arXiv subjects

Sangmin Lee

Publications and source records attributed to Sangmin Lee.

At least 73 records · Page 4Linked to original sources

arXiv2023

Pulsed-mode metalorganic vapor-phase epitaxy of GaN on graphene-coated c-sapphire for freestanding GaN thin films

We report the growth of high-quality GaN epitaxial thin films on graphene-coated c-sapphire substrates using pulsed-mode metalorganic vapor-phase epitaxy, together with the fabrication of freestanding GaN films by simple mechanical exfoliation for transferable light-emitting diodes (LEDs). High-quality GaN films grown on the graphene-coated sapphire substrates were easily lifted off using thermal release tape and transferred onto foreign substrates. Furthermore, we revealed that the pulsed operation of ammonia flow during GaN growth was a critical factor for the fabrication of high-quality freestanding GaN films. These films, exhibiting excellent single crystallinity, were utilized to fabricate transferable GaN LEDs by heteroepitaxially growing InxGa1-xN/GaN multiple quantum wells and a p-GaN layer on the GaN films, showing their potential application in advanced optoelectronic devices.

cond-mat.mtrl-sci↗

arXiv2023

Can We Utilize Pre-trained Language Models within Causal Discovery Algorithms?

Scaling laws have allowed Pre-trained Language Models (PLMs) into the field of causal reasoning. Causal reasoning of PLM relies solely on text-based descriptions, in contrast to causal discovery which aims to determine the causal relationships between variables utilizing data. Recently, there has been current research regarding a method that mimics causal discovery by aggregating the outcomes of repetitive causal reasoning, achieved through specifically designed prompts. It highlights the usefulness of PLMs in discovering cause and effect, which is often limited by a lack of data, especially when dealing with multiple variables. Conversely, the characteristics of PLMs which are that PLMs do not analyze data and they are highly dependent on prompt design leads to a crucial limitation for directly using PLMs in causal discovery. Accordingly, PLM-based causal reasoning deeply depends on the prompt design and carries out the risk of overconfidence and false predictions in determining causal relationships. In this paper, we empirically demonstrate the aforementioned limitations of PLM-based causal reasoning through experiments on physics-inspired synthetic data. Then, we propose a new framework that integrates prior knowledge obtained from PLM with a causal discovery algorithm. This is accomplished by initializing an adjacency matrix for causal discovery and incorporating regularization using prior knowledge. Our proposed framework not only demonstrates improved performance through the integration of PLM and causal discovery but also suggests how to leverage PLM-extracted prior knowledge with existing causal discovery algorithms.

cs.AI↗

arXiv2023

Which is better? Exploring Prompting Strategy For LLM-based Metrics

This paper describes the DSBA submissions to the Prompting Large Language Models as Explainable Metrics shared task, where systems were submitted to two tracks: small and large summarization tracks. With advanced Large Language Models (LLMs) such as GPT-4, evaluating the quality of Natural Language Generation (NLG) has become increasingly paramount. Traditional similarity-based metrics such as BLEU and ROUGE have shown to misalign with human evaluation and are ill-suited for open-ended generation tasks. To address this issue, we explore the potential capability of LLM-based metrics, especially leveraging open-source LLMs. In this study, wide range of prompts and prompting techniques are systematically analyzed with three approaches: prompting strategy, score aggregation, and explainability. Our research focuses on formulating effective prompt templates, determining the granularity of NLG quality scores and assessing the impact of in-context examples on LLM-based evaluation. Furthermore, three aggregation strategies are compared to identify the most reliable method for aggregating NLG quality scores. To examine explainability, we devise a strategy that generates rationales for the scores and analyzes the characteristics of the explanation produced by the open-source LLMs. Extensive experiments provide insights regarding evaluation capabilities of open-source LLMs and suggest effective prompting strategies.

cs.CL↗

arXiv2023

Self-supervised debiasing using low rank regularization

Spurious correlations can cause strong biases in deep neural networks, impairing generalization ability. While most existing debiasing methods require full supervision on either spurious attributes or target labels, training a debiased model from a limited amount of both annotations is still an open question. To address this issue, we investigate an interesting phenomenon using the spectral analysis of latent representations: spuriously correlated attributes make neural networks inductively biased towards encoding lower effective rank representations. We also show that a rank regularization can amplify this bias in a way that encourages highly correlated features. Leveraging these findings, we propose a self-supervised debiasing framework potentially compatible with unlabeled samples. Specifically, we first pretrain a biased encoder in a self-supervised manner with the rank regularization, serving as a semantic bottleneck to enforce the encoder to learn the spuriously correlated attributes. This biased encoder is then used to discover and upweight bias-conflicting samples in a downstream task, serving as a boosting to effectively debias the main model. Remarkably, the proposed debiasing framework significantly improves the generalization performance of self-supervised learning baselines and, in some cases, even outperforms state-of-the-art supervised debiasing approaches.

cs.LG↗

arXiv2023

Symplectic Perturbation Theory in Massive Ambitwistor Space: A Zig-Zag Theory of Massive Spinning Particles

We develop a theory of massive spinning particles interacting with background fields in four spacetime dimensions in which holomorphy and chirality play a central role. Applying a perturbation theory of symplectic forms to the massive twistor space as a Kähler manifold, we find that the spin precession behavior of a massive spinning particle is directly determined from the manner in which self-dual and anti-self-dual field strengths permeate into "complex spacetime." Especially, the particle shows the minimally coupled precession behavior if self-dual field strength continues holomorphically into the complex: the Newman-Janis shift. In general, computing the momentum impulse shows that the parameters that control generic non-holomorphic continuations are directly related to the coupling constants in the massive-massive-massless spinning on-shell amplitude of Arkani-Hamed, Huang, and Huang, and thus they are interpreted as the single-curvature Wilson coefficients given by Levi and Steinhoff, redefined on complex worldlines. Finally, exact expressions for Kerr and $\sqrt{\text{Kerr}}$ actions are bootstrapped in monochromatic self-dual plane-wave backgrounds from symplectivity and a matching between