arXiv subjects

Shiyu Li

Publications and source records attributed to Shiyu Li.

63 records · Page 4Linked to original sources

arXiv2021

Neural Image Beauty Predictor Based on Bradley-Terry Model

Image beauty assessment is an important subject of computer vision. Therefore, building a model to mimic the image beauty assessment becomes an important task. To better imitate the behaviours of the human visual system (HVS), a complete survey about images of different categories should be implemented. This work focuses on image beauty assessment. In this study, the pairwise evaluation method was used, which is based on the Bradley-Terry model. We believe that this method is more accurate than other image rating methods within an image group. Additionally, Convolution neural network (CNN), which is fit for image quality assessment, is used in this work. The first part of this study is a survey about the image beauty comparison of different images. The Bradley-Terry model is used for the calculated scores, which are the target of CNN model. The second part of this work focuses on the results of the image beauty prediction, including landscape images, architecture images and portrait images. The models are pretrained by the AVA dataset to improve the performance later. Then, the CNN model is trained with the surveyed images and corresponding scores. Furthermore, this work compares the results of four CNN base networks, i.e., Alex net, VGG net, Squeeze net and LSiM net, as discussed in literature. In the end, the model is evaluated by the accuracy in pairs, correlation coefficient and relative error calculated by survey results. Satisfactory results are achieved by our proposed methods with about 70 percent accuracy in pairs. Our work sheds more light on the novel image beauty assessment method. While more studies should be conducted, this method is a promising step.

cs.CV↗

arXiv2020

NASGEM: Neural Architecture Search via Graph Embedding Method

Neural Architecture Search (NAS) automates and prospers the design of neural networks. Estimator-based NAS has been proposed recently to model the relationship between architectures and their performance to enable scalable and flexible search. However, existing estimator-based methods encode the architecture into a latent space without considering graph similarity. Ignoring graph similarity in node-based search space may induce a large inconsistency between similar graphs and their distance in the continuous encoding space, leading to inaccurate encoding representation and/or reduced representation capacity that can yield sub-optimal search results. To preserve graph correlation information in encoding, we propose NASGEM which stands for Neural Architecture Search via Graph Embedding Method. NASGEM is driven by a novel graph embedding method equipped with similarity measures to capture the graph topology information. By precisely estimating the graph distance and using an auxiliary Weisfeiler-Lehman kernel to guide the encoding, NASGEM can utilize additional structural information to get more accurate graph representation to improve the search efficiency. GEMNet, a set of networks discovered by NASGEM, consistently outperforms networks crafted by existing search methods in classification tasks, i.e., with 0.4%-3.6% higher accuracy while having 11%- 21% fewer Multiply-Accumulates. We further transfer GEMNet for COCO object detection. In both one-stage and twostage detectors, our GEMNet surpasses its manually-crafted and automatically-searched counterparts.

cs.AI↗

arXiv2020

PENNI: Pruned Kernel Sharing for Efficient CNN Inference

Although state-of-the-art (SOTA) CNNs achieve outstanding performance on various tasks, their high computation demand and massive number of parameters make it difficult to deploy these SOTA CNNs onto resource-constrained devices. Previous works on CNN acceleration utilize low-rank approximation of the original convolution layers to reduce computation cost. However, these methods are very difficult to conduct upon sparse models, which limits execution speedup since redundancies within the CNN model are not fully exploited. We argue that kernel granularity decomposition can be conducted with low-rank assumption while exploiting the redundancy within the remaining compact coefficients. Based on this observation, we propose PENNI, a CNN model compression framework that is able to achieve model compactness and hardware efficiency simultaneously by (1) implementing kernel sharing in convolution layers via a small number of basis kernels and (2) alternately adjusting bases and coefficients with sparse constraints. Experiments show that we can prune 97% parameters and 92% FLOPs on ResNet18 CIFAR10 with no accuracy loss, and achieve 44% reduction in run-time memory consumption and a 53% reduction in inference latency.

cs.CV↗

arXiv2020

Resonant Asymmetric All-Dielectric Metasurface for Boosting Third-Harmonic Generation

Resonant metasurfaces have received extensive attention due to their sharp spectral feature and extraordinary field enhancement. In this work, by breaking the in-plane symmetry of silicon nanopillars, we achieve a sharp Fano resonance. The far-field radiation and near-field distribution of metasurfaces are calculated and analyzed to further uncover the resonant performance of metasurfaces. Moreover, the theoretical derivation and simulation exhibit an inverse quadratic dependence of Q-factors on asymmetry parameters, revealing that the resonance is governed by the symmetry-protected bound states in the continuum. Finally we experimentally demonstrate the sharp resonance, and employ it to effciently boost the third-harmonic generation. This enhancement can be attributed to the strong optical intensity enhancement inside the metasurface.

physics.optics↗

arXiv2020

Opportunities to Search for Extra-Terrestrial Intelligence with the Five-hundred-meter Aperture Spherical radio Telescope

The discovery of ubiquitous habitable extrasolar planets, combined with revolutionary advances in instrumentation and observational capabilities, has ushered in a renaissance in the search for extra-terrestrial intelligence (SETI). Large scale SETI activities are now underway at numerous international facilities. The Five-hundred-meter Aperture Spherical radio Telescope (FAST) is the largest single-aperture radio telescope in the world, well positioned to conduct sensitive searches for radio emission indicative of exo-intelligence. SETI is one of the five key science goals specified in the original FAST project plan. A collaboration with the Breakthrough Listen Initiative has been initiated in 2016 with a joint statement signed both by Dr. Jun Yan, the then director of the National Astronomical Observatories, Chinese Academy of Sciences (NAOC), and Dr. Peter Worden, the Chairman of the Breakthrough Prize Foundation. In this paper, we highlight some of the unique features of FAST that will allow for novel SETI observations. We identify and describe three different signal types indicative of a technological source, namely, narrow-band, wide-band artificially dispersed, and modulated signals. We here propose observations with FAST to achieve sensitivities never before explored.

