arXiv subjects

Sungsoo Kim

Publications and source records attributed to Sungsoo Kim.

14 recordsLinked to original sources

arXiv2026

The Diagrammar of Quantum Magnusian

The logarithm of the time-evolution operator has been termed Magnusian, on account of the fact that its expansion describes the Magnus series. The diagrammatic expansion and computation of the classical Magnusian have been completely established in terms of tree graphs and their Hopf algebra. Recent works initiated extensions into quantum field theory, revealing general structures of loop expansions while finding intriguing relations between different diagrams. In this work, we advance the loop expansion further by providing an efficient diagrammatic algorithm to calculate the weight factor of each graph in the quantum Magnusian, known as the Murua coefficient. This is achieved by incorporating two complementary perspectives on the Magnusian at the same time: the color basis and the black-and-white basis. We extract the Murua coefficients from the Magnus series by utilizing these two bases while implementing an exponentiated Wick contraction. In turn, we identify the loop-level extension of Murua's recursive formula. Eventually, we establish a set of edge-contraction rules which facilitate a direct recursive computation of the Murua coefficients at the purely diagrammatic level, without referencing or directly manipulating the underlying Magnus expansion. This shows that the matrix elements of the quantum Magnusian can be computed from graph manipulations alone.

hep-th↗

arXiv2026

On the integrability of root-Kerr probe dynamics

In the background of a Kerr-Newman black hole, the motion of a scalar particle is integrable by virtue of an extra conserved charge known as Carter charge. When the particle is endowed with spin, it is known that another conserved charge, the R\"udiger charge, maintains the integrability at least at low orders in the spin magnitude. We explore the extent of this integrability in a simpler model where both the source and the probe are root-Kerr particles, the non-gravitating limit of the Kerr-Newman black hole. At the leading order in the probe charge, the integrability holds to all orders in the spin magnitude if the interaction vertices of the probe are dictated by the Newman-Janis shift. At the second order in the probe charge, the integrability can be extended to the spin-squared order but begins to fail at the spin-cubic order. An argument based on asymptotic conservation suggests that it is impossible to restore the conservation at the spin-cubic order by a further deformation of the probe action. We compare our results with related observations for Kerr black holes with gravitational interactions.

hep-th↗

arXiv2025

Classical eikonal in relativistic scattering

The classical eikonal is defined to be the generator of all scattering observables in a scattering problem in classical mechanics. It was originally introduced as the log of the quantum S-matrix in the classical limit. But its classical nature calls for a definition and computational methods independent of quantum mechanics. In this paper, we formulate a classical interaction picture which serves as the foundation of the classical eikonal. Our emphasis is on generality. In perturbation theories, both Hamiltonian deformation and symplectic deformation are considered. Particles and fields are treated on a similar footing. The causality prescription of the propagator is essentially the same for non-relativistic and relativistic kinematics. For a probe particle in electromagnetic or gravitational background, we present all order formulas for the perturbative eikonal. In the electromagnetic setting, we also illustrate how the eikonal encodes the information on radiation of external fields.

hep-th↗

arXiv2024

Towards Maximum Likelihood Training for Transducer-based Streaming Speech Recognition

Transducer neural networks have emerged as the mainstream approach for streaming automatic speech recognition (ASR), offering state-of-the-art performance in balancing accuracy and latency. In the conventional framework, streaming transducer models are trained to maximize the likelihood function based on non-streaming recursion rules. However, this approach leads to a mismatch between training and inference, resulting in the issue of deformed likelihood and consequently suboptimal ASR accuracy. We introduce a mathematical quantification of the gap between the actual likelihood and the deformed likelihood, namely forward variable causal compensation (FoCC). We also present its estimator, FoCCE, as a solution to estimate the exact likelihood. Through experiments on the LibriSpeech dataset, we show that FoCCE training improves the accuracy of the streaming transducers.

eess.AS↗

arXiv2024

Classical eikonal from Magnus expansion

In a classical scattering problem, the classical eikonal is defined as the generator of the canonical transformation that maps in-states to out-states. It can be regarded as the classical limit of the log of the quantum S-matrix. In a classical analog of the Born approximation in quantum mechanics, the classical eikonal admits an expansion in oriented tree graphs, where oriented edges denote retarded/advanced worldline propagators. The Magnus expansion, which takes the log of a time-ordered exponential integral, offers an efficient method to compute the coefficients of the tree graphs to all orders. We exploit a Hopf algebra structure behind the Magnus expansion to develop a fast algorithm which can compute the tree coefficients up to the 12th order (over half a million trees) in less than an hour. In a relativistic setting, our methods can be applied to the post-Minkowskian (PM) expansion for gravitational binaries in the worldline formalism. We demonstrate the methods by computing the 3PM eikonal and find agreement with previous results based on amplitude methods. Importantly, the Magnus expansion yields a finite eikonal, while the na\"ive eikonal based on the time-symmetric propagator is infrared-divergent from 3PM on.

