arXiv subjects

Kuan-Wen Chen

Publications and source records attributed to Kuan-Wen Chen.

27 records · Page 2Linked to original sources

arXiv2022

Neural Architecture Search using Progressive Evolution

Vanilla neural architecture search using evolutionary algorithms (EA) involves evaluating each architecture by training it from scratch, which is extremely time-consuming. This can be reduced by using a supernet to estimate the fitness of every architecture in the search space due to its weight sharing nature. However, the estimated fitness is very noisy due to the co-adaptation of the operations in the supernet. In this work, we propose a method called pEvoNAS wherein the whole neural architecture search space is progressively reduced to smaller search space regions with good architectures. This is achieved by using a trained supernet for architecture evaluation during the architecture search using genetic algorithm to find search space regions with good architectures. Upon reaching the final reduced search space, the supernet is then used to search for the best architecture in that search space using evolution. The search is also enhanced by using weight inheritance wherein the supernet for the smaller search space inherits its weights from previous trained supernet for the bigger search space. Exerimentally, pEvoNAS gives better results on CIFAR-10 and CIFAR-100 while using significantly less computational resources as compared to previous EA-based methods. The code for our paper can be found in https://github.com/nightstorm0909/pEvoNAS

cs.NE↗

arXiv2021

Neural Architecture Search using Covariance Matrix Adaptation Evolution Strategy

Evolution-based neural architecture search requires high computational resources, resulting in long search time. In this work, we propose a framework of applying the Covariance Matrix Adaptation Evolution Strategy (CMA-ES) to the neural architecture search problem called CMANAS, which achieves better results than previous evolution-based methods while reducing the search time significantly. The architectures are modelled using a normal distribution, which is updated using CMA-ES based on the fitness of the sampled population. We used the accuracy of a trained one shot model (OSM) on the validation data as a prediction of the fitness of an individual architecture to reduce the search time. We also used an architecture-fitness table (AF table) for keeping record of the already evaluated architecture, thus further reducing the search time. CMANAS finished the architecture search on CIFAR-10 with the top-1 test accuracy of 97.44% in 0.45 GPU day and on CIFAR-100 with the top-1 test accuracy of 83.24% for 0.6 GPU day on a single GPU. The top architectures from the searches on CIFAR-10 and CIFAR-100 were then transferred to ImageNet, achieving the top-5 accuracy of 92.6% and 92.1%, respectively.

cs.NE↗

arXiv2021

Unusual high-field metal in a Kondo insulator

Within condensed-matter systems, strong electronic interactions often lead to exotic quantum phases. A recent manifestation of this is the unexpected observation of magnetic quantum oscillations and metallic thermal transport, both properties of systems with Fermi surfaces of itinerant quasiparticles, in the Kondo insulators SmB6 and YbB$_{12}$. To understand these phenomena, it is informative to study their evolution as the energy gap of the Kondo-Insulator state is closed by a large magnetic field. We show here that both the quantum-oscillation frequency and the cyclotron mass display a strong field dependence in the resulting high-field metallic state in $_{12}$. By tracking the Fermi-surface area, we conclude that the same quasiparticle band gives rise to the quantum oscillations in both insulating and metallic states. These data are understood most simply using a two-fluid picture where unusual quasiparticles, contributing little or nothing to charge transport, coexist with conventional fermions. In the metallic state this leads to a heavy-fermion bad metal with negligible magnetoresistance, relatively high resistivity and a very large Kadowaki-Woods ratio, underlining the exotic nature of the fermion ensemble inhabiting $_{12}$.

cond-mat.str-el↗

arXiv2020

Evolving Neural Architecture Using One Shot Model

Neural Architecture Search (NAS) is emerging as a new research direction which has the potential to replace the hand-crafted neural architectures designed for specific tasks. Previous evolution based architecture search requires high computational resources resulting in high search time. In this work, we propose a novel way of applying a simple genetic algorithm to the NAS problem called EvNAS (Evolving Neural Architecture using One Shot Model) which reduces the search time significantly while still achieving better result than previous evolution based methods. The architectures are represented by using the architecture parameter of the one shot model which results in the weight sharing among the architectures for a given population of architectures and also weight inheritance from one generation to the next generation of architectures. We propose a decoding technique for the architecture parameter which is used to divert majority of the gradient information towards the given architecture and is also used for improving the performance prediction of the given architecture from the one shot model during the search process. Furthermore, we use the accuracy of the partially trained architecture on the validation data as a prediction of its fitness in order to reduce the search time. EvNAS searches for the architecture on the proxy dataset i.e. CIFAR-10 for 4.4 GPU day on a single GPU and achieves top-1 test error of 2.47% with 3.63M parameters which is then transferred to CIFAR-100 and ImageNet achieving top-1 error of 16.37% and top-5 error of 7.4% respectively. All of these results show the potential of evolutionary methods in solving the architecture search problem.

