arXiv subjects

Yu Lu

Publications and source records attributed to Yu Lu.

At least 127 records · Page 7Linked to original sources

arXiv2021

Reconfigurable Intelligent Surface Based Hybrid Precoding for THz Communications

Benefiting from the growth of the bandwidth, Terahertz (THz) communication can support the new application with explosive requirements of the ultra-high-speed rates for future 6G wireless systems. In order to compensate for the path loss of high frequency, massive multiple-input multiple-output (MIMO) can be utilized for high array gains by beamforming. However, since a large number of analog phase shifters should be used to realize the analog beamforming, the existing THz communication with massive MIMO has very high energy consumption. To solve this problem, a reconfigurable intelligent surface (RIS)-based hybrid precoding architecture for THz communication is developed in this paper, where the energy-hungry phased array is replaced by the energy-efficient RIS to realize the analog beamforming of the hybrid precoding. Then, based on the proposed RIS-based architecture, a sum-rate maximization problem for hybrid precoding is investigated. Since the phase shifts implemented by RIS in practice are often discrete, this sum-rate maximization problem with a non-convex constraint is challenging. Next, the sum-rate maximization problem is reformulated as a parallel deep neural network (DNN)-based classification problem, which can be solved by the proposed low-complexity deep learning-based multiple discrete classification (DL-MDC) hybrid precoding scheme. Finally, we provide numerous simulation results to show that the proposed DL-MDC scheme works well both in the theoretical Saleh-Valenzuela channel model and practical 3GPP channel model. Compared with existing iterative search algorithms, the can proposed DL-MDC scheme reduces the runtime significantly with a negligible performance loss.

cs.IT↗

arXiv2021

Federated Two-stage Learning with Sign-based Voting

Federated learning is a distributed machine learning mechanism where local devices collaboratively train a shared global model under the orchestration of a central server, while keeping all private data decentralized. In the system, model parameters and its updates are transmitted instead of raw data, and thus the communication bottleneck has become a key challenge. Besides, recent larger and deeper machine learning models also pose more difficulties in deploying them in a federated environment. In this paper, we design a federated two-stage learning framework that augments prototypical federated learning with a cut layer on devices and uses sign-based stochastic gradient descent with the majority vote method on model updates. Cut layer on devices learns informative and low-dimension representations of raw data locally, which helps reduce global model parameters and prevents data leakage. Sign-based SGD with the majority vote method for model updates also helps alleviate communication limitations. Empirically, we show that our system is an efficient and privacy preserving federated learning scheme and suits for general application scenarios.

cs.DC↗

arXiv2021

Development of a GPU-accelerated Monte Carlo dose calculation module for nuclear medicine, ARCHER-NM: Demonstration for a PET/CT imaging procedure

This paper describes the development and validation of a Monte Carlo (MC) dose computing module dedicated to organ dose calculations of patients undergoing nuclear medicine (NM) internal radiation exposures involving 18F-FDG PET/CT examination. This new module extends the more-than-10-years-long ARCHER project that developed a GPU-accelerated MC dose engine by adding dedicated NM source-definition features. To validate the code, we compared dose distributions from the 0.511-MeV point photon source calculated for a water phantom as well as a patient PET/CT phantom against a well-tested MC code, GATE. The water-phantom results show excellent agreement, suggesting that the radiation physics module in the new NM code is adequate. To demonstrate the clinical utility and advantage of ARCHER-NM, one set of PET/CT data for an adult male NM patient is calculated using the new code. Radiosensitive organs in the CT dataset are segmented using a CNN-based tool called DeepViewer. The PET image intensity maps are converted to radioactivity distributions to allow for MC radiation transport dose calculations at the voxel level. The dose rate maps and corresponding statistical uncertainties were calculated for the duration of PET image acquisition. The dose rate results of the 18F-FDG PET imaging patient show that ARCHER-NM's results agree very well with those of the GATE within 0.58% to 4.11%. Most impressively, ARCHER-NM obtains such results in less than 0.5 minutes while it takes GATE as much as 376 minutes. This is the first study presenting GPU-accelerated patient-specific MC internal radiation dose rate calculations for clinically realistic 18F-FDG PET/CT imaging cases involving auto-segmentation of whole-body PET/CT images. This study suggests that modern computing tools -- ARCHER-NM and DeepViewer -- are accurate and fast enough for routine internal dosimetry in NM clinics.

