arXiv subjects

Xin Wen

Publications and source records attributed to Xin Wen.

83 records · Page 5Linked to original sources

arXiv2020

Laser Intensity Noise Suppression for Preparing Audio-Frequency 795 nm Squeezed Vacuum State of Light at Rubidium D1 Line

Laser intensity noise suppression has essential effects on preparation and characterization of the audio-frequency squeezed vacuum state of light based on a sub-threshold optical parametric oscillator (OPO).We have implemented two feedback loops by using relevant acousto-optical modulators (AOM) to stabilize the intensity of 795-nm near infrared (NIR) fundamental laser and 397.5-nm ultraviolet (UV) laser generated by cavity-enhanced frequency doubling.Typical peak-to-peak laser intensity fluctuation with a bandwidth of $\sim10$ kHz in a half hour has been improved from $\pm7.45$$\%$ to $\pm0.06$$\%$ for 795-nm NIR laser beam, and from $\pm9.04$$\%$ to $\pm0.05$$\%$ for 397.5-nm UV laser beam, respectively. The squeezing level of the squeezed vacuum state at 795 nm prepared by the sub-threshold OPO with a PPKTP crystal has been improved from -3.3 to -4.0 dB around 3$\sim$9 kHz of audio analysis frequency range.

physics.atom-ph↗

arXiv2019

Point2SpatialCapsule: Aggregating Features and Spatial Relationships of Local Regions on Point Clouds using Spatial-aware Capsules

Learning discriminative shape representation directly on point clouds is still challenging in 3D shape analysis and understanding. Recent studies usually involve three steps: first splitting a point cloud into some local regions, then extracting corresponding feature of each local region, and finally aggregating all individual local region features into a global feature as shape representation using simple max pooling. However, such pooling-based feature aggregation methods do not adequately take the spatial relationships between local regions into account, which greatly limits the ability to learn discriminative shape representation. To address this issue, we propose a novel deep learning network, named Point2SpatialCapsule, for aggregating features and spatial relationships of local regions on point clouds, which aims to learn more discriminative shape representation. Compared with traditional max-pooling based feature aggregation networks, Point2SpatialCapsule can explicitly learn not only geometric features of local regions but also spatial relationships among them. It consists of two modules. To resolve the disorder problem of local regions, the first module, named geometric feature aggregation, is designed to aggregate the local region features into the learnable cluster centers, which explicitly encodes the spatial locations from the original 3D space. The second module, named spatial relationship aggregation, is proposed for further aggregating clustered features and the spatial relationships among them in the feature space using the spatial-aware capsules developed in this paper. Compared to the previous capsule network based methods, the feature routing on the spatial-aware capsules can learn more discriminative spatial relationships among local regions for point clouds, which establishes a direct mapping between log priors and the spatial locations through feature clusters.

cs.CV↗

arXiv2019

L2G Auto-encoder: Understanding Point Clouds by Local-to-Global Reconstruction with Hierarchical Self-Attention

Auto-encoder is an important architecture to understand point clouds in an encoding and decoding procedure of self reconstruction. Current auto-encoder mainly focuses on the learning of global structure by global shape reconstruction, while ignoring the learning of local structures. To resolve this issue, we propose Local-to-Global auto-encoder (L2G-AE) to simultaneously learn the local and global structure of point clouds by local to global reconstruction. Specifically, L2G-AE employs an encoder to encode the geometry information of multiple scales in a local region at the same time. In addition, we introduce a novel hierarchical self-attention mechanism to highlight the important points, scales and regions at different levels in the information aggregation of the encoder. Simultaneously, L2G-AE employs a recurrent neural network (RNN) as decoder to reconstruct a sequence of scales in a local region, based on which the global point cloud is incrementally reconstructed. Our outperforming results in shape classification, retrieval and upsampling show that L2G-AE can understand point clouds better than state-of-the-art methods.

cs.CV↗

arXiv2019

Adversarial Cross-Modal Retrieval via Learning and Transferring Single-Modal Similarities

Cross-modal retrieval aims to retrieve relevant data across different modalities (e.g., texts vs. images). The common strategy is to apply element-wise constraints between manually labeled pair-wise items to guide the generators to learn the semantic relationships between the modalities, so that the similar items can be projected close to each other in the common representation subspace. However, such constraints often fail to preserve the semantic structure between unpaired but semantically similar items (e.g. the unpaired items with the same class label are more similar than items with different labels). To address the above problem, we propose a novel cross-modal similarity transferring (CMST) method to learn and preserve the semantic relationships between unpaired items in an unsupervised way. The key idea is to learn the quantitative similarities in single-modal representation subspace, and then transfer them to the common representation subspace to establish the semantic relationships between unpaired items across modalities. Experiments show that our method outperforms the state-of-the-art approaches both in the class-based and pair-based retrieval tasks.