cs.NE↗

arXiv2020

IF-Net: An Illumination-invariant Feature Network

Feature descriptor matching is a critical step is many computer vision applications such as image stitching, image retrieval and visual localization. However, it is often affected by many practical factors which will degrade its performance. Among these factors, illumination variations are the most influential one, and especially no previous descriptor learning works focus on dealing with this problem. In this paper, we propose IF-Net, aimed to generate a robust and generic descriptor under crucial illumination changes conditions. We find out not only the kind of training data important but also the order it is presented. To this end, we investigate several dataset scheduling methods and propose a separation training scheme to improve the matching accuracy. Further, we propose a ROI loss and hard-positive mining strategy along with the training scheme, which can strengthen the ability of generated descriptor dealing with large illumination change conditions. We evaluate our approach on public patch matching benchmark and achieve the best results compared with several state-of-the-arts methods. To show the practicality, we further evaluate IF-Net on the task of visual localization under large illumination changes scenes, and achieves the best localization accuracy.

cs.CV↗

arXiv2020

Multiple Dirac Nodes and Symmetry Protected Dirac Nodal Line in Orthorhombic $α$-RhSi

Owing to their chiral cubic structure, exotic multifold topological excitations have been predicted and recently observed in transition metal silicides like $β$-RhSi. Herein, we report that the topological character of RhSi is also observed in its orthorhombic $α$-phase which displays multiple types of Dirac nodes very close to the Fermi level ($\varepsilon_F$) with the near absence of topologically trivial carriers. We discuss the symmetry analysis, band connectivity along high-symmetry lines using group representations, the band structure, and the nature of the Dirac points and nodal lines occurring near $\varepsilon_F$. The de Haas-van Alphen effect (dHvA) indicates a Fermi surface in agreement with the calculations. We find an elliptically-shaped nodal line very close to $\varepsilon_F$ around and near the $S$-point on the $k_y-k_z$ plane that results from the intersection of two upside-down Dirac cones. The two Dirac points of the participating Kramers degenerate bands are only 5 meV apart, hence an accessible magnetic field might induce a crossing between the spin-up partner of the upper-Dirac cone and the spin-down partner of the lower Dirac cone, possibly explaining the anomalies observed in the magnetic torque.

cond-mat.str-el↗

arXiv2017

Magnetic Properties of New Triangular Lattice Magnets A${_4}$B'B${_2}$O$_{12}$

The geometrically frustrated two dimensional triangular lattice magnets A${_4}$B'B${_2}$O$_{12}$ (A = Ba, Sr, La; B' = Co, Ni, Mn; B = W, Re) have been studied by x-ray diffraction, AC and DC susceptibilities, powder neutron diffraction, and specific heat measurements. The results reveal that (i) the samples containing Co$^{2+}$ (effective spin-1/2) and Ni$^{2+}$ (spin-1) ions with small spin numbers exhibit ferromagnetic (FM) ordering while the sample containing Mn$^{2+}$ (spin-5/2) ions with a large spin number exhibits antiferromagnetic (AFM) ordering. We ascribe these spin number manipulated ground states to the competition between the AFM B'-O-O-B' and FM B'-O-B-O-B' superexchange interactions; (ii) the chemical pressure introduced into the Co containing samples through the replacement of different size ions on the A site finely tunes the FM ordering temperature of the system. We attribute this effect to the modification of the FM interaction strength induced by the change of the O-B-O angle through chemical pressure.

cond-mat.mtrl-sci↗

arXiv2015

Non-stoichiometry and Defects in the Weyl Semimetals TaAs, TaP, NbP, and NbAs

We report a structural study of the Weyl semimetals TaAs, TaP, NbP, and NbAs, utilizing diffraction techniques (single crystal x-ray diffraction and energy dispersive spectroscopy) and imaging techniques (transmission electron microscopy/scanning transmission electron microscopy). We observe defects of various degrees, leading to non-stoichiometric single crystals of all four semimetals. While TaP displays a large pnictide deficiency with composition TaP$_{0.83(3)}$, and stacking faults accompanied by anti-site disorder and site vacancies, TaAs displays transition metal deficiency with composition Ta$_{0.92(2)}$As and a high density of stacking faults. NbP also displays pnictide deficiency, yielding composition NbP$_{0.95(2)}$, and lastly, NbAs display very little deviation from a 1:1 composition, NbAs$_{1.00(3)}$, and is therefore recommended to serve as the model compound for these semimetals.

cond-mat.mtrl-sci↗

arXiv2015

To Know Where We Are: Vision-Based Positioning in Outdoor Environments

Augmented reality (AR) displays become more and more popular recently, because of its high intuitiveness for humans and high-quality head-mounted display have rapidly developed. To achieve such displays with augmented information, highly accurate image registration or ego-positioning are required, but little attention have been paid for out-door environments. This paper presents a method for ego-positioning in outdoor environments with low cost monocular cameras. To reduce the computational and memory requirements as well as the communication overheads, we formulate the model compression algorithm as a weighted k-cover problem for better preserving model structures. Specifically for real-world vision-based positioning applications, we consider the issues with large scene change and propose a model update algorithm to tackle these problems. A long- term positioning dataset with more than one month, 106 sessions, and 14,275 images is constructed. Based on both local and up-to-date models constructed in our approach, extensive experimental results show that high positioning accuracy (mean ~ 30.9cm, stdev. ~ 15.4cm) can be achieved, which outperforms existing vision-based algorithms.

cs.CV↗