arXiv subjects

Yang He

Publications and source records attributed to Yang He.

At least 127 records · Page 7Linked to original sources

arXiv2020

Synthetic Convolutional Features for Improved Semantic Segmentation

Recently, learning-based image synthesis has enabled to generate high-resolution images, either applying popular adversarial training or a powerful perceptual loss. However, it remains challenging to successfully leverage synthetic data for improving semantic segmentation with additional synthetic images. Therefore, we suggest to generate intermediate convolutional features and propose the first synthesis approach that is catered to such intermediate convolutional features. This allows us to generate new features from label masks and include them successfully into the training procedure in order to improve the performance of semantic segmentation. Experimental results and analysis on two challenging datasets Cityscapes and ADE20K show that our generated feature improves performance on segmentation tasks.

cs.CV↗

arXiv2020

Lithium niobate photonic-crystal electro-optic modulator

Modern advanced photonic integrated circuits require dense integration of high-speed electro-optic functional elements on a compact chip that consumes only moderate power. Energy efficiency, operation speed, and device dimension are thus crucial metrics underlying almost all current developments of photonic signal processing units. Recently, thin-film lithium niobate (LN) emerges as a promising platform for photonic integrated circuits. Here we make an important step towards miniaturizing functional components on this platform, reporting probably the smallest high-speed LN electro-optic modulators, based upon photonic crystal nanobeam resonators. The devices exhibit a significant tuning efficiency up to 1.98 GHz/V, a broad modulation bandwidth of 17.5 GHz, while with a tiny electro-optic modal volume of only 0.58 $μ{\rm m}^3$. The modulators enable efficient electro-optic driving of high-Q photonic cavity modes in both adiabatic and non-adiabatic regimes, and allow us to achieve electro-optic switching at 11 Gb/s with a bit-switching energy as low as 22 fJ. The demonstration of energy efficient and high-speed electro-optic modulation at the wavelength scale paves a crucial foundation for realizing large-scale LN photonic integrated circuits that are of immense importance for broad applications in data communication, microwave photonics, and quantum photonics.

physics.app-ph↗

arXiv2020

Consistency between ARPES and STM measurements on SmB$_6$

Strongly correlated topological surface states are promising platforms for next-generation quantum applications, but they remain elusive in real materials. The correlated Kondo insulator SmB$_6$ is one of the most promising candidates, with theoretically predicted heavy Dirac surface states supported by transport and scanning tunneling microscopy (STM) experiments. However, a puzzling discrepancy appears between STM and angle-resolved photoemission (ARPES) experiments on SmB$_6$. Although ARPES detects spin-textured surface states, their velocity is an order of magnitude higher than expected, while the Dirac point -- the hallmark of any topological system -- can only be inferred deep within the bulk valence band. A significant challenge is that SmB$_6$ lacks a natural cleavage plane, resulting in ordered surface domains limited to 10s of nanometers. Here we use STM to show that surface band bending can shift energy features by 10s of meV between domains. Starting from our STM spectra, we simulate the full spectral function as an average over multiple domains with different surface potentials. Our simulation shows excellent agreement with ARPES data, and thus resolves the apparent discrepancy between large-area measurements that average over multiple band-shifted domains and atomically-resolved measurements within a single domain.

cond-mat.str-el↗

arXiv2020

Stochastic Model Pruning via Weight Dropping Away and Back

Deep neural networks have dramatically achieved great success on a variety of challenging tasks. However, most successful DNNs have an extremely complex structure, leading to extensive research on model compression.As a significant area of progress in model compression, traditional gradual pruning approaches involve an iterative prune-retrain procedure and may suffer from two critical issues: local importance judgment, where the pruned weights are merely unimportant in the current model; and an irretrievable pruning process, where the pruned weights have no chance to come back. Addressing these two issues, this paper proposes the Drop Pruning approach, which leverages stochastic optimization in the pruning process by introducing a drop strategy at each pruning step, namely, drop away, which stochastically deletes some unimportant weights, and drop back, which stochastically recovers some pruned weights. The suitable choice of drop probabilities decreases the model size during the pruning process and helps it flow to the target sparsity. Compared to the Bayesian approaches that stochastically train a compact model for pruning, we directly aim at stochastic gradual pruning. We provide a detailed analysis showing that the drop away and drop back approaches have individual contributions. Moreover, Drop Pruning can achieve competitive compression performance and accuracy on many benchmark tasks compared with state-of-the-art weights pruning and Bayesian training approaches.

