arXiv subjects

Hang Yu

Publications and source records attributed to Hang Yu.

At least 163 records · Page 9Linked to original sources

arXiv2022

Anti-Delay Kalman Filter Fusion Algorithm for Vehicle-borne Sensor Network with Finite-Time Convergence

Intelligent vehicles in autonomous driving and obstacle avoidance, the precise relative state of vehicles put forward a higher demand. For a vehicle-borne sensor network with time-varying transmission delays, the problem of coordinate fusion of vehicle state is the focus of this paper. By the ingeniously designed low-complexity integration with a consensus strategy and buffer technology, an anti-delay distributed Kalman filter (DKF) with finite-time convergence is proposed.By introducing the matrix weight to assess local estimates, the optimal fusion state result is available in the sense of linear minimum variance. In addition, to accommodate practical engineering in intelligent vehicles, the communication weight coefficient and directed topology with unidirectional transmission are also considered. From a theoretical perspective, the proof of error covariances upper bounds with different communication topologies with delays are presented. Furthermore, the maximum allowable delays of vehicle-borne sensor network is derived backwards. Simulations verify that while considering various non-ideal factors above, the proposed DFK algorithm produces more accurate and robust fusion estimation state results than existing algorithms, making it more valuable in practical applications. Simultaneously, a mobile car trajectory tracking experiment is carried out, which further verifies the feasibility of the proposed algorithm.

eess.SY↗

arXiv2022

Rethinking Task Sampling for Few-shot Vision-Language Transfer Learning

Despite achieving state-of-the-art zero-shot performance, existing vision-language models still fall short of few-shot transfer ability on domain-specific problems. Classical fine-tuning often fails to prevent highly expressive models from exploiting spurious correlations. Although model-agnostic meta-learning (MAML) presents as a natural alternative for few-shot transfer learning, the expensive computation due to implicit second-order optimization limits its use on large-scale vision-language models such as CLIP. While much literature has been devoted to exploring alternative optimization strategies, we identify another essential aspect towards effective few-shot transfer learning, task sampling, which is previously only be viewed as part of data pre-processing in MAML. To show the impact of task sampling, we propose a simple algorithm, Model-Agnostic Multitask Fine-tuning (MAMF), which differentiates classical fine-tuning only on uniformly sampling multiple tasks. Despite its simplicity, we show that MAMF consistently outperforms classical fine-tuning on five few-shot vision-language classification tasks. We further show that the effectiveness of the bi-level optimization in MAML is highly sensitive to the zero-shot performance of a task in the context of few-shot vision-language classification. The goal of this paper is to provide new insights on what makes few-shot learning work, and encourage more research into investigating better task sampling strategies.

cs.MM↗

arXiv2022

Uncertainty and bias of cosmology and astrophysical population model from statistical dark sirens

Gravitational-wave (GW) radiation from a coalescing compact binary is a standard siren as the luminosity distance of each event can be directly measured from the amplitude of the signal. One possibility to constrain cosmology using the GW siren is to perform statistical inference on a population of binary black hole (BBH) events. In essence, this statistical method can be viewed as follows. We can modify the shape of the distribution of observed BBH events by changing cosmological parameters until it eventually matches the distribution constructed from an astrophysical population model, thereby allowing us to determine the cosmological parameters. In this work, we derive the Cramér-Rao bound for both cosmological parameters and those governing the astrophysical population model from this statistical dark siren method by examining the Fisher information contained in the event distribution. Our study provides analytical insights and enables fast yet accurate estimations of the statistical accuracy of dark siren cosmology. Furthermore, we consider the bias in cosmology due to unmodeled substructures in the merger rate and the mass distribution. We find a $1\%$ deviation in the astrophysical model can lead to a more than $1\%$ error in the Hubble constant. This could limit the accuracy of dark siren cosmology when there are more than $10^4$ BBH events detected.

