arXiv subjects

Feng Jin

Publications and source records attributed to Feng Jin.

67 records · Page 4Linked to original sources

arXiv2019

Multiple Patients Behavior Detection in Real-time using mmWave Radar and Deep CNNs

To address potential gaps noted in patient monitoring in the hospital, a novel patient behavior detection system using mmWave radar and deep convolution neural network (CNN), which supports the simultaneous recognition of multiple patients' behaviors in real-time, is proposed. In this study, we use an mmWave radar to track multiple patients and detect the scattering point cloud of each one. For each patient, the Doppler pattern of the point cloud over a time period is collected as the behavior signature. A three-layer CNN model is created to classify the behavior for each patient. The tracking and point clouds detection algorithm was also implemented on an mmWave radar hardware platform with an embedded graphics processing unit (GPU) board to collect Doppler pattern and run the CNN model. A training dataset of six types of behavior were collected, over a long duration, to train the model using Adam optimizer with an objective to minimize cross-entropy loss function. Lastly, the system was tested for real-time operation and obtained a very good inference accuracy when predicting each patient's behavior in a two-patient scenario.

eess.SP↗

arXiv2019

Automotive Radar Interference Mitigation Using Adaptive Noise Canceller

Interference among frequency modulated continues wave automotive radars can either increase the noise floor, which occurs in the most cases, or generate a ghost target in rare situations. To address the increment of noise floor due to interference, we proposed a low calculation cost method using adaptive noise canceller to increase the signal-to-interference ratio. In a quadrature receiver, the interference in the positive half of frequency spectrum is correlated to that in the negative half of frequency spectrum, whereas the beat frequencies from real targets are always present in the positive frequency. Thus, we estimated the power of the negative frequency as an indication of interference, and fed the positive frequency and negative frequency components into the primary and reference channel of an adaptive noise canceller, respectively. The least mean square algorithm was used to solve for the optimum filter solution. As a result, both the simulation and experiment showed a good interference mitigation performance.

eess.SP↗

arXiv2019

Experimental identification of electric dipoles induced by magnetic monopoles in Tb2Ti2O7

The fundamental principles of electrodynamics allow an electron carrying both electric monopole (charge) and magnetic dipole (spin) but prohibit its magnetic counterpart. Recently it was predicted that the magnetic "monopoles" carrying emergent magnetic charges in spin ice systems can induce electric dipoles. The inspiring prediction offers a novel way to study magnetic monopole excitations and magnetoelectric coupling. However, no clear example has been identified up to now. Here, we report the experimental evidence for electric dipoles induced by magnetic monopoles in spin frustrated Tb2Ti2O7. The magnetic field applied to pyrochlore Tb2Ti2O7 along [111] direction, brings out a "3-in-1-out" magnetic monopole configuration, and then induces a subtle structural phase transition at Hc~2.3 T. The transition is evidenced by the non-linear phonon splitting under magnetic fields and the anomalous crystal-field excitations of Tb3+ ions. The observations consistently point to the displacement of the oxygen O" anions along [111] axis which gives rise to the formation of electric dipoles. The finding demonstrates that the scenario of magnetic monopole having both magnetic charge and electric dipole is realized in Tb2Ti2O7 and sheds light into the coupling between electricity and magnetism of magnetic monopoles in spin frustrated systems.

cond-mat.str-el↗

arXiv2019

Dynamics of a Persistent Insulator-to-Metal Transition in Strained Manganite Films

Transition metal oxides possess complex free energy surfaces with competing degrees of freedom. Photoexcitation allows shaping of such rich energy landscapes. In epitaxially strained $\mathrm{La_{0.67}Ca_{0.33}MnO_3}$, optical excitation with a sub-100 fs pulse above $2\ \mathrm{mJ/cm^2}$ leads to a persistent metallic phase below 100 K. Using single-shot optical and terahertz spectroscopy, we show that this phase transition is a multi-step process. We conclude that the phase transition is driven by partial charge order melting, followed by growth of the persistent metallic phase on longer timescales. A time-dependent Ginzburg-Landau model can describe the fast dynamics of the reflectivity, followed by longer timescale in-growth of the metallic phase.

cond-mat.str-el↗

arXiv2018

Deep Diabetologist: Learning to Prescribe Hyperglycemia Medications with Hierarchical Recurrent Neural Networks