cs.LG↗

arXiv2020

Progressive Local Filter Pruning for Image Retrieval Acceleration

This paper focuses on network pruning for image retrieval acceleration. Prevailing image retrieval works target at the discriminative feature learning, while little attention is paid to how to accelerate the model inference, which should be taken into consideration in real-world practice. The challenge of pruning image retrieval models is that the middle-level feature should be preserved as much as possible. Such different requirements of the retrieval and classification model make the traditional pruning methods not that suitable for our task. To solve the problem, we propose a new Progressive Local Filter Pruning (PLFP) method for image retrieval acceleration. Specifically, layer by layer, we analyze the local geometric properties of each filter and select the one that can be replaced by the neighbors. Then we progressively prune the filter by gradually changing the filter weights. In this way, the representation ability of the model is preserved. To verify this, we evaluate our method on two widely-used image retrieval datasets,i.e., Oxford5k and Paris6K, and one person re-identification dataset,i.e., Market-1501. The proposed method arrives with superior performance to the conventional pruning methods, suggesting the effectiveness of the proposed method for image retrieval.

cs.CV↗

arXiv2020

Probing Image Potential States on Topological Semimetal Antimony Surface

A point charge near the surface of a topological insulator (TI) with broken time-reversal symmetry is predicted to generate an image magnetic charge in addition to an image electric charge. We use scanning tunneling spectroscopy to study the image potential states (IPS) of the topological semimetal Sb(111) surface. We observe five IPS with discrete energy levels that are well described by a one-dimensional model. The spatial variation of the IPS energies and lifetimes near surface step edges shows the first local signature of resonant interband scattering between IPS, which suggests that image charges too may interact. Our work motivates the exploration of the TI surface geometry necessary to realize and manipulate a magnetic charge.

cond-mat.mes-hall↗

arXiv2019

Asymptotic Soft Filter Pruning for Deep Convolutional Neural Networks

Deeper and wider Convolutional Neural Networks (CNNs) achieve superior performance but bring expensive computation cost. Accelerating such over-parameterized neural network has received increased attention. A typical pruning algorithm is a three-stage pipeline, i.e., training, pruning, and retraining. Prevailing approaches fix the pruned filters to zero during retraining, and thus significantly reduce the optimization space. Besides, they directly prune a large number of filters at first, which would cause unrecoverable information loss. To solve these problems, we propose an Asymptotic Soft Filter Pruning (ASFP) method to accelerate the inference procedure of the deep neural networks. First, we update the pruned filters during the retraining stage. As a result, the optimization space of the pruned model would not be reduced but be the same as that of the original model. In this way, the model has enough capacity to learn from the training data. Second, we prune the network asymptotically. We prune few filters at first and asymptotically prune more filters during the training procedure. With asymptotic pruning, the information of the training set would be gradually concentrated in the remaining filters, so the subsequent training and pruning process would be stable. Experiments show the effectiveness of our ASFP on image classification benchmarks. Notably, on ILSVRC-2012, our ASFP reduces more than 40% FLOPs on ResNet-50 with only 0.14% top-5 accuracy degradation, which is higher than the soft filter pruning (SFP) by 8%.

cs.CV↗

arXiv2019

Regularized Adversarial Sampling and Deep Time-aware Attention for Click-Through Rate Prediction

Improving the performance of click-through rate (CTR) prediction remains one of the core tasks in online advertising systems. With the rise of deep learning, CTR prediction models with deep networks remarkably enhance model capacities. In deep CTR models, exploiting users' historical data is essential for learning users' behaviors and interests. As existing CTR prediction works neglect the importance of the temporal signals when embed users' historical clicking records, we propose a time-aware attention model which explicitly uses absolute temporal signals for expressing the users' periodic behaviors and relative temporal signals for expressing the temporal relation between items. Besides, we propose a regularized adversarial sampling strategy for negative sampling which eases the classification imbalance of CTR data and can make use of the strong guidance provided by the observed negative CTR samples. The adversarial sampling strategy significantly improves the training efficiency, and can be co-trained with the time-aware attention model seamlessly. Experiments are conducted on real-world CTR datasets from both in-station and out-station advertising places.

cs.LG↗

arXiv2019

Perfect soliton crystals on demand

Recent advance of soliton microcombs has shown great promise to revolutionize many important areas such as optical communication, spectroscopic sensing, optical clock, and frequency synthesis. A largely tunable comb line spacing is crucial for the practical application of soliton microcombs, which unfortunately is challenging to realize for an on-chip monolithic microresonator. The recently discovered perfect soliton crystal (PSC) offers a convenient route to tune the comb line spacing. However, excitation of a PSC is generally stochastic by its nature and accessing a certain PSC state requires delicate tuning procedure. Here we demonstrate the on-demand generation of PSCs in a lithium niobate microresonator. The unique device characteristics allow us to produce a variety of PCSs and to switch between different PSC states, deterministically and repetitively. We utilize the device to show arbitrary dialing of the comb line spacing from 1 to 11 times of the free-spectral range of the resonator. The demonstration of PCSs on demand may now open up a great avenue for flexibly controlling the repetition rate of soliton pulses, which would significantly enhance and extend the application potential of soliton microcombs for communication, signal processing, and sensing.