astro-ph.CO↗

arXiv2022

Exposing Gravitational Waves below the Quantum Shot Noise

The sensitivities of ground-based gravitational-wave (GW) detectors are limited by quantum shot noise at a few hundred Hertz and above. Nonetheless, one can use a quantum-correlation technique proposed by Martynov, et al. [Phys. Rev. A 95, 043831 (2017)] to remove the expectation value of the shot noise, thereby exposing underlying classical signals in the cross spectrum formed by cross-correlating the two outputs in a GW interferometer's anti-symmetric port. We explore here the prospects and analyze the sensitivity of using quantum correlation to detect astrophysical GW signals. Conceptually, this technique is similar to the correlation of two different GW detectors as it utilizes the fact that a GW signal will be correlated in the two outputs but the shot noise will be uncorrelated. Quantum correlation also has its unique advantages as it requires only a single interferometer to make a detection. Therefore, quantum correlation could increase the duty cycle, enhance the search efficiency, and enable the detection of highly polarized signals. In particular, we show that quantum correlation could be especially useful for detecting post-merger remnants of binary neutron stars with both short ($< 1\,{\rm s}$) and intermediate ($\sim 10-10^4\,{\rm s}$) durations and setting upper limits on continuous emissions from unknown pulsars.

gr-qc↗

arXiv2022

Controllable Production of Degenerate Fermi Gases of $^6$Li Atoms in the 2D-3D Crossover

The many-body physics in the dimensional crossover regime attracts much attention in cold atom experiments, but yet to explore systematically. One of the technical difficulties existed in the experiments is the lack of the experimental technique to quantitatively tune the atom occupation ratio of the different lattice bands. In this letter, we report such techniques in a process of transferring a 3D Fermi gas into a 1D optical lattice, where the capability of tuning the occupation of the energy band is realized by varying the trapping potentials of the optical dipole trap (ODT) and the lattice, respectively. We could tune a Fermi gas with the occupation in the lowest band from unity to 50$\%$ quantitatively. This provides a route to experimentally study the dependence of many-body interaction on the dimensionality in a Fermi gas.

cond-mat.quant-gas↗

arXiv2022

Impact of Side Chain Hydrophilicity on Packing, Swelling and Ion Interactions in Oxy-bithiophene Semiconductors

Exchanging hydrophobic alkyl-based side chains to hydrophilic glycol-based side chains is a widely adopted method for improving mixed-transport device performance, despite the impact on solid state packing and polymer-electrolyte interactions being poorly understood. Presented here is a Molecular Dynamics (MD) force field for modelling alkoxylated and glycolated polythiophenes. The force field is validated against known packing motifs for their monomer crystals. MD simulations, coupled with X-ray Diffraction (XRD), show that alkoxylated polythiophenes will pack with a `tilted stack' and straight interdigitating side chains, whilst their glycolated counterpart will pack with a `deflected stack' and an s-bend side chain configuration. MD simulations reveal water penetration pathways into the alkoxylated and glycolated crystals - through the π-stack and through the lamellar stack respectively. Finally, the two distinct ways tri-ethylene glycol polymers can bind to cations are revealed, showing the formation of a meta-stable single bound state, or an energetically deep double bound state, both with a strong side chain length dependance. The minimum energy pathways for the formation of the chelates are identified, showing the physical process through which cations can bind to one or two side chains of a glycolated polythiophene, with consequences for ion transport in bithiophene semiconductors.

cond-mat.soft↗

arXiv2021

Tidal evolution and diffusive growth during high-eccentricity planet migration: revisiting the eccentricity distribution of hot Jupiters

High-eccentricity tidal migration is a potential formation channel for hot Jupiters. During this process, the planetary f-mode may experience a phase of diffusive growth, allowing its energy to quickly build up to large values. In Yu et al. (2021, ApJ, 917, 31), we demonstrated that nonlinear mode interactions between a parent f-mode and daughter f- and p-modes expand the parameter space over which the diffusive growth of the parent is triggered. We extend that study by incorporating (1) the angular momentum transfer between the orbit and the mode, and consequently the evolution of the pericenter distance, (2) a phenomenological correction to the nonlinear frequency shift at high parent mode energies, and (3) dissipation of the parent's energy due to both turbulent convective damping of the daughter modes and strongly nonlinear wave-breaking events. The new ingredients allow us to follow the coupled evolution of the mode and orbit over $\gtrsim 10^4$ years, covering the diffusive evolution from its onset to its termination. We find that the semi-major axis shrinks by a factor of nearly ten over $10^4$ years, corresponding to a tidal quality factor $\mathcal{Q}\sim10$. The f-mode's diffusive growth terminates while the eccentricity is still high, at around e=0.8-0.95. Using these results, we revisit the eccentricity distribution of proto-hot Jupiters. We estimate that less than 1 proto-HJ with eccentricity >0.9 should be expected in Kepler's data once the diffusive regime is accounted for, explaining the observed paucity of this population.