hep-th↗

arXiv2024

TiVaT: A Transformer with a Single Unified Mechanism for Capturing Asynchronous Dependencies in Multivariate Time Series Forecasting

Multivariate time series (MTS) forecasting is vital across various domains but remains challenging due to the need to simultaneously model temporal and inter-variate dependencies. Existing channel-dependent models, where Transformer-based models dominate, process these dependencies separately, limiting their capacity to capture complex interactions such as lead-lag dynamics. To address this issue, we propose TiVaT (Time-variate Transformer), a novel architecture incorporating a single unified module, a Joint-Axis (JA) attention module, that concurrently processes temporal and variate modeling. The JA attention module dynamically selects relevant features to particularly capture asynchronous interactions. In addition, we introduce distance-aware time-variate sampling in the JA attention, a novel mechanism that extracts significant patterns through a learned 2D embedding space while reducing noise. Extensive experiments demonstrate TiVaT's overall performance across diverse datasets, particularly excelling in scenarios with intricate asynchronous dependencies.

cs.LG↗

arXiv2022

Two-Pass End-to-End ASR Model Compression

Speech recognition on smart devices is challenging owing to the small memory footprint. Hence small size ASR models are desirable. With the use of popular transducer-based models, it has become possible to practically deploy streaming speech recognition models on small devices [1]. Recently, the two-pass model [2] combining RNN-T and LAS modules has shown exceptional performance for streaming on-device speech recognition. In this work, we propose a simple and effective approach to reduce the size of the two-pass model for memory-constrained devices. We employ a popular knowledge distillation approach in three stages using the Teacher-Student training technique. In the first stage, we use a trained RNN-T model as a teacher model and perform knowledge distillation to train the student RNN-T model. The second stage uses the shared encoder and trains a LAS rescorer for student model using the trained RNN-T+LAS teacher model. Finally, we perform deep-finetuning for the student model with a shared RNN-T encoder, RNN-T decoder, and LAS rescorer. Our experimental results on standard LibriSpeech dataset show that our system can achieve a high compression rate of 55% without significant degradation in the WER compared to the two-pass teacher model.

eess.AS↗

arXiv2021

A review of on-device fully neural end-to-end automatic speech recognition algorithms

In this paper, we review various end-to-end automatic speech recognition algorithms and their optimization techniques for on-device applications. Conventional speech recognition systems comprise a large number of discrete components such as an acoustic model, a language model, a pronunciation model, a text-normalizer, an inverse-text normalizer, a decoder based on a Weighted Finite State Transducer (WFST), and so on. To obtain sufficiently high speech recognition accuracy with such conventional speech recognition systems, a very large language model (up to 100 GB) is usually needed. Hence, the corresponding WFST size becomes enormous, which prohibits their on-device implementation. Recently, fully neural network end-to-end speech recognition algorithms have been proposed. Examples include speech recognition systems based on Connectionist Temporal Classification (CTC), Recurrent Neural Network Transducer (RNN-T), Attention-based Encoder-Decoder models (AED), Monotonic Chunk-wise Attention (MoChA), transformer-based speech recognition systems, and so on. These fully neural network-based systems require much smaller memory footprints compared to conventional algorithms, therefore their on-device implementation has become feasible. In this paper, we review such end-to-end speech recognition models. We extensively discuss their structures, performance, and advantages compared to conventional algorithms.

cs.LG↗

arXiv2021

Sequential Routing Framework: Fully Capsule Network-based Speech Recognition

Capsule networks (CapsNets) have recently gotten attention as a novel neural architecture. This paper presents the sequential routing framework which we believe is the first method to adapt a CapsNet-only structure to sequence-to-sequence recognition. Input sequences are capsulized then sliced by a window size. Each slice is classified to a label at the corresponding time through iterative routing mechanisms. Afterwards, losses are computed by connectionist temporal classification (CTC). During routing, the required number of parameters can be controlled by the window size regardless of the length of sequences by sharing learnable weights across the slices. We additionally propose a sequential dynamic routing algorithm to replace traditional dynamic routing. The proposed technique can minimize decoding speed degradation caused by the routing iterations since it can operate in a non-iterative manner without dropping accuracy. The method achieves a 1.1% lower word error rate at 16.9% on the Wall Street Journal corpus compared to bidirectional long short-term memory-based CTC networks. On the TIMIT corpus, it attains a 0.7% lower phone error rate at 17.5% compared to convolutional neural network-based CTC networks (Zhang et al., 2016).

eess.AS↗

arXiv2020

Attention based on-device streaming speech recognition with large speech corpus

In this paper, we present a new on-device automatic speech recognition (ASR) system based on monotonic chunk-wise attention (MoChA) models trained with large (> 10K hours) corpus. We attained around 90% of a word recognition rate for general domain mainly by using joint training of connectionist temporal classifier (CTC) and cross entropy (CE) losses, minimum word error rate (MWER) training, layer-wise pre-training and data augmentation methods. In addition, we compressed our models by more than 3.4 times smaller using an iterative hyper low-rank approximation (LRA) method while minimizing the degradation in recognition accuracy. The memory footprint was further reduced with 8-bit quantization to bring down the final model size to lower than 39 MB. For on-demand adaptation, we fused the MoChA models with statistical n-gram models, and we could achieve a relatively 36% improvement on average in word error rate (WER) for target domains including the general domain.

