arXiv subjects

Yifan Zhu

Publications and source records attributed to Yifan Zhu.

At least 145 records · Page 8Linked to original sources

arXiv2023

Looking Through the Glass: Neural Surface Reconstruction Against High Specular Reflections

Neural implicit methods have achieved high-quality 3D object surfaces under slight specular highlights. However, high specular reflections (HSR) often appear in front of target objects when we capture them through glasses. The complex ambiguity in these scenes violates the multi-view consistency, then makes it challenging for recent methods to reconstruct target objects correctly. To remedy this issue, we present a novel surface reconstruction framework, NeuS-HSR, based on implicit neural rendering. In NeuS-HSR, the object surface is parameterized as an implicit signed distance function (SDF). To reduce the interference of HSR, we propose decomposing the rendered image into two appearances: the target object and the auxiliary plane. We design a novel auxiliary plane module by combining physical assumptions and neural networks to generate the auxiliary plane appearance. Extensive experiments on synthetic and real-world datasets demonstrate that NeuS-HSR outperforms state-of-the-art approaches for accurate and robust target surface reconstruction against HSR. Code is available at https://github.com/JiaxiongQ/NeuS-HSR.

cs.CV↗

arXiv2023

An Information-Theoretic Framework for Supervised Learning

Each year, deep learning demonstrates new and improved empirical results with deeper and wider neural networks. Meanwhile, with existing theoretical frameworks, it is difficult to analyze networks deeper than two layers without resorting to counting parameters or encountering sample complexity bounds that are exponential in depth. Perhaps it may be fruitful to try to analyze modern machine learning under a different lens. In this paper, we propose a novel information-theoretic framework with its own notions of regret and sample complexity for analyzing the data requirements of machine learning. With our framework, we first work through some classical examples such as scalar estimation and linear regression to build intuition and introduce general techniques. Then, we use the framework to study the sample complexity of learning from data generated by deep neural networks with ReLU activation units. For a particular prior distribution on weights, we establish sample complexity bounds that are simultaneously width independent and linear in depth. This prior distribution gives rise to high-dimensional latent representations that, with high probability, admit reasonably accurate low-dimensional approximations. We conclude by corroborating our theoretical results with experimental analysis of random single-hidden-layer neural networks.

cs.LG↗

arXiv2022

Deep Learning Assisted Optimization for 3D Reconstruction from Single 2D Line Drawings

In this paper, we revisit the long-standing problem of automatic reconstruction of 3D objects from single line drawings. Previous optimization-based methods can generate compact and accurate 3D models, but their success rates depend heavily on the ability to (i) identifying a sufficient set of true geometric constraints, and (ii) choosing a good initial value for the numerical optimization. In view of these challenges, we propose to train deep neural networks to detect pairwise relationships among geometric entities (i.e., edges) in the 3D object, and to predict initial depth value of the vertices. Our experiments on a large dataset of CAD models show that, by leveraging deep learning in a geometric constraint solving pipeline, the success rate of optimization-based 3D reconstruction can be significantly improved.

cs.CV↗

arXiv2022

Is Stochastic Gradient Descent Near Optimal?

The success of neural networks over the past decade has established them as effective models for many relevant data generating processes. Statistical theory on neural networks indicates graceful scaling of sample complexity. For example, Joen & Van Roy (arXiv:2203.00246) demonstrate that, when data is generated by a ReLU teacher network with $W$ parameters, an optimal learner needs only $\tilde{O}(W/ε)$ samples to attain expected error $ε$. However, existing computational theory suggests that, even for single-hidden-layer teacher networks, to attain small error for all such teacher networks, the computation required to achieve this sample complexity is intractable. In this work, we fit single-hidden-layer neural networks to data generated by single-hidden-layer ReLU teacher networks with parameters drawn from a natural distribution. We demonstrate that stochastic gradient descent (SGD) with automated width selection attains small expected error with a number of samples and total number of queries both nearly linear in the input dimension and width. This suggests that SGD nearly achieves the information-theoretic sample complexity bounds of Joen & Van Roy (arXiv:2203.00246) in a computationally efficient manner. An important difference between our positive empirical results and the negative theoretical results is that the latter address worst-case error of deterministic algorithms, while our analysis centers on expected error of a stochastic algorithm.