astro-ph.EP↗

arXiv2021

Nonlinear noise regression in gravitational-wave detectors with convolutional neural networks

Currently, the sub-60 Hz sensitivity of gravitational-wave (GW) detectors like Advanced LIGO is limited by the control noises from auxiliary degrees of freedom, which nonlinearly couple to the main GW readout. One particularly promising way to tackle this contamination is to perform nonlinear noise mitigation using machine-learning-based convolutional neural networks (CNNs), which we examine in detail in this study. As in many cases the noise coupling is bilinear and can be viewed as a few fast channels' outputs modulated by some slow channels, we show that we can utilize this knowledge of the physical system and adopt an explicit "slow$\times$fast" structure in the design of the CNN to enhance its performance of noise subtraction. We then examine the requirement in the signal-to-noise ratio (SNR) in both the target channel (i.e., the main GW readout) and in the auxiliary sensors in order to reduce the noise by at least a factor of a few. In the case of limited SNR in the target channel, we further demonstrate that the CNN can still reach a good performance if we adopt curriculum learning techniques, which in reality can be achieved by combining data from quiet times and those from periods with active noise injections.

astro-ph.IM↗

arXiv2021

AP-10K: A Benchmark for Animal Pose Estimation in the Wild

Accurate animal pose estimation is an essential step towards understanding animal behavior, and can potentially benefit many downstream applications, such as wildlife conservation. Previous works only focus on specific animals while ignoring the diversity of animal species, limiting the generalization ability. In this paper, we propose AP-10K, the first large-scale benchmark for mammal animal pose estimation, to facilitate the research in animal pose estimation. AP-10K consists of 10,015 images collected and filtered from 23 animal families and 54 species following the taxonomic rank and high-quality keypoint annotations labeled and checked manually. Based on AP-10K, we benchmark representative pose estimation models on the following three tracks: (1) supervised learning for animal pose estimation, (2) cross-domain transfer learning from human pose estimation to animal pose estimation, and (3) intra- and inter-family domain generalization for unseen animals. The experimental results provide sound empirical evidence on the superiority of learning from diverse animals species in terms of both accuracy and generalization ability. It opens new directions for facilitating future research in animal pose estimation. AP-10k is publicly available at https://github.com/AlexTheBad/AP10K.

cs.CV↗

arXiv2021

CvT-ASSD: Convolutional vision-Transformer Based Attentive Single Shot MultiBox Detector

Due to the success of Bidirectional Encoder Representations from Transformers (BERT) in natural language process (NLP), the multi-head attention transformer has been more and more prevalent in computer-vision researches (CV). However, it still remains a challenge for researchers to put forward complex tasks such as vision detection and semantic segmentation. Although multiple Transformer-Based architectures like DETR and ViT-FRCNN have been proposed to complete object detection task, they inevitably decreases discrimination accuracy and brings down computational efficiency caused by the enormous learning parameters and heavy computational complexity incurred by the traditional self-attention operation. In order to alleviate these issues, we present a novel object detection architecture, named Convolutional vision Transformer Based Attentive Single Shot MultiBox Detector (CvT-ASSD), that built on the top of Convolutional vision Transormer (CvT) with the efficient Attentive Single Shot MultiBox Detector (ASSD). We provide comprehensive empirical evidence showing that our model CvT-ASSD can leads to good system efficiency and performance while being pretrained on large-scale detection datasets such as PASCAL VOC and MS COCO. Code has been released on public github repository at https://github.com/albert-jin/CvT-ASSD.

cs.CV↗

arXiv2021

Circular arc polygons, numerical conformal mappings, and moduli of quadrilaterals

