arXiv subjects

Jaewon Lee

Publications and source records attributed to Jaewon Lee.

66 records · Page 4Linked to original sources

arXiv2022

Mapping the Minimum Detectable Activities of Gamma-Ray Sources in a 3-D Scene

The ability to formulate maps of minimum detectable activities (MDAs) that describe the sensitivity of an ad hoc measurement that used one or more freely moving radiation detector systems would be significantly beneficial for the conduct and understanding of many radiological search activities. In a real-time scenario with a free-moving detector system, an MDA map can provide useful feedback to the operator about which areas have not been searched as thoroughly as others, thereby allowing the operator to prioritize future actions. Similarly, such a calculation could be used to inform subsequent navigation decisions of autonomous platforms. Here we describe a near real-time MDA mapping approach that can be applied when searching for point sources using detected events in a spectral region of interest while assuming a constant, unknown background rate. We show the application of this MDA mapping method to a real scenario, a survey of the interior of a small building using a handheld detector system. Repeated measurements with no sources and with Cs-137 sources of different strengths yield results consistent with the estimated thresholds and MDA values; namely, that for background-only measurements no sources are seen above threshold anywhere in the scene, while when sources are present they are detected above the thresholds calculated for their locations.

physics.ins-det↗

arXiv2022

Learning Local Implicit Fourier Representation for Image Warping

Image warping aims to reshape images defined on rectangular grids into arbitrary shapes. Recently, implicit neural functions have shown remarkable performances in representing images in a continuous manner. However, a standalone multi-layer perceptron suffers from learning high-frequency Fourier coefficients. In this paper, we propose a local texture estimator for image warping (LTEW) followed by an implicit neural representation to deform images into continuous shapes. Local textures estimated from a deep super-resolution (SR) backbone are multiplied by locally-varying Jacobian matrices of a coordinate transformation to predict Fourier responses of a warped image. Our LTEW-based neural function outperforms existing warping methods for asymmetric-scale SR and homography transform. Furthermore, our algorithm well generalizes arbitrary coordinate transformations, such as homography transform with a large magnification factor and equirectangular projection (ERP) perspective transform, which are not provided in training.

cs.CV↗

arXiv2022

Local Texture Estimator for Implicit Representation Function

Recent works with an implicit neural function shed light on representing images in arbitrary resolution. However, a standalone multi-layer perceptron shows limited performance in learning high-frequency components. In this paper, we propose a Local Texture Estimator (LTE), a dominant-frequency estimator for natural images, enabling an implicit function to capture fine details while reconstructing images in a continuous manner. When jointly trained with a deep super-resolution (SR) architecture, LTE is capable of characterizing image textures in 2D Fourier space. We show that an LTE-based neural function achieves favorable performance against existing deep SR methods within an arbitrary-scale factor. Furthermore, we demonstrate that our implementation takes the shortest running time compared to previous works.

cs.CV↗

arXiv2021

COX: CUDA on X86 by Exposing Warp-Level Functions to CPUs

As CUDA programs become the de facto program among data parallel applications such as high-performance computing or machine learning applications, running CUDA on other platforms has been a compelling option. Although several efforts have attempted to support CUDA on other than NVIDIA GPU devices, due to extra steps in the translation, the support is always behind a few years from supporting CUDA's latest features. The examples are DPC, Hipfy, where CUDA source code have to be translated to their native supporting language and then they are supported. In particular, the new CUDA programming model exposes the warp concept in the programming language, which greatly changes the way the CUDA code should be mapped to CPU programs. In this paper, hierarchical collapsing that \emph{correctly} supports CUDA warp-level functions on CPUs is proposed. Based on hierarchical collapsing, a framework, COX, is developed that allows CUDA programs with the latest features to be executed efficiently on CPU platforms. COX consists of a compiler IR transformation (new LLVM pass) and a runtime system to execute the transformed programs on CPU devices. COX can support the most recent CUDA features, and the application coverage is much higher (90%) than for previous frameworks (68%) with comparable performance. We also show that the warp-level functions in CUDA can be efficiently executed by utilizing CPU SIMD (AVX) instructions.

cs.DC↗

arXiv2021

Point Cloud Augmentation with Weighted Local Transformations

Despite the extensive usage of point clouds in 3D vision, relatively limited data are available for training deep neural networks. Although data augmentation is a standard approach to compensate for the scarcity of data, it has been less explored in the point cloud literature. In this paper, we propose a simple and effective augmentation method called PointWOLF for point cloud augmentation. The proposed method produces smoothly varying non-rigid deformations by locally weighted transformations centered at multiple anchor points. The smooth deformations allow diverse and realistic augmentations. Furthermore, in order to minimize the manual efforts to search the optimal hyperparameters for augmentation, we present AugTune, which generates augmented samples of desired difficulties producing targeted confidence scores. Our experiments show our framework consistently improves the performance for both shape classification and part segmentation tasks. Particularly, with PointNet++, PointWOLF achieves the state-of-the-art 89.7 accuracy on shape classification with the real-world ScanObjectNN dataset.

cs.CV↗

arXiv2021

Supporting CUDA for an extended RISC-V GPU architecture

With the rapid development of scientific computation, more and more researchers and developers are committed to implementing various workloads/operations on different devices. Among all these devices, NVIDIA GPU is the most popular choice due to its comprehensive documentation and excellent development tools. As a result, there are abundant resources for hand-writing high-performance CUDA codes. However, CUDA is mainly supported by only commercial products and there has been no support for open-source H/W platforms. RISC-V is the most popular choice for hardware ISA, thanks to its elegant design and open-source license. In this project, we aim to utilize these existing CUDA codes with RISC-V devices. More specifically, we design and implement a pipeline that can execute CUDA source code on an RISC-V GPU architecture. We have succeeded in executing CUDA kernels with several important features, like multi-thread and atomic instructions, on an RISC-V GPU architecture.