physics.med-ph↗

arXiv2021

Exploring a quantum-information-relevant magnonic material: Ultralow damping at low temperature in the organic ferrimagnet V[TCNE]x

Quantum information science and engineering requires novel low-loss magnetic materials for magnon-based quantum-coherent operations. The search for low-loss magnetic materials, traditionally driven by applications in microwave electronics near room-temperature, has gained additional constraints from the need to operate at cryogenic temperatures for many applications in quantum information science and technology. Whereas yttrium iron garnet (YIG) has been the material of choice for decades, the emergence of molecule-based materials with robust magnetism and ultra-low damping has opened new avenues for exploration. Specifically, thin-films of vanadium tetracyanoethylene (V[TCNE]x) can be patterned into the multiple, connected structures needed for hybrid quantum elements and have shown room-temperature Gilbert damping (α = 4 \times 10^-5) that rivals the intrinsic (bulk) damping otherwise seen only in highly-polished YIG spheres (far more challenging to integrate into arrays). Here, we present a comprehensive and systematic study of the low-temperature magnetization dynamics for V[TCNE]x thin films, with implications for their application in quantum systems. These studies reveal a temperature-driven, strain-dependent magnetic anisotropy that compensates the thin-film shape anisotropy, and the recovery of a magnetic resonance linewidth at 5 K that is comparable to room-temperature values (roughly 2 G at 9.4 GHz). We can account for these variations of the V[TCNE]x linewidth within the context of scattering from very dilute paramagnetic impurities, and anticipate additional linewidth narrowing as the temperature is further reduced.

physics.app-ph↗

arXiv2021

Heavy quark spin partners of the $Y(4260)$ in coupled-channel formalism

The charmoniumlike state $Y(4260)$ is described as predominantly a $D_1 \bar{D}$ molecule in a coupled-channel quark model [Phys.\,Rev.\,D\,\textbf{96},\,114022\,(2017)]. The heavy quark spin symmetry (HQSS) thus implies the possible emergence of its heavy quark spin partners with molecular configuration as $D_1 \bar{D}^*$ and $D_2^* \bar{D}^*$ below these charmed mesons' thresholds. We analyze the probabilities of various intermediate charmed meson loops for $J^{PC}=1^{--}$ exotic state $Y(4360)$ and find that the channel $D_1 \bar{D}^*$ couples more strongly around its mass regime, and the coupling behavior remains the same even if the mass of $Y(4360)$ is pushed closer to $D_1 \bar{D}^*$ threshold. This enlightens that the most favorable molecular scenario for the $Y(4360)$ could be $D_1 \bar{D}^*$, and hence it can be interpreted as HQSS partner of the $Y(4260)$. We also find the strong coupling behavior of $D_2^* \bar{D}^*$ channel with the $ψ(4415)$, which makes it a good candidate for a dominant $D_2^* \bar{D}^*$ molecule. We discuss the important decay patterns of these resonances to disentangle their long- and short-distance structures.

hep-ph↗

arXiv2021

Interpreting Deep Knowledge Tracing Model on EdNet Dataset

With more deep learning techniques being introduced into the knowledge tracing domain, the interpretability issue of the knowledge tracing models has aroused researchers' attention. Our previous study(Lu et al. 2020) on building and interpreting the KT model mainly adopts the ASSISTment dataset(Feng, Heffernan, and Koedinger 2009),, whose size is relatively small. In this work, we perform the similar tasks but on a large and newly available dataset, called EdNet(Choi et al. 2020). The preliminary experiment results show the effectiveness of the interpreting techniques, while more questions and tasks are worthy to be further explored and accomplished.