cs.MM↗

arXiv2018

Flexoelectret: An Electret with Tunable Flexoelectric-like Response

Because of the flexoelectric effect, dielectric materials usually polarize in response to a strain gradient. Soft materials are good candidates for developing large strain gradient because of their good deformability. However, they always suffer from lower flexoelectric coefficients compared to ceramics. In this work, a flexoelectric-like effect is introduced to enhance the effective flexoelectricity of a polydimethylsiloxane (PDMS) bar. The flexoelectric-like effect is realized by depositing a layer of net charges on the middle plane of the bar to form an electret. Experiments show that the enhancement of the flexoelectricity depends on the density of inserted net charges. It is found that a charged layer with surface potential of -5723V results in 100 times increase of the material's flexoelectric coefficient. We also show that the enhancement is proportional to the thickness of electrets. This work provides a new way of enhancing flexoelectricity in soft materials and further prompt the application of soft materials in electromechanical transducers.

cond-mat.soft↗

arXiv2017

Generation of polarization squeezed light with an optical parametric amplifier at 795 nm

We report the experimental demonstration of polarization squeezed beam at 795 nm by combining a quadrature squeezed beam with a coherent beam. The quadrature squeezed beam is generated by a degenerate optical parametric amplifier based on a PPKTP crystal. Stokes parameter squeezing of -3.8 dB and anti squeezing of +5.0 dB is observed. This polarization squeezed beam resonant to rubidium D1 line has potential applications in quantum information networks and precise measurement beyond the shot noise limit.

quant-ph↗

arXiv2017

Polarization squeezing at the audio frequency band for the Rubidium D1 line

A 2.8-dB polarization squeezing of the Stokes operator S2 for the rubidium D1 line (795 nm) is achieved, with the lowest squeezing band at an audio frequency of 2.6 kHz. Two methods are applied for improving the squeezing towards low frequencies: an orthogonal-polarized locking beam that precludes residual unwanted interference and quantum noise locking method that stabilizes the relative phase between the squeezed vacuum and the local oscillator. The squeezing level is limited by absorption-induced losses at short wavelengths, here is 397.5 nm. The generated polarization squeezed light can be used in a quantum enhanced magnetometer to increase the measurement sensitivity.

quant-ph↗

arXiv2016

Improved Sufficient Conditions for Exact Convex Relaxation of Storage-Concerned ED

To avoid simultaneous charging and discharging of storages, complementarity constraints are introduced to storage-concerned economic dispatch (ED), which makes the problem non-convex. This letter concerns the conditions under which the convex relaxation of storage-concerned ED with complementarity constraints is exact. Two new sufficient conditions are proposed, proved and verified to significantly reduce the conservatism of recent results [3], [4].

math.OC↗

arXiv2016

Comparison and characterization of efficient frequency doubling at 397.5 nm with PPKTP, LBO and BiBO crystals

A continuous-wave Ti:sapphire laser at 795 nm is frequency doubled in a bow-tie type enhancement four-mirror ring cavity with LiB3O5 (LBO), BiB3O6 (BiBO), and periodically polled KTiOPO4 (PPKTP) crystals, respectively. The properties of 397.5 nm ultra-violet (UV) output power, beam quality, stability for these different nonlinear crystals are investigated and compared. For PPKTP crystal, the highest doubling efficiency of 58.1% is achieved from 191 mW of 795 nm mode-matched fundamental power to 111 mW of 397.5 nm UV output. For LBO crystal, with 1.34 W of mode-matched 795 nm power, 770 mW of 397.5 nm UV output is achieved, implying a doubling efficiency of 57.4%. For BiBO crystal, with 323 mW of mode-matched 795 nm power, 116 mW of 397.5 nm UV output is achieved, leading to a doubling efficiency of 35.9%. The generated UV radiation has potential applications in the fields of quantum physics

physics.optics↗

arXiv2016

Improvement of vacuum squeezing resonant on the rubidium D1 line at 795 nm

We report on efficient generation of second harmonic laser and single-mode vacuum squeezed light of 795 nm with periodically poled KTiOPO4 (PPKTP) crystals. We achieved 111 mW of ultra-violet (UV) light at 397.5 nm from 191 mW of fundamental light with a PPKTP crystal in a doubling cavity, corresponding to a conversion efficiency of 58.1%. Using the UV light to pump an optical parametric oscillator with a PPKTP crystal, we realized -5.6 dB of a maximum squeezing. We analyzed the pump power dependence of the squeezing level and concluded that the UV light induced losses limit the improvement of the squeezing level. The generated squeezed light has huge potential application in quantum memory and ultra-precise measurement.

quant-ph↗

arXiv2015

General Rank Multiuser Downlink Beamforming With Shaping Constraints Using Real-valued OSTBC

In this paper we consider optimal multiuser downlink beamforming in the presence of a massive number of arbitrary quadratic shaping constraints. We combine beamforming with full-rate high dimensional real-valued orthogonal space time block coding (OSTBC) to increase the number of beamforming weight vectors and associated degrees of freedom in the beamformer design. The original multi-constraint beamforming problem is converted into a convex optimization problem using semidefinite relaxation (SDR) which can be solved efficiently. In contrast to conventional (rank-one) beamforming approaches in which an optimal beamforming solution can be obtained only when the SDR solution (after rank reduction) exhibits the rank-one property, in our approach optimality is guaranteed when a rank of eight is not exceeded. We show that our approach can incorporate up to 79 additional shaping constraints for which an optimal beamforming solution is guaranteed as compared to a maximum of two additional constraints that bound the conventional rank-one downlink beamforming designs. Simulation results demonstrate the flexibility of our proposed beamformer design.

cs.IT↗