In healthcare, applying deep learning models to electronic health records (EHRs) has drawn considerable attention. EHR data consist of a sequence of medical visits, i.e. a multivariate time series of diagnosis, medications, physical examinations, lab tests, etc. This sequential nature makes EHR well matching the power of Recurrent Neural Network (RNN). In this paper, we propose "Deep Diabetologist" - using RNNs for EHR sequential data modelling, to provide the personalized hyperglycemia medication prediction for diabetic patients. Particularly, we develop a hierarchical RNN to capture the heterogeneous sequential information in the EHR data. Our experimental results demonstrate the improved performance, compared with a baseline classifier using logistic regression. Moreover, hierarchical RNN models outperform basic ones, providing deeper data insights for clinical decision support.

cs.LG↗

arXiv2017

Neural Network Multitask Learning for Traffic Flow Forecasting

Traditional neural network approaches for traffic flow forecasting are usually single task learning (STL) models, which do not take advantage of the information provided by related tasks. In contrast to STL, multitask learning (MTL) has the potential to improve generalization by transferring information in training signals of extra tasks. In this paper, MTL based neural networks are used for traffic flow forecasting. For neural network MTL, a backpropagation (BP) network is constructed by incorporating traffic flows at several contiguous time instants into an output layer. Nodes in the output layer can be seen as outputs of different but closely related STL tasks. Comprehensive experiments on urban vehicular traffic flow data and comparisons with STL show that MTL in BP neural networks is a promising and effective approach for traffic flow forecasting.

cs.LG↗

arXiv2017

Raman phonons in the ferroelectric-like metal LiOsO$_3$

The novel ferroelectric-like structural transition observed in metallic LiOsO$_3$ [Y. Shi et al., Nat. Mater. 12, 1024 (2013)], has invoked many theoretical and experimental interests. In this work, we have performed polarized and temperature-dependent Raman scattering measurements on high-quality single crystal LiOsO$_3$ and identified Raman-active modes in both centrosymmetric phase (300 K, R$\bar{3}$c) and non-centrosymmetric phase (10 K, R3c). Only four phonon peaks are observed in the former phase, while there are twelve peaks in the latter phase because of the reduction of crystal symmetry. With the help of careful symmetry analysis and first-principles calculations, we can make a systematic assignment for the observed Raman modes in both phases. The significant changes in line-width and the continuous evolution of Raman frequencies with temperatures were observed for the E$_g$ modes around the transition temperature, which suggests that the ferroelectric-like structural transition is a continuous order-disorder transition. The result sheds light on the coexistence of ferroelectricity and metallicity in the compound.

cond-mat.mtrl-sci↗

arXiv2017

Raman scattering study of large magnetoresistance semimetals TaAs$_2$ and NbAs$_2$

We have performed polarized and temperature-dependent Raman scattering measurements on extremely large magnetoresitance compounds TaAs$_2$ and NbAs$_2$. In both crystals, all the Raman active modes, including six A$_g$ modes and three B$_g$ modes, are clearly observed and well assigned with the combination of symmetry analysis and first-principles calculations. The well-resolved periodic intensity modulations of the observed modes with rotating crystal orientations, verify the symmetry of each assigned mode and are fitted to experimentally determine the elements of Raman tensor matrixes. The broadening of two A$_g$ modes seen in both compounds allows us to estimate electron-phonon coupling constant, which suggests a relatively small electron-phonon coupling in the semimetals TaAs$_2$ and NbAs$_2$. The present study provides the fundamental lattice dynamics information on TaAs$_2$ and NbAs$_2$ and may shed light on the understanding of their extraordinary large magnetoresistance.

cond-mat.mtrl-sci↗

arXiv2016

Raman scattering in transition metal dichalcogenides MTe2 (M = Mo, W)

We performed comparable polarized Raman scattering studies of MoTe2 and WTe2. By rotating crystals to tune the angle between the principal axis of the crystals and the polarization of the incident/scattered light, we obtained the angle dependence of the intensities for all the observed modes, which is perfectly consistent with careful symmetry analysis. Combining these results with first-principles calculations, we clearly identified the observed phonon modes in the different phases of both crystals. Fifteen Raman-active phonon modes (10Ag+5Bg) in the high-symmetry phase 1T'-MoTe2 (300 K) were well assigned, and all the symmetry-allowed Raman modes (11A1+6A2) in the low-symmetry phase Td-MoTe2 (10 K) and 12 Raman phonons (8A1+4A2) in Td-WTe2 were observed and identified. The present work provides basic information about the lattice dynamics in transition-metal dichalcogenides and may shed some light on the understanding of the extremely large magnetoresistance (MR) in this class of materials.