cs.LG↗

arXiv2022

Heterogeneous Graph-Based Multimodal Brain Network Learning

Graph neural networks (GNNs) provide powerful insights for brain neuroimaging technology from the view of graphical networks. However, most existing GNN-based models assume that the neuroimaging-produced brain connectome network is a homogeneous graph with single types of nodes and edges. In fact, emerging studies have reported and emphasized the significance of heterogeneity among human brain activities, especially between the two cerebral hemispheres. Thus, homogeneous-structured brain network-based graph methods are insufficient for modelling complicated cerebral activity states. To overcome this problem, in this paper, we present a heterogeneous graph neural network (HebrainGNN) for multimodal brain neuroimaging fusion learning. We first model the brain network as a heterogeneous graph with multitype nodes (i.e., left and right hemispheric nodes) and multitype edges (i.e., intra- and interhemispheric edges). Then, we propose a self-supervised pretraining strategy based on a heterogeneous brain network to address the potential overfitting problem caused by the conflict between a large parameter size and a small medical data sample size. Our results show the superiority of the proposed model over other existing methods in brain-related disease prediction tasks. Ablation experiments show that our heterogeneous graph-based model attaches more importance to hemispheric connections that may be neglected due to their low strength by previous homogeneous graph models. Other experiments also indicate that our proposed model with a pretraining strategy alleviates the problem of limited labelled data and yields a significant improvement in accuracy.

cs.LG↗

arXiv2022

Phononic Skyrmions

Skyrmions with topologically stable configurations have shown a promising route toward magnetic and photonic materials for information processing due to their defect-immune and low-driven energy. However, the practical application of magnetic skyrmions is severely hindered by their harsh cryogenic environment and complex carriers. In addition, the narrowband nature of magnetic and photonic skyrmions leads to lower data rate transmissions, restricting the development of high-speed information processing technologies. Here, we introduce and demonstrate the concept of phononic skyrmion as new topological structures to break the above barriers. The phononic skyrmion can be produced in any solid structure at room temperature, including chip-scale structures, with high robustness and ultra-bandwidth, which could pave a new path for high-speed and topological information processing technologies. We experimentally demonstrate the existence of phononic skyrmion formed by breaking the rotational symmetry of the three-dimensional hybrid spin of elastic waves. The frequency-independent spin configuration leads to the remarkable ultra-broadband and tunable feature of phononic skyrmions. We further experimentally show the excellent robustness of the flexibly movable phononic skyrmion lattices against local defects of disorder, sharp corners, and even rectangular holes. Our research also opens a vibrant horizon towards an unprecedented way for elastic wave manipulation and structuration by spin configuration, and offers a promising lever for alternative phononic technologies, including quantum information, biomedical testing, and wave engineering.

physics.app-ph↗

arXiv2022

AMinerGNN: Heterogeneous Graph Neural Network for Paper Click-through Rate Prediction with Fusion Query

Paper recommendation with user-generated keyword is to suggest papers that simultaneously meet user's interests and are relevant to the input keyword. This is a recommendation task with two queries, a.k.a. user ID and keyword. However, existing methods focus on recommendation according to one query, a.k.a. user ID, and are not applicable to solving this problem. In this paper, we propose a novel click-through rate (CTR) prediction model with heterogeneous graph neural network, called AMinerGNN, to recommend papers with two queries. Specifically, AMinerGNN constructs a heterogeneous graph to project user, paper, and keyword into the same embedding space by graph representation learning. To process two queries, a novel query attentive fusion layer is designed to recognize their importances dynamically and then fuse them as one query to build a unified and end-to-end recommender system. Experimental results on our proposed dataset and online A/B tests prove the superiority of AMinerGNN.

cs.IR↗

arXiv2022

Indexing Metric Spaces for Exact Similarity Search

With the continued digitization of societal processes, we are seeing an explosion in available data. This is referred to as big data. In a research setting, three aspects of the data are often viewed as the main sources of challenges when attempting to enable value creation from big data: volume, velocity, and variety. Many studies address volume or velocity, while fewer studies concern the variety. Metric spaces are ideal for addressing variety because they can accommodate any data as long as it can be equipped with a distance notion that satisfies the triangle inequality. To accelerate search in metric spaces, a collection of indexing techniques for metric data have been proposed. However, existing surveys offer limited coverage, and a comprehensive empirical study exists has yet to be reported. We offer a comprehensive survey of existing metric indexes that support exact similarity search: we summarize existing partitioning, pruning, and validation techniques used by metric indexes to support exact similarity search; we provide the time and space complexity analyses of index construction; and we offer an empirical comparison of their query processing performance. Empirical studies are important when evaluating metric indexing performance, because performance can depend highly on the effectiveness of available pruning and validation as well as on the data distribution, which means that complexity analyses often offer limited insights. This article aims at revealing strengths and weaknesses of different indexing techniques to offer guidance on selecting an appropriate indexing technique for a given setting, and to provide directions for future research on metric indexing.

cs.DB↗

arXiv2022

Hybrid integration of deterministic quantum dots-based single-photon sources with CMOS-compatible silicon carbide photonics