physics.optics↗

arXiv2019

Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Acceleration

Previous works utilized ''smaller-norm-less-important'' criterion to prune filters with smaller norm values in a convolutional neural network. In this paper, we analyze this norm-based criterion and point out that its effectiveness depends on two requirements that are not always met: (1) the norm deviation of the filters should be large; (2) the minimum norm of the filters should be small. To solve this problem, we propose a novel filter pruning method, namely Filter Pruning via Geometric Median (FPGM), to compress the model regardless of those two requirements. Unlike previous methods, FPGM compresses CNN models by pruning filters with redundancy, rather than those with ''relatively less'' importance. When applied to two image classification benchmarks, our method validates its usefulness and strengths. Notably, on CIFAR-10, FPGM reduces more than 52% FLOPs on ResNet-110 with even 2.69% relative accuracy improvement. Moreover, on ILSVRC-2012, FPGM reduces more than 42% FLOPs on ResNet-101 without top-5 accuracy drop, which has advanced the state-of-the-art. Code is publicly available on GitHub: https://github.com/he-y/filter-pruning-geometric-median

cs.CV↗

arXiv2019

Density wave probes cuprate quantum phase transition

In cuprates, the strong correlations in proximity to the antiferromagnetic Mott insulating state give rise to an array of unconventional phenomena beyond high temperature superconductivity. Developing a complete description of the ground state evolution is crucial to decoding the complex phase diagram. Here we use the structure of broken translational symmetry, namely $d$-form factor charge modulations in (Bi,Pb)$_2$(Sr,La)$_2$CuO$_{6+δ}$, as a probe of the ground state reorganization that occurs at the transition from truncated Fermi arcs to a large Fermi surface. We use real space imaging of nanoscale electronic inhomogeneity as a tool to access a range of dopings within each sample, and we definitively validate the spectral gap $Δ$ as a proxy for local hole doping. From the $Δ$-dependence of the charge modulation wavevector, we discover a commensurate to incommensurate transition that is coincident with the Fermi surface transition from arcs to large hole pocket, demonstrating the qualitatively distinct nature of the electronic correlations governing the two sides of this quantum phase transition. Furthermore, the doping dependence of the incommensurate wavevector on the overdoped side is at odds with a simple Fermi surface driven instability.

cond-mat.supr-con↗

arXiv2019

Photon-level tuning of photonic nanocavities

Energy-efficient optical control of photonic device properties is crucial for diverse photonic signal processing. Here we demonstrate extremely efficient optical tuning of photonic nanocavities, with only photon-level optical energy. With a lithium niobate photonic crystal nanocavity with an optical Q up to 1.41 million and an effective mode volume down to 0.78$(λ/n)^3$, we are able to achieve a resonance tuning rate of about 88.4~MHz/photon (0.67~GHz/aJ), which allows us to tune across the whole cavity resonance with only about 2.4 photons on average inside the cavity. Such a photon-level resonance tuning is of great potential for energy-efficient optical switching, wavelength routing, and reconfiguration of photonic devices/circuits that are indispensable for future photonic interconnect.

physics.optics↗

arXiv2018

Unexpected Zero Bias Conductance Peak on the Topological Semimetal Sb(111) with a Broken Bilayer

The long-sought Majorana fermion is expected to manifest in a topological-superconductor heterostructure as a zero bias conductance peak (ZBCP). As one promising platform for such heterostructures, we investigate the cleaved surface of the topological semimetal Sb(111) using scanning tunneling microscopy and spectroscopy. Remarkably, we find a robust ZBCP on some terraces of the cleaved surface, although no superconductor is present. Using quasiparticle interference imaging, Landau level spectroscopy and density functional theory, we show that the ZBCP originates from a van Hove singularity pushed up to the Fermi level by a sub-surface stacking fault. Amidst the sprint to stake claims on new Majorana fermion systems, our finding highlights the importance of using a local probe together with detailed modeling to check thoroughly for crystal imperfections that may give rise to a trivial ZBCP unrelated to Majorana physics.

