arXiv subjects

Jiho Lee

Publications and source records attributed to Jiho Lee.

26 records · Page 2Linked to original sources

arXiv2022

Revisiting Image Pyramid Structure for High Resolution Salient Object Detection

Salient object detection (SOD) has been in the spotlight recently, yet has been studied less for high-resolution (HR) images. Unfortunately, HR images and their pixel-level annotations are certainly more labor-intensive and time-consuming compared to low-resolution (LR) images and annotations. Therefore, we propose an image pyramid-based SOD framework, Inverse Saliency Pyramid Reconstruction Network (InSPyReNet), for HR prediction without any of HR datasets. We design InSPyReNet to produce a strict image pyramid structure of saliency map, which enables to ensemble multiple results with pyramid-based image blending. For HR prediction, we design a pyramid blending method which synthesizes two different image pyramids from a pair of LR and HR scale from the same image to overcome effective receptive field (ERF) discrepancy. Our extensive evaluations on public LR and HR SOD benchmarks demonstrate that InSPyReNet surpasses the State-of-the-Art (SotA) methods on various SOD metrics and boundary accuracy.

cs.CV↗

arXiv2022

Small solutions of the Einstein-Boltzmann-scalar field system with Bianchi symmetry

We show that small homogeneous solutions to the Einstein-Boltzmann-scalar field system exist globally towards the future and tend to the de Sitter solution in a suitable sense. More specifically, we assume that the spacetime is of Bianchi type I--VIII, that the matter is described by Israel particles and that there exists a scalar field with a potential which has a positive lower bound. This represents a generalization of the work [19], where a cosmological constant was considered, and a generalization of [16], where a spatially flat FLRW spacetime was considered. We obtain the global existence and asymptotic behavior of classical solutions to the Einstein-Boltzmann-scalar field system for small initial data.

gr-qc↗

arXiv2022

Better Quality Estimation for Low Resource Corpus Mining

Quality Estimation (QE) models have the potential to change how we evaluate and maybe even train machine translation models. However, these models still lack the robustness to achieve general adoption. We show that State-of-the-art QE models, when tested in a Parallel Corpus Mining (PCM) setting, perform unexpectedly bad due to a lack of robustness to out-of-domain examples. We propose a combination of multitask training, data augmentation and contrastive learning to achieve better and more robust QE performance. We show that our method improves QE performance significantly in the MLQE challenge and the robustness of QE models when tested in the Parallel Corpus Mining setup. We increase the accuracy in PCM by more than 0.80, making it on par with state-of-the-art PCM methods that use millions of sentence pairs to train their models. In comparison, we use a thousand times less data, 7K parallel sentences in total, and propose a novel low resource PCM method.

cs.CL↗

arXiv2021

$α$-Stable convergence of heavy-tailed infinitely-wide neural networks

We consider infinitely-wide multi-layer perceptrons (MLPs) which are limits of standard deep feed-forward neural networks. We assume that, for each layer, the weights of an MLP are initialized with i.i.d. samples from either a light-tailed (finite variance) or heavy-tailed distribution in the domain of attraction of a symmetric $α$-stable distribution, where $α\in(0,2]$ may depend on the layer. For the bias terms of the layer, we assume i.i.d. initializations with a symmetric $α$-stable distribution having the same $α$ parameter of that layer. We then extend a recent result of Favaro, Fortini, and Peluchetti (2020), to show that the vector of pre-activation values at all nodes of a given hidden layer converges in the limit, under a suitable scaling, to a vector of i.i.d. random variables with symmetric $α$-stable distributions.

stat.ML↗

arXiv2021

A de Finetti-type representation of joint hierarchically exchangeable arrays on directed acyclic graphs

We define joint exchangeability on arrays indexed by a vector of natural numbers with coordinates being the vertices of directed acyclic graphs (DAGs) using local isomorphisms. The notion provides a new version of exchangeability, which is a joint version of hierarchical exchangeability defined in Jung, L., Staton, Yang (2020). We also prove the existence of a generic representation by independent uniform random variables.

math.PR↗

arXiv2020

Deep learning-based citation recommendation system for patents

In this study, we address the challenges in developing a deep learning-based automatic patent citation recommendation system. Although deep learning-based recommendation systems have exhibited outstanding performance in various domains (such as movies, products, and paper citations), their validity in patent citations has not been investigated, owing to the lack of a freely available high-quality dataset and relevant benchmark model. To solve these problems, we present a novel dataset called PatentNet that includes textual information and metadata for approximately 110,000 patents from the Google Big Query service. Further, we propose strong benchmark models considering the similarity of textual information and metadata (such as cooperative patent classification code). Compared with existing recommendation methods, the proposed benchmark method achieved a mean reciprocal rank of 0.2377 on the test set, whereas the existing state-of-the-art recommendation method achieved 0.2073.

cs.IR↗

arXiv2020

A Generalization of Hierarchical Exchangeability on Trees to Directed Acyclic Graphs

Motivated by the problem of designing inference-friendly Bayesian nonparametric models in probabilistic programming languages, we introduce a general class of partially exchangeable random arrays which generalizes the notion of hierarchical exchangeability introduced in Austin and Panchenko (2014). We say that our partially exchangeable arrays are DAG-exchangeable since their partially exchangeable structure is governed by a collection of Directed Acyclic Graphs. More specifically, such a random array is indexed by $\mathbb{N}^{|V|}$ for some DAG $G=(V,E)$, and its exchangeability structure is governed by the edge set $E$. We prove a representation theorem for such arrays which generalizes the Aldous-Hoover and Austin-Panchenko representation theorems.

math.PR↗

arXiv2019

Grid dissections of tangential quadrilaterals

For any integer $n\ge 2$, a square can be partitioned into $n^2$ smaller squares via a checkerboard-type dissection. Does there such a class-preserving grid dissection exist for some other types of quadrilaterals? For instance, is it true that a tangential quadrilateral can be partitioned into $n^2$ smaller tangential quadrilaterals using an $n\times n$ grid dissection? We prove that the answer is affirmative for every integer $n\ge 2$.

math.MG↗