astro-ph.IM↗

arXiv2019

SwiftNet: Using Graph Propagation as Meta-knowledge to Search Highly Representative Neural Architectures

Designing neural architectures for edge devices is subject to constraints of accuracy, inference latency, and computational cost. Traditionally, researchers manually craft deep neural networks to meet the needs of mobile devices. Neural Architecture Search (NAS) was proposed to automate the neural architecture design without requiring extensive domain expertise and significant manual efforts. Recent works utilized NAS to design mobile models by taking into account hardware constraints and achieved state-of-the-art accuracy with fewer parameters and less computational cost measured in Multiply-accumulates (MACs). To find highly compact neural architectures, existing works relies on predefined cells and directly applying width multiplier, which may potentially limit the model flexibility, reduce the useful feature map information, and cause accuracy drop. To conquer this issue, we propose GRAM(GRAph propagation as Meta-knowledge) that adopts fine-grained (node-wise) search method and accumulates the knowledge learned in updates into a meta-graph. As a result, GRAM can enable more flexible search space and achieve higher search efficiency. Without the constraints of predefined cell or blocks, we propose a new structure-level pruning method to remove redundant operations in neural architectures. SwiftNet, which is a set of models discovered by GRAM, outperforms MobileNet-V2 by 2.15x higher accuracy density and 2.42x faster with similar accuracy. Compared with FBNet, SwiftNet reduces the search cost by 26x and achieves 2.35x higher accuracy density and 1.47x speedup while preserving similar accuracy. SwiftNetcan obtain 63.28% top-1 accuracy on ImageNet-1K with only 53M MACs and 2.07M parameters. The corresponding inference latency is only 19.09 ms on Google Pixel 1.

cs.LG↗

arXiv2018

LEASGD: an Efficient and Privacy-Preserving Decentralized Algorithm for Distributed Learning

Distributed learning systems have enabled training large-scale models over large amount of data in significantly shorter time. In this paper, we focus on decentralized distributed deep learning systems and aim to achieve differential privacy with good convergence rate and low communication cost. To achieve this goal, we propose a new learning algorithm LEASGD (Leader-Follower Elastic Averaging Stochastic Gradient Descent), which is driven by a novel Leader-Follower topology and a differential privacy model.We provide a theoretical analysis of the convergence rate and the trade-off between the performance and privacy in the private setting.The experimental results show that LEASGD outperforms state-of-the-art decentralized learning algorithm DPSGD by achieving steadily lower loss within the same iterations and by reducing the communication cost by 30%. In addition, LEASGD spends less differential privacy budget and has higher final accuracy result than DPSGD under private setting.

cs.LG↗

arXiv2014

The Distribution of Gaps between Summands in Generalized Zeckendorf Decompositions

Zeckendorf proved that any integer can be decomposed uniquely as a sum of non-adjacent Fibonacci numbers, $F_n$. Using continued fractions, Lekkerkerker proved the average number of summands of an $m \in [F_n, F_{n+1})$ is essentially $n/(φ^2 +1)$, with $φ$ the golden ratio. Miller-Wang generalized this by adopting a combinatorial perspective, proving that for any positive linear recurrence the number of summands in decompositions for integers in $[G_n, G_{n+1})$ converges to a Gaussian distribution. We prove the probability of a gap larger than the recurrence length converges to decaying geometrically, and that the distribution of the smaller gaps depends in a computable way on the coefficients of the recurrence. These results hold both for the average over all $m \in [G_n, G_{n+1})$, as well as holding almost surely for the gap measure associated to individual $m$. The techniques can also be used to determine the distribution of the longest gap between summands, which we prove is similar to the distribution of the longest gap between heads in tosses of a biased coin. It is a double exponential strongly concentrated about the mean, and is on the order of $\log n$ with computable constants depending on the recurrence.

math.CO↗

arXiv2012

The Average Gap Distribution for Generalized Zeckendorf Decompositions

An interesting characterization of the Fibonacci numbers is that, if we write them as $F_1 = 1$, $F_2 = 2$, $F_3 = 3$, $F_4 = 5, ...$, then every positive integer can be written uniquely as a sum of non-adjacent Fibonacci numbers. This is now known as Zeckendorf's theorem [21], and similar decompositions exist for many other sequences ${G_{n+1} = c_1 G_{n} + ... + c_L G_{n+1-L}}$ arising from recurrence relations. Much more is known. Using continued fraction approaches, Lekkerkerker [15] proved the average number of summands needed for integers in $[G_n, G_{n+1})$ is on the order of $C_{\rm Lek} n$ for a non-zero constant; this was improved by others to show the number of summands has Gaussian fluctuations about this mean. Kolo$\breve{\rm g}$lu, Kopp, Miller and Wang [17, 18] recently recast the problem combinatorially, reproving and generalizing these results. We use this new perspective to investigate the distribution of gaps between summands. We explore the average behavior over all $m \in [G_n, G_{n+1})$ for special choices of the $c_i$'s. Specifically, we study the case where each $c_i \in {0,1}$ and there is a $g$ such that there are always exactly $g-1$ zeros between two non-zero $c_i$'s; note this includes the Fibonacci, Tribonacci and many other important special cases. We prove there are no gaps of length less than $g$, and the probability of a gap of length $j > g$ decays geometrically, with the decay ratio equal to the largest root of the recurrence relation. These methods are combinatorial and apply to related problems; we end with a discussion of similar results for far-difference (i.e., signed) decompositions.

math.NT↗