cs.AI↗

arXiv2021

Towards Heterogeneous Clients with Elastic Federated Learning

Federated learning involves training machine learning models over devices or data silos, such as edge processors or data warehouses, while keeping the data local. Training in heterogeneous and potentially massive networks introduces bias into the system, which is originated from the non-IID data and the low participation rate in reality. In this paper, we propose Elastic Federated Learning (EFL), an unbiased algorithm to tackle the heterogeneity in the system, which makes the most informative parameters less volatile during training, and utilizes the incomplete local updates. It is an efficient and effective algorithm that compresses both upstream and downstream communications. Theoretically, the algorithm has convergence guarantee when training on the non-IID data at the low participation rate. Empirical experiments corroborate the competitive performance of EFL framework on the robustness and the efficiency.

cs.LG↗

arXiv2021

Efficient Speech Emotion Recognition Using Multi-Scale CNN and Attention

Emotion recognition from speech is a challenging task. Re-cent advances in deep learning have led bi-directional recur-rent neural network (Bi-RNN) and attention mechanism as astandard method for speech emotion recognition, extractingand attending multi-modal features - audio and text, and thenfusing them for downstream emotion classification tasks. Inthis paper, we propose a simple yet efficient neural networkarchitecture to exploit both acoustic and lexical informationfrom speech. The proposed framework using multi-scale con-volutional layers (MSCNN) to obtain both audio and text hid-den representations. Then, a statistical pooling unit (SPU)is used to further extract the features in each modality. Be-sides, an attention module can be built on top of the MSCNN-SPU (audio) and MSCNN (text) to further improve the perfor-mance. Extensive experiments show that the proposed modeloutperforms previous state-of-the-art methods on IEMOCAPdataset with four emotion categories (i.e., angry, happy, sadand neutral) in both weighted accuracy (WA) and unweightedaccuracy (UA), with an improvement of 5.0% and 5.2% respectively under the ASR setting.

cs.SD↗

arXiv2021

RevCore: Review-augmented Conversational Recommendation

Existing conversational recommendation (CR) systems usually suffer from insufficient item information when conducted on short dialogue history and unfamiliar items. Incorporating external information (e.g., reviews) is a potential solution to alleviate this problem. Given that reviews often provide a rich and detailed user experience on different interests, they are potential ideal resources for providing high-quality recommendations within an informative conversation. In this paper, we design a novel end-to-end framework, namely, Review-augmented Conversational Recommender (RevCore), where reviews are seamlessly incorporated to enrich item information and assist in generating both coherent and informative responses. In detail, we extract sentiment-consistent reviews, perform review-enriched and entity-based recommendations for item suggestions, as well as use a review-attentive encoder-decoder for response generation. Experimental results demonstrate the superiority of our approach in yielding better performance on both recommendation and conversation responding.

cs.CL↗

arXiv2021

Comparison of Observed Galaxy Properties with Semianalytic Model Predictions using Machine Learning

With current and upcoming experiments such as WFIRST, Euclid and LSST, we can observe up to billions of galaxies. While such surveys cannot obtain spectra for all observed galaxies, they produce galaxy magnitudes in color filters. This data set behaves like a high-dimensional nonlinear surface, an excellent target for machine learning. In this work, we use a lightcone of semianalytic galaxies tuned to match CANDELS observations from Lu et al. (2014) to train a set of neural networks on a set of galaxy physical properties. We add realistic photometric noise and use trained neural networks to predict stellar masses and average star formation rates on real CANDELS galaxies, comparing our predictions to SED fitting results. On semianalytic galaxies, we are nearly competitive with template-fitting methods, with biases of $0.01$ dex for stellar mass, $0.09$ dex for star formation rate, and $0.04$ dex for metallicity. For the observed CANDELS data, our results are consistent with template fits on the same data at $0.15$ dex bias in $M_{\rm star}$ and $0.61$ dex bias in star formation rate. Some of the bias is driven by SED-fitting limitations, rather than limitations on the training set, and some is intrinsic to the neural network method. Further errors are likely caused by differences in noise properties between the semianalytic catalogs and data. Our results show that galaxy physical properties can in principle be measured with neural networks at a competitive degree of accuracy and precision to template-fitting methods.