Thin film 4H-silicon carbide (4H-SiC) is emerging as a contender for realizing large-scale optical quantum circuits due to its high CMOS technology compatibility and large optical nonlinearities. Though, challenges remain in producing wafer-scale 4H-SiC thin film on insulator (4H-SiCOI) for dense integration of photonic circuits, and in efficient coupling of deterministic quantum emitters that are essential for scalable quantum photonics. Here we demonstrate hybrid integration of self-assembled InGaAs quantum dots (QDs) based single-photon sources (SPSs) with wafer-scale 4H-SiC photonic chips prepared by ion slicing technique. By designing a bilayer vertical coupler, we realize generation and highly efficient routing of single-photon emission in the hybrid quantum photonic chip. Furthermore, we realize a chip-integrated beamsplitter operation for triggered single photons through fabricating a 1x2 multi-mode interferometer (MMI) with a symmetric power splitting ratio of 50:50. The successful demonstration of heterogeneously integrating QDs-based SPSs on 4H-SiC photonic chip prepared by ion slicing technique constitutes an important step toward CMOS-compatible, fast reconfigurable quantum photonic circuits with deterministic SPSs.

physics.optics↗

arXiv2022

Excavation Reinforcement Learning Using Geometric Representation

Excavation of irregular rigid objects in clutter, such as fragmented rocks and wood blocks, is very challenging due to their complex interaction dynamics and highly variable geometries. In this paper, we adopt reinforcement learning (RL) to tackle this challenge and learn policies to plan for a sequence of excavation trajectories for irregular rigid objects, given point clouds of excavation scenes. Moreover, we separately learn a compact representation of the point cloud on geometric tasks that do not require human labeling. We show that using the representation reduces training time for RL, while achieving similar asymptotic performance compare to an end-to-end RL algorithm. When using a policy trained in simulation directly on a real scene, we show that the policy trained with the representation outperforms end-to-end RL. To our best knowledge, this paper presents the first application of RL to plan a sequence of excavation trajectories of irregular rigid objects in clutter.

cs.RO↗

arXiv2022

Automated Heart and Lung Auscultation in Robotic Physical Examinations

This paper presents the first implementation of autonomous robotic auscultation of heart and lung sounds. To select auscultation locations that generate high-quality sounds, a Bayesian Optimization (BO) formulation leverages visual anatomical cues to predict where high-quality sounds might be located, while using auditory feedback to adapt to patient-specific anatomical qualities. Sound quality is estimated online using machine learning models trained on a database of heart and lung stethoscope recordings. Experiments on 4 human subjects show that our system autonomously captures heart and lung sounds of similar quality compared to tele-operation by a human trained in clinical auscultation. Surprisingly, one of the subjects exhibited a previously unknown cardiac pathology that was first identified using our robot, which demonstrates the potential utility of autonomous robotic auscultation for health screening.

cs.RO↗

arXiv2021

Over-the-Air Aggregation for Federated Learning: Waveform Superposition and Prototype Validation

In this paper, we develop an orthogonal-frequency-division-multiplexing (OFDM)-based over-the-air (OTA) aggregation solution for wireless federated learning (FL). In particular, the local gradients in massive IoT devices are modulated by an analog waveform and are then transmitted using the same wireless resources. To this end, achieving perfect waveform superposition is the key challenge, which is difficult due to the existence of frame timing offset (TO) and carrier frequency offset (CFO). In order to address these issues, we propose a two-stage waveform pre-equalization technique with a customized multiple access protocol that can estimate and then mitigate the TO and CFO for the OTA aggregation. Based on the proposed solution, we develop a hardware transceiver and application software to train a real-world FL task, which learns a deep neural network to predict the received signal strength with global positioning system information. Experiments verify that the proposed OTA aggregation solution can achieve comparable performance to offline learning procedures with high prediction accuracy.

eess.SP↗

arXiv2021

Fully Three-dimensional Radial Visualization

We develop methodology for three-dimensional (3D) radial visualization (RadViz) of multidimensional datasets. The classical two-dimensional (2D) RadViz visualizes multivariate data in the 2D plane by mapping every observation to a point inside the unit circle. Our tool, RadViz3D, distributes anchor points uniformly on the 3D unit sphere. We show that this uniform distribution provides the best visualization with minimal artificial visual correlation for data with uncorrelated variables. However, anchor points can be placed exactly equi-distant from each other only for the five Platonic solids, so we provide equi-distant anchor points for these five settings, and approximately equi-distant anchor points via a Fibonacci grid for the other cases. Our methodology, implemented in the R package $radviz3d$, makes fully 3D RadViz possible and is shown to improve the ability of this nonlinear technique in more faithfully displaying simulated data as well as the crabs, olive oils and wine datasets. Additionally, because radial visualization is naturally suited for compositional data, we use RadViz3D to illustrate (i) the chemical composition of Longquan celadon ceramics and their Jingdezhen imitation over centuries, and (ii) US regional SARS-Cov-2 variants' prevalence in the Covid-19 pandemic during the summer 2021 surge of the Delta variant.