eess.AS↗

arXiv2019

end-to-end training of a large vocabulary end-to-end speech recognition system

In this paper, we present an end-to-end training framework for building state-of-the-art end-to-end speech recognition systems. Our training system utilizes a cluster of Central Processing Units(CPUs) and Graphics Processing Units (GPUs). The entire data reading, large scale data augmentation, neural network parameter updates are all performed "on-the-fly". We use vocal tract length perturbation [1] and an acoustic simulator [2] for data augmentation. The processed features and labels are sent to the GPU cluster. The Horovod allreduce approach is employed to train neural network parameters. We evaluated the effectiveness of our system on the standard Librispeech corpus [3] and the 10,000-hr anonymized Bixby English dataset. Our end-to-end speech recognition system built using this training infrastructure showed a 2.44 % WER on test-clean of the LibriSpeech test set after applying shallow fusion with a Transformer language model (LM). For the proprietary English Bixby open domain test set, we obtained a WER of 7.92 % using a Bidirectional Full Attention (BFA) end-to-end model after applying shallow fusion with an RNN-LM. When the monotonic chunckwise attention (MoCha) based approach is employed for streaming speech recognition, we obtained a WER of 9.95 % on the same Bixby open domain test set.

eess.AS↗

arXiv2018

Adversarial Video Compression Guided by Soft Edge Detection

We propose a video compression framework using conditional Generative Adversarial Networks (GANs). We rely on two encoders: one that deploys a standard video codec and another which generates low-level maps via a pipeline of down-sampling, a newly devised soft edge detector, and a novel lossless compression scheme. For decoding, we use a standard video decoder as well as a neural network based one, which is trained using a conditional GAN. Recent "deep" approaches to video compression require multiple videos to pre-train generative networks to conduct interpolation. In contrast to this prior work, our scheme trains a generative decoder on pairs of a very limited number of key frames taken from a single video and corresponding low-level maps. The trained decoder produces reconstructed frames relying on a guidance of low-level maps, without any interpolation. Experiments on a diverse set of 131 videos demonstrate that our proposed GAN-based compression engine achieves much higher quality reconstructions at very low bitrates than prevailing standard codecs such as H.264 or HEVC.

eess.IV↗

arXiv2014

Formation of Warped Disks by Galactic Fly-by Encounters. I. Stellar Disks

Warped disks are almost ubiquitous among spiral galaxies. Here we revisit and test the `fly-by scenario' of warp formation, in which impulsive encounters between galaxies are responsible for warped disks. Based on N-body simulations, we investigate the morphological and kinematical evolution of the stellar component of disks when galaxies undergo fly-by interactions with adjacent dark matter halos. We find that the so-called `S'-shaped warps can be excited by fly-bys and sustained for even up to a few billion years, and that this scenario provides a cohesive explanation for several key observations. We show that disk warp properties are governed primarily by the following three parameters; (1) the impact parameter, i.e., the minimum distance between two halos, (2) the mass ratio between two halos, and (3) the incident angle of the fly-by perturber. The warp angle is tied up with all three parameters, yet the warp lifetime is particularly sensitive to the incident angle of the perturber. Interestingly, the modeled S-shaped warps are often non-symmetric depending on the incident angle. We speculate that the puzzling U- and L-shaped warps are geometrically superimposed S-types produced by successive fly-bys with different incident angles, including multiple interactions with a satellite on a highly elongated orbit.

astro-ph.GA↗

arXiv2012

Initial Size Distribution of the Galactic Globular Cluster System

Despite the importance of their size evolution in understanding the dynamical evolution of globular clusters (GCs) of the Milky Way, studies are rare that focus specifically on this issue. Based on the advanced, realistic Fokker-Planck (FP) approach, we predict theoretically the initial size distribution (SD) of the Galactic GCs along with their initial mass function and radial distribution. Over one thousand FP calculations in a wide parameter space have pinpointed the best-fit initial conditions for the SD, mass function, and radial distribution. Our best-fit model shows that the initial SD of the Galactic GCs is of larger dispersion than today's SD, and that typical projected half-light radius of the initial GCs is ~4.6 pc, which is 1.8 times larger than that of the present-day GCs (~2.5 pc). Their large size signifies greater susceptibility to the Galactic tides: the total mass of destroyed GCs reaches 3-5x10^8 M_sun$, several times larger than the previous estimates. Our result challenges a recent view that the Milky Way GCs were born compact on the sub-pc scale, and rather implies that (1) the initial GCs are generally larger than the typical size of the present-day GCs, (2) the initially large GCs mostly shrink and/or disrupt as a result of the galactic tides, and (3) the initially small GCs expand by two-body relaxation, and later shrink by the galactic tides.

astro-ph.GA↗