astro-ph.GA↗

arXiv2020

GINet: Graph Interaction Network for Scene Parsing

Recently, context reasoning using image regions beyond local convolution has shown great potential for scene parsing. In this work, we explore how to incorporate the linguistic knowledge to promote context reasoning over image regions by proposing a Graph Interaction unit (GI unit) and a Semantic Context Loss (SC-loss). The GI unit is capable of enhancing feature representations of convolution networks over high-level semantics and learning the semantic coherency adaptively to each sample. Specifically, the dataset-based linguistic knowledge is first incorporated in the GI unit to promote context reasoning over the visual graph, then the evolved representations of the visual graph are mapped to each local representation to enhance the discriminated capability for scene parsing. GI unit is further improved by the SC-loss to enhance the semantic representations over the exemplar-based semantic graph. We perform full ablation studies to demonstrate the effectiveness of each component in our approach. Particularly, the proposed GINet outperforms the state-of-the-art approaches on the popular benchmarks, including Pascal-Context and COCO Stuff.

cs.CV↗

arXiv2020

Seamless maps of major elements of the Moon: Results from high-resolution geostationary satellite

Major elements such as Fe, Ti, Mg, Al, Ca, and Si play very important roles in understanding the origin and evolution of the Moon. Previous maps of these major elements derived from orbital data are based on mosaic images or low-resolution Gamma ray data. The hue variations and gaps among orbital boundaries in the mosaic images are not conducive to geological studies. This paper aims to produce seamless and homogenous distribution maps of major elements using the single-exposure image of the whole lunar disk obtained by China's high-resolution geostationary satellite, Gaofen-4, with a spatial resolution of ~500 m. The elemental contents of soil samples returned by Apollo and Luna missions were used as ground truth, and were correlated with the reflectance of the sampling sites extracted from Gaofen-4 data. The final distribution maps of these major oxides are generated with the statistical regression model. With these products the average contents and proportions of the major elements for maria and highlands were estimated and compared. The results showed that SiO2 and TiO2 have the highest and lowest fractions in mare and highland areas, respectively. Besides, the relative concentrations of these elements could serve as indicators of geologic processes, e.g., the obviously asymmetric distributions of Al2O3, CaO, and SiO2 around Tycho crater may suggest that Tycho crater was formed by an oblique impact from the southwest direction.

astro-ph.EP↗

arXiv2020

A random-walk model for dark matter halo spins

We extend the random-walk model of Vitvitska et al. for predicting the spins of dark matter halos from their merger histories. Using updated merger rates, orbital parameter distributions, and N-body constraints we show that this model can accurately reproduce the distribution of spin parameters measured in N-body simulations when we include a weak correlation between the spins of halos and the angular momenta of infalling subhalos. We further show that this model is in approximate agreement with the correlation of the spin magnitude over time as determined from N-body simulations, while it slightly underpredicts the correlation in the direction of the spin vector measured from the same simulations. This model is useful for predicting spins from merger histories derived from non-N-body sources, thereby circumventing the need for very high resolution simulations to permit accurate measurements of spins. It may be particularly relevant to modeling systems which accumulate angular momentum from halos over time (such as galactic discs) - we show that this model makes small but significant changes in the distribution of galactic disc sizes computed using the Galacticus semi-analytic galaxy formation model.

astro-ph.GA↗

arXiv2020

Quark structure of the $χ_{\rm c}(3P)$ and $X(4274)$ resonances and their strong and radiative decays

We calculate the masses of $χ_{\rm c}(3P)$ states with threshold corrections in a coupled-channel model. The model was recently applied to the description of the properties of $χ_{\rm c}(2P)$ and $χ_{\rm b}(3P)$ multiplets [Phys.\ Lett.\ B {\bf 789}, 550 (2019)]. We also compute the open-charm strong decay widths of the $χ_{\rm c}(3P)$ states and their radiative transitions. According to our predictions, the $χ_{\rm c}(3P)$ states should be dominated by the charmonium core, but they may also show small meson-meson components. The $X(4274)$ is interpreted as a $c \bar c$ $χ_{\rm c1}(3P)$ state. More informations on the other members of the $χ_{\rm c}(3P)$ multiplet, as well as a more rigorous analysis of the $X(4274)$'s decay modes, are needed to provide further indications on the quark structure of the previous resonance.