cs.PL↗

arXiv2021

First-Generation Inference Accelerator Deployment at Facebook

In this paper, we provide a deep dive into the deployment of inference accelerators at Facebook. Many of our ML workloads have unique characteristics, such as sparse memory accesses, large model sizes, as well as high compute, memory and network bandwidth requirements. We co-designed a high-performance, energy-efficient inference accelerator platform based on these requirements. We describe the inference accelerator platform ecosystem we developed and deployed at Facebook: both hardware, through Open Compute Platform (OCP), and software framework and tooling, through Pytorch/Caffe2/Glow. A characteristic of this ecosystem from the start is its openness to enable a variety of AI accelerators from different vendors. This platform, with six low-power accelerator cards alongside a single-socket host CPU, allows us to serve models of high complexity that cannot be easily or efficiently run on CPUs. We describe various performance optimizations, at both platform and accelerator level, which enables this platform to serve production traffic at Facebook. We also share deployment challenges, lessons learned during performance optimization, as well as provide guidance for future inference hardware co-design.

cs.AR↗

arXiv2020

The role of charge recombination to spin-triplet excitons in non-fullerene acceptor organic solar cells

The power conversion efficiencies (PCEs) of organic solar cells (OSCs) using non-fullerene acceptors (NFAs) have now reached 18%. However, this is still lower than inorganic solar cells, for which PCEs >20% are commonplace. A key reason is that OSCs still show low open-circuit voltages (Voc) relative to their optical band gaps, attributed to non-radiative recombination. For OSCs to compete with inorganics in efficiency, all non-radiative loss pathways must be identified and where possible, removed. Here, we show that in most NFA OSCs, the majority of charge recombination at open-circuit proceeds via formation of non-emissive NFA triplet excitons (T1); in the benchmark PM6:Y6 blend, this fraction reaches 90%, contributing 60 mV to the reduction of Voc. We develop a new design to prevent recombination via this non-radiative channel through the engineering of significant hybridisation between the NFA T1 and the spin-triplet charge transfer exciton (3CTE). We model that the rate of the back charge transfer from 3CTE to T1 can be reduced by an order of magnitude, allowing re-dissociation of the 3CTE. We then demonstrate NFA systems where T1 formation is suppressed. This work therefore provides a clear design pathway for improved OSC performance to 20% PCE and beyond.

physics.app-ph↗

arXiv2020

Applying GPGPU to Recurrent Neural Network Language Model based Fast Network Search in the Real-Time LVCSR

Recurrent Neural Network Language Models (RNNLMs) have started to be used in various fields of speech recognition due to their outstanding performance. However, the high computational complexity of RNNLMs has been a hurdle in applying the RNNLM to a real-time Large Vocabulary Continuous Speech Recognition (LVCSR). In order to accelerate the speed of RNNLM-based network searches during decoding, we apply the General Purpose Graphic Processing Units (GPGPUs). This paper proposes a novel method of applying GPGPUs to RNNLM-based graph traversals. We have achieved our goal by reducing redundant computations on CPUs and amount of transfer between GPGPUs and CPUs. The proposed approach was evaluated on both WSJ corpus and in-house data. Experiments shows that the proposed approach achieves the real-time speed in various circumstances while maintaining the Word Error Rate (WER) to be relatively 10% lower than that of n-gram models.

cs.CL↗

arXiv2018

Accelerating recurrent neural network language model based online speech recognition system

This paper presents methods to accelerate recurrent neural network based language models (RNNLMs) for online speech recognition systems. Firstly, a lossy compression of the past hidden layer outputs (history vector) with caching is introduced in order to reduce the number of LM queries. Next, RNNLM computations are deployed in a CPU-GPU hybrid manner, which computes each layer of the model on a more advantageous platform. The added overhead by data exchanges between CPU and GPU is compensated through a frame-wise batching strategy. The performance of the proposed methods evaluated on LibriSpeech test sets indicates that the reduction in history vector precision improves the average recognition speed by 1.23 times with minimum degradation in accuracy. On the other hand, the CPU-GPU hybrid parallelization enables RNNLM based real-time recognition with a four times improvement in speed.

cs.CL↗

arXiv2017

Using highly uniform and smooth Selenium colloids as low-loss magnetodielectric building blocks of optical metafluids

We systematically analyzed magnetodielectric resonances of Se colloids for the first time to exploit the possibility for use as building blocks of all-dielectric optical metafluids. By taking synergistic advantages of Se colloids, including (i) high-refractive-index at optical frequencies, (ii) unprecedented structural uniformity, and (iii) versatile access to copious quantities, the Kerker-type directional light scattering resulting from efficient coupling between strong electric and magnetic resonances were observed directly from Se colloidal suspension. Thus, the use of Se colloid as a generic magnetodielectric building block highlights an opportunity for the fluidic low-loss optical antenna, which can be processed via spin-coating and painting.

physics.optics↗

arXiv2014

Inequalities for generalized normalized $δ$-Casorati curvatures of slant submanifolds in quaternionic space forms

In this paper we prove two sharp inequalities involving the normalized scalar curvature and the generalized normalized $δ$-Casorati curvatures for slant submanifolds in quaternionic space forms. We also characterize those submanifolds for which the equality cases hold. These results are a generalization of some recent results concerning the Casorati curvature for a slant submanifold in a quaternionic space form obtained by Slesar et al.: J. Inequal. Appl. 2014, 2014:123.

math.DG↗