stat.ME↗

arXiv2021

Visualization of Labeled Mixed-featured Datasets

We develop methodology for visualization of labeled mixed-featured datasets. We first investigate datasets with continuous features where our Max-Ratio Projection (MRP) method utilizes the group information in high dimensions to provide distinctive lower-dimensional projections that are then displayed using Radviz3D. Our methodology is extended to datasets with discrete and continuous features where a Gaussianized distributional transform is used in conjunction with copula models before applying MRP and visualizing the result using RadViz3D. A R package $radviz3d$ implementing our complete methodology is available.

stat.ML↗

arXiv2021

Multi-window SRS imaging using a rapid widely tunable fiber laser

Spectroscopic stimulated Raman scattering (SRS) imaging has become a useful tool finding a broad range of applications. Yet, wider adoption is hindered by the bulky and environmentally-sensitive solid-state optical parametric oscillator (OPO) in current SRS microscope. Moreover, chemically-informative multi-window SRS imaging across C-H, C-D and fingerprint Raman regions is challenging due to the slow wavelength tuning speed of the solid-state OPO. In this work, we present a multi-window SRS imaging system based on a compact and robust fiber laser with rapid and widely tuning capability. To address the relative intensity noise intrinsic to fiber laser, we implemented auto-balanced detection which enhances the signal-to-noise ratio of stimulated Raman loss imaging by 23 times. We demonstrate high-quality SRS metabolic imaging of fungi, cancer cells, and Caenorhabditis elegans across the C-H, C-D and fingerprint Raman windows. Our re-sults showcase the potential of the compact multi-window SRS system for a broad range of applications.

physics.optics↗

arXiv2021

An Ultrahigh-Q Microresonator on 4H-silicon-carbide-on-insulator Platform for Multiple Harmonics, Cascaded Raman Lasing and Kerr Comb Generations

The realization of ultrahigh quality (Q) resonators regardless of the underpinning material platforms has been a ceaseless pursuit, because the high Q resonators provide an extreme environment of storage of light to enable observations of many unconventional nonlinear optical phenomenon with high efficiencies. Here, we demonstrate an ultra-high Q factor (7.1*10^6) microresonator on the 4H-silicon-carbide-on-insulator (4H-SiCOI) platform in which both \c{hi}^(2) and \c{hi}^(3) nonlinear processes of high efficiencies have been generated. Broadband frequency conversions, including second-, third-, fourth-harmonic generation were observed. Cascaded Raman lasing was demonstrated in the SiC microresonator for the first time to the best of our knowledge. Broadband Kerr frequency combs covering from 1300 to 1700 nm were achieved using a dispersion-engineered SiC microresonator. Our demonstration is a significant milestone in the development of SiC photonic integrated devices.

physics.optics↗

arXiv2021

Elastic Bound State in the Continuum with Perfect Mode Conversion

The partial or complete confinement of waves in an open system is omnipresent in nature and in wave-based materials and technology. Here, we theoretically analyze and experimentally observe the formation of a trapped mode with perfect mode conversion (TMPC) between flexural waves and longitudinal waves, by achieving a quasi-bound state in the continuum (BIC) in an open elastic wave system. The latter allows a quasi-BIC in a semi-infinite background plate when Fano resonance hybridizes flexural and longitudinal waves and balances their radiative decay rates. We demonstrate that when the Fabry-Pérot resonance of the longitudinal wave is realized simultaneously, the TMPC formed by the elastic BIC approaches infinite quality factor. Furthermore, we show that quasi-BIC can be tuned continuously to BIC through the critical frequency of mode conversion, which offers the possibility of TMPC with an arbitrarily high quality factor. Our reported concept and physical mechanism open new routes to achieve perfect mode conversion with tunable high quality factor in elastic systems.

physics.app-ph↗

arXiv2021

Visible and Near-infrared Microdisk Resonators on a 4H-Silicon-Carbide-on-Insulator Platform

Wavelength-sized microdisk resonators were fabricated on a single crystalline 4H-silicon-carbide-oninsulator platform (4H-SiCOI). By carrying out microphotoluminescence measurements at room temperature, we show that the microdisk resonators support whispering-gallery modes (WGMs) with quality factors up to $5.25 \times 10^3$ and mode volumes down to $2.69 \times(λ/n)^3$ at the visible and near-infrared wavelengths. Moreover, the demonstrated wavelength-sized microdisk resonators exhibit WGMs whose resonant wavelengths compatible with the zero-phonon lines of spin defects in 4H-SiCOI, making them a promising candidate for applications in cavity quantum electrodynamics and integrated quantum photonic circuits.

physics.optics↗