cond-mat.mes-hall↗

arXiv2018

Imaging emergent heavy Dirac fermions of a topological Kondo insulator

Kondo insulators are primary candidates in the search for strongly correlated topological quantum phases, which may host topological order, fractionalization, and non-Abelian statistics. Within some Kondo insulators, the hybridization gap is predicted to protect a nontrivial topological invariant and to harbor emergent heavy Dirac fermion surface modes. We use high-energy-resolution spectroscopic imaging in real and momentum space on the Kondo insulator, SmB$_6$. On cooling through $T^*_Δ\approx$ 35 K we observe the opening of an insulating gap that expands to $Δ\approx$ 10 meV at 2 K. Within the gap, we image the formation of linearly dispersing surface states with effective masses reaching $m^* = (410\pm20)m_e$. We thus demonstrate existence of a strongly correlated topological Kondo insulator phase hosting the heaviest known Dirac fermions.

cond-mat.str-el↗

arXiv2018

Optical parametric generation in a lithium niobate microring with modal phase matching

The lithium niobate integrated photonic platform has recently shown great promise in nonlinear optics on a chip scale. Here, we report second-harmonic generation in a high-Q lithium niobate microring resonator through modal phase matching, with a conversion efficiency of 1,500% W$^{-1}$. Our device also allows us to observe difference-frequency generation in the telecom band. Our work demonstrates the great potential of the lithium niobate integrated platform for nonlinear wavelength conversion with high efficiencies.

physics.optics↗

arXiv2018

Semi-nonlinear nanophotonic waveguides for highly efficient second-harmonic generation

Quadratic optical parametric processes form the foundation for a variety of applications related to classical and quantum frequency conversion, which have attracted significant interest recently in on-chip implementation. These processes rely on phase matching among the interacting guided modes, and refractive index engineering has been a primary approach for this purpose. Unfortunately, the modal phase matching approaches developed so far only produce parametric generation with fairly low efficiencies, due to the intrinsic modal mismatch of spatial symmetries. Here we propose a universal design and operation principle for highly efficient optical parametric generation on integrated photonic platforms. By introducing spatial symmetry breaking into the optical nonlinearity of the device, we are able to dramatically enhance the nonlinear parametric interaction to realize an extremely high efficiency. We employ this approach to design and fabricate a heterogeneous titanium oxide/lithium niobate nanophotonic waveguide that is able to offer second-harmonic generation with a theoretical conversion efficiency as high as 2900% W$^{-1}$cm$^{-2}$, which enables us to experimentally achieve a conversion efficiency of 36.0% W$^{-1}$ in a waveguide only 2.35 mm long, corresponding to a recorded normalized efficiency of 650% W$^{-1}$cm$^{-2}$ that is significantly beyond the reach of conventional modal phase matching approaches. Unlike nonlinearity domain engineering that is material selective, the proposed operation principle can be flexibly applied to any other on-chip quadratic nonlinear platform to support ultra-highly efficient optical parametric interactions, thus opening up a great avenue towards extreme nonlinear and quantum optics with great potentials for broad applications in energy efficient nonlinear and quantum photonic signal processing.

physics.optics↗

arXiv2018

Soft Filter Pruning for Accelerating Deep Convolutional Neural Networks

This paper proposed a Soft Filter Pruning (SFP) method to accelerate the inference procedure of deep Convolutional Neural Networks (CNNs). Specifically, the proposed SFP enables the pruned filters to be updated when training the model after pruning. SFP has two advantages over previous works: (1) Larger model capacity. Updating previously pruned filters provides our approach with larger optimization space than fixing the filters to zero. Therefore, the network trained by our method has a larger model capacity to learn from the training data. (2) Less dependence on the pre-trained model. Large capacity enables SFP to train from scratch and prune the model simultaneously. In contrast, previous filter pruning methods should be conducted on the basis of the pre-trained model to guarantee their performance. Empirically, SFP from scratch outperforms the previous filter pruning methods. Moreover, our approach has been demonstrated effective for many advanced CNN architectures. Notably, on ILSCRC-2012, SFP reduces more than 42% FLOPs on ResNet-101 with even 0.2% top-5 accuracy improvement, which has advanced the state-of-the-art. Code is publicly available on GitHub: https://github.com/he-y/soft-filter-pruning

cs.CV↗

arXiv2018

Diverse Conditional Image Generation by Stochastic Regression with Latent Drop-Out Codes

Recent advances in Deep Learning and probabilistic modeling have led to strong improvements in generative models for images. On the one hand, Generative Adversarial Networks (GANs) have contributed a highly effective adversarial learning procedure, but still suffer from stability issues. On the other hand, Conditional Variational Auto-Encoders (CVAE) models provide a sound way of conditional modeling but suffer from mode-mixing issues. Therefore, recent work has turned back to simple and stable regression models that are effective at generation but give up on the sampling mechanism and the latent code representation. We propose a novel and efficient stochastic regression approach with latent drop-out codes that combines the merits of both lines of research. In addition, a new training objective increases coverage of the training distribution leading to improvements over the state of the art in terms of accuracy as well as diversity.

cs.CV↗