We study numerical conformal mappings of planar Jordan domains with boundaries consisting of finitely many circular arcs and compute the moduli of quadrilaterals for these domains. Experimental error estimates are provided and, when possible, comparison to exact values or other methods are given. The main ingredients of the computation are boundary integral equations combined with the fast multipole method.

math.CV↗

arXiv2021

Approaching the motional ground state of a 10 kg object

The motion of a mechanical object -- even a human-sized object -- should be governed by the rules of quantum mechanics. Coaxing them into a quantum state is, however, difficult: the thermal environment masks any quantum signature of the object's motion. Indeed, the thermal environment also masks effects of proposed modifications of quantum mechanics at large mass scales. We prepare the center-of-mass motion of a 10 kg mechanical oscillator in a state with an average phonon occupation of 10.8. The reduction in temperature, from room temperature to 77 nK, is commensurate with an 11 orders-of-magnitude suppression of quantum back-action by feedback -- and a 13 orders-of-magnitude increase in the mass of an object prepared close to its motional ground state. This begets the possibility of probing gravity on massive quantum systems.

quant-ph↗

arXiv2021

Fixed-time Cooperative Guidance Law for Multiple Missiles in Simultaneous Attacks

To enable multiple missiles to attack a maneuvering target simultaneously, fixed-time distributed cooperative guidance laws are proposed in this paper. Here, we present a novel fixed-time fast nonsingular terminal sliding mode surface (FNTSMS). In particular, the sliding mode surface not only avoids singularities but also has the characteristic of a settling time boundary regardless of the initial conditions. Based on the FNTSMS, we have developed a distributed guidance law, which has the characteristic of fixed-time convergence. The guidance law achieves the consensus of range-to-go, relative velocity along and perpendicular to the line of sight (LOS) direction to realize the simultaneous attack. In addition, a saturation function is introduced to avoid the chattering problem caused by the commonly used sign function. Furthermore, the distributed cooperative guidance law with communication failure is considered and proved theoretically, which shows that the proposed guidance law has excellent performance. Finally, the simulation results verify the performance of the distributed guidance law and its robustness against communication topology mutations, and explain the phenomenon in detail.

math.DS↗

arXiv2021

BWCP: Probabilistic Learning-to-Prune Channels for ConvNets via Batch Whitening

This work presents a probabilistic channel pruning method to accelerate Convolutional Neural Networks (CNNs). Previous pruning methods often zero out unimportant channels in training in a deterministic manner, which reduces CNN's learning capacity and results in suboptimal performance. To address this problem, we develop a probability-based pruning algorithm, called batch whitening channel pruning (BWCP), which can stochastically discard unimportant channels by modeling the probability of a channel being activated. BWCP has several merits. (1) It simultaneously trains and prunes CNNs from scratch in a probabilistic way, exploring larger network space than deterministic methods. (2) BWCP is empowered by the proposed batch whitening tool, which is able to empirically and theoretically increase the activation probability of useful channels while keeping unimportant channels unchanged without adding any extra parameters and computational cost in inference. (3) Extensive experiments on CIFAR-10, CIFAR-100, and ImageNet with various network architectures show that BWCP outperforms its counterparts by achieving better accuracy given limited computational budgets. For example, ResNet50 pruned by BWCP has only 0.70\% Top-1 accuracy drop on ImageNet, while reducing 43.1\% FLOPs of the plain ResNet50.

cs.LG↗

arXiv2021

Automatic Learning to Detect Concept Drift

Many methods have been proposed to detect concept drift, i.e., the change in the distribution of streaming data, due to concept drift causes a decrease in the prediction accuracy of algorithms. However, the most of current detection methods are based on the assessment of the degree of change in the data distribution, cannot identify the type of concept drift. In this paper, we propose Active Drift Detection with Meta learning (Meta-ADD), a novel framework that learns to classify concept drift by tracking the changed pattern of error rates. Specifically, in the training phase, we extract meta-features based on the error rates of various concept drift, after which a meta-detector is developed via a prototypical neural network by representing various concept drift classes as corresponding prototypes. In the detection phase, the learned meta-detector is fine-tuned to adapt to the corresponding data stream via stream-based active learning. Hence, Meta-ADD uses machine learning to learn to detect concept drifts and identify their types automatically, which can directly support drift understand. The experiment results verify the effectiveness of Meta-ADD.