cond-mat.mtrl-sci↗

arXiv2015

Cooperative photoinduced metastable phase control in strained manganite films

A major challenge in condensed matter physics is active control of quantum phases. Dynamic control with pulsed electromagnetic fields can overcome energetic barriers enabling access to transient or metastable states that are not thermally accessible. Here we demonstrate strain-engineered tuning of La2/3Ca1/3MnO3 into an emergent charge-ordered insulating phase with extreme photo-susceptibility where even a single optical pulse can initiate a transition to a long-lived metastable hidden metallic phase. Comprehensive single-shot pulsed excitation measurements demonstrate that the transition is cooperative and ultrafast, requiring a critical absorbed photon density to activate local charge excitations that mediate magnetic-lattice coupling that, in turn, stabilize the metallic phase. These results reveal that strain engineering can tune emergent functionality towards proximal macroscopic states to enable dynamic ultrafast optical phase switching and control.

cond-mat.str-el↗

arXiv2015

Gapless quantum spin liquid ground state in the two-dimensional spin-1/2 triangular antiferromagnet YbMgGaO$_4$

Quantum spin liquid (QSL) is a novel state of matter which refuses the conventional spin freezing even at 0 K. Experimentally searching for the structurally perfect candidates is a big challenge in condensed matter physics. Here we report the successful synthesis of a new spin-1/2 triangular antiferromagnet YbMgGaO$_4$ with R$\bar{3}$m symmetry. The compound with an ideal two-dimensional and spatial isotropic magnetic triangular-lattice has no site-mixing magnetic defects and no antisymmetric Dzyaloshinsky-Moriya (DM) interactions. No spin freezing down to 60 mK (despite $Θ$$_w$ $\sim$ -4 K), the low-T power-law temperature dependence of heat capacity and nonzero susceptibility suggest that YbMgGaO$_4$ is a promising gapless ($\leq$ $|$$Θ$$_w$$|$/100) QSL candidate. The residual spin entropy, which is accurately determined with a non-magnetic reference LuMgGaO$_4$, approaches zero ($<$ 0.6 \%). This indicates that the possible QSL ground state (GS) of the frustrated spin system has been experimentally achieved at the lowest measurement temperatures.

cond-mat.str-el↗

arXiv2015

Raman scattering in superconducting NdO1-xFxBiS2 crystals

The recently discovered layered BiS2-based superconductors have attracted a great deal of interest due to their structural similarity to cuprate and iron-pnictide superconductors. We have performed Raman scattering measurements on two superconducting crystals NdO0.5F0.5BiS2 (Tc = 4.5 K) and NdO0.7F0.3BiS2 (Tc = 4.8 K). The observed Raman phonon modes are assigned with the aid of first-principles calculations. The asymmetrical phonon mode around 118 cm-1 reveals a small electron-phonon (e-ph) coupling constant 0.16, which is insufficient to generate superconductivity at ~ 4.5 K. In the Raman spectra there exists a clear temperature-dependent hump around 100 cm-1, which can be well understood in term of inter-band vertical transitions around Fermi surface. The transitions get boosted when the particular rectangular-like Fermi surface meets band splitting caused by spin-orbit coupling. It enables a unique and quantitative insight into the band splitting.

cond-mat.supr-con↗

arXiv2014

Simultaneously plasmon lasing and spasing behavior in a silver grating-film geometry

By using a self-consistent Maxwell-Bloch method, we demonstrate the simultaneously lasing and spasing behavior in a simple metal grating-film nanostructure, which can be attributed to spatial hole burning and the gain competition of different modes at the band edge and in the plasmonic band gap. We show three modes: one spaser mode in gap with quality factor as high as 248.54, one plasmon lasing mode at band edge which emit vertically from the grating surface, and the other plasmon lasing mode at band edge which is suppressed by the spaser mode. This method may find significant applications in coherent light and surface plasmon sources with low threshold, surface enhanced Raman scattering, solid-state lighting emission, etc.

physics.optics↗