hep-ph↗

arXiv2020

Towards Interpretable Deep Learning Models for Knowledge Tracing

As an important technique for modeling the knowledge states of learners, the traditional knowledge tracing (KT) models have been widely used to support intelligent tutoring systems and MOOC platforms. Driven by the fast advancements of deep learning techniques, deep neural network has been recently adopted to design new KT models for achieving better prediction performance. However, the lack of interpretability of these models has painfully impeded their practical applications, as their outputs and working mechanisms suffer from the intransparent decision process and complex inner structures. We thus propose to adopt the post-hoc method to tackle the interpretability issue for deep learning based knowledge tracing (DLKT) models. Specifically, we focus on applying the layer-wise relevance propagation (LRP) method to interpret RNN-based DLKT model by backpropagating the relevance from the model's output layer to its input layer. The experiment results show the feasibility using the LRP method for interpreting the DLKT model's predictions, and partially validate the computed relevance scores from both question level and concept level. We believe it can be a solid step towards fully interpreting the DLKT models and promote their practical applications in the education domain.

cs.LG↗

arXiv2020

C-DLinkNet: considering multi-level semantic features for human parsing

Human parsing is an essential branch of semantic segmentation, which is a fine-grained semantic segmentation task to identify the constituent parts of human. The challenge of human parsing is to extract effective semantic features to resolve deformation and multi-scale variations. In this work, we proposed an end-to-end model called C-DLinkNet based on LinkNet, which contains a new module named Smooth Module to combine the multi-level features in Decoder part. C-DLinkNet is capable of producing competitive parsing performance compared with the state-of-the-art methods with smaller input sizes and no additional information, i.e., achiving mIoU=53.05 on the validation set of LIP dataset.

cs.CV↗

arXiv2020

Unveiling the secrets of the mid-infrared Moon

The Moon's optical characteristics in visible and long-wavelength infrared (LWIR) have long been observed with our eyes or with instruments. What the mid-infrared (MIR) Moon looks like is still a mystery. For the first time we present detailed appearance of the MIR Moon observed by a high-resolution geostationary satellite and reveal the essence behind its appearance. The appearance of the MIR Moon is opposite to its normal visible appearance. In addition the MIR Moon shows limb darkening. Both the absolute and the relative brightness distribution of the MIR lunar disk changes with the solar incidence angle. The signatures of the MIR Moon are controlled by both the reflection and emission of the lunar surface. We also show first-ever brightness temperature maps of the lunar disk without needing a mosaic, which better show the temperature variation across the lunar disk. They reveal that the relationship between brightness temperature and solar incidence angle i is cos1/bi, and the power parameter is smaller than the Lambertian temperature model of cos1/4i observed for lunar orbit-based measurements. The slower decrease of the brightness temperature when moving away from the sub-solar point than the Lambertian model is due to topographic effects. The brightness temperature is dominated by albedo and the solar incidence angle and influenced by the topography. Our results indicate that the Moon in the MIR exhibits many interesting phenomena which were previously unknown, and contains abundant information about lunar reflection and thermal emission for future study.

astro-ph.EP↗

arXiv2020

Object Instance Mining for Weakly Supervised Object Detection

Weakly supervised object detection (WSOD) using only image-level annotations has attracted growing attention over the past few years. Existing approaches using multiple instance learning easily fall into local optima, because such mechanism tends to learn from the most discriminative object in an image for each category. Therefore, these methods suffer from missing object instances which degrade the performance of WSOD. To address this problem, this paper introduces an end-to-end object instance mining (OIM) framework for weakly supervised object detection. OIM attempts to detect all possible object instances existing in each image by introducing information propagation on the spatial and appearance graphs, without any additional annotations. During the iterative learning process, the less discriminative object instances from the same class can be gradually detected and utilized for training. In addition, we design an object instance reweighted loss to learn larger portion of each object instance to further improve the performance. The experimental results on two publicly available databases, VOC 2007 and 2012, demonstrate the efficacy of proposed approach.

cs.CV↗