockets near the zone center and an electron pocket near the zone corner are resolved in the case of Fe$_{0.9}$Ni$_{0.1}$Te$_{0.5}$Se$_{0.5}$, suggesting that the hole pocket having predominantly the $xy$ orbital character is very sensitive to the impurity scattering. Upon electron doping, the size of the hole pockets decrease and the size of the electron pockets increase as compared to the host compound. However, the observed changes in the size of the electron and hole pockets are not consistent with the rigid-band model. Moreover, the effective mass of the hole pockets is reduced near the zone center and of the electron pockets is increased near the zone corner in the doped Fe$_{0.9}$Ni$_{0.1}$Te$_{0.5}$Se$_{0.5}$ as compared to FeTe$_{0.5}$Se$_{0.5}$. We refer these observations to the changes of the spectral function due to the effect of the impurity potential of the dopants., Comment: 8 pages, 3 figures

Observation of ferroelectricity among non-d0 systems, which was believed for a long time an unrealistic concept, led to various proposals for the mechanisms to explain the same (i.e. magnetically induced ferroelectricity) during last decade. Here, we provide support for ferroelectricity of a displacive-type possibly involving magnetic ions due to short-range magnetic correlations within a spin-chain, through the demonstration of magnetoelectric coupling in a Haldane spin-chain compound Er2BaNiO5 well above its N\'eel temperature of (TN=) 32K. There is a distinct evidence for electric polarization setting in near 60 K around which there is an evidence for short-range magnetic correlations from other experimental methods. Raman studies also establish a softening of phonon modes in the same temperature (T) range and T-dependent x-ray diffraction (XRD) patterns also reveal lattice parameters anomalies. Density-functional theory based calculations establish a displacive component (similar to d0-ness) as the root-cause of ferroelectricity from (magnetic) NiO6 chain, thereby offering a new route to search for similar materials near room temperature to enable applications.