cs.AI↗

arXiv2021

Early warning of coalescing neutron-star and neutron-star-black-hole binaries from nonstationary noise background using neural networks

The success of the multi-messenger astronomy relies on gravitational-wave observatories like LIGO and Virgo to provide prompt warning of merger events involving neutron stars (including both binary neutron stars and neutron-star-black-holes), which further depends critically on the low-frequency sensitivity of LIGO as a typical binary neutron star stays in this band for minutes. However, the current sub-60 Hz sensitivity of LIGO has not yet reached its design target and the excess noise can be more than an order of magnitude below 20 Hz. It is limited by nonlinearly coupled noises from auxiliary control loops which are also nonstationary, posing challenges to realistic early-warning pipelines. Nevertheless, machine-learning-based neural networks provide ways to simultaneously improve the low-frequency sensitivity and mitigate its nonstationarity, and detect the real-time gravitational-wave signal with a very short computational time. We propose to achieve this by inputting both the main gravitational-wave readout and key auxiliary witnesses to a compound neural network. Using simulated data with characteristic representing the real LIGO detectors, our machine-learning-based neural networks can reduce nonlinearly coupled noise by about a factor of 5 and allows a typical binary neutron star (neutron-star-black-hole) to be detected 100 s (10 s) before the merger at a distance of 40 Mpc (160 Mpc). If one can further reduce the noise to the fundamental limit, our neural networks can achieve detection out to a distance of 80 Mpc and 240 Mpc for binary neutron stars and neutron-star-black-holes, respectively. It thus demonstrates that utilizing machine-learning-based neural networks is a promising direction for the timely detection of the coalescence of electromagnetically bright LIGO/Virgo sources.

gr-qc↗

arXiv2021

Tides in the high-eccentricity migration of hot Jupiters: Triggering diffusive growth by nonlinear mode interactions

High eccentricity migration is a possible formation channel for hot Jupiters. However, in order for it to be consistent with the observed population of planets, tides must circularize the orbits in less than $\approx$ a Myr. A potential mechanism for such rapid circularization is the diffusive growth of the tidally driven planetary f-mode. Such growth occurs if the f-mode's phase at pericenter varies chaotically from one pericenter passage to the next. Previous studies focused on the variation of the orbital period due to tidal back-reaction on the orbit as the source of chaos. Here we show that nonlinear mode interactions can also be an important source. Specifically, we show that nonlinear interactions between a parent f-mode and daughter f-/p-modes induce an energy-dependent shift in the oscillation frequency of the parent. This frequency shift varies randomly from orbit to orbit because the parent's energy varies. As a result, the parent's phase at pericenter varies randomly, which we find can trigger it to grow diffusively. We show that the phase shift induced by nonlinear mode interactions in fact dominates the shift induced by tidal back-reaction and significantly lowers the one-kick energy threshold for diffusive growth by about a factor of 5 compared to the linear theory's prediction. Nonlinear interactions could thus enhance the formation rate of hot Jupiters through the high-eccentricity migration channel and potentially mitigate the discrepancy between the observed and predicted occurrence rates for close-in gas giants as compared to those further from the star.

astro-ph.EP↗

arXiv2021

A General Framework of Nonparametric Feature Selection in High-Dimensional Data

Nonparametric feature selection in high-dimensional data is an important and challenging problem in statistics and machine learning fields. Most of the existing methods for feature selection focus on parametric or additive models which may suffer from model misspecification. In this paper, we propose a new framework to perform nonparametric feature selection for both regression and classification problems. In this framework, we learn prediction functions through empirical risk minimization over a reproducing kernel Hilbert space. The space is generated by a novel tensor product kernel which depends on a set of parameters that determine the importance of the features. Computationally, we minimize the empirical risk with a penalty to estimate the prediction and kernel parameters at the same time. The solution can be obtained by iteratively solving convex optimization problems. We study the theoretical property of the kernel feature space and prove both the oracle selection property and the Fisher consistency of our proposed method. Finally, we demonstrate the superior performance of our approach compared to existing methods via extensive simulation studies and application to a microarray study of eye disease in animals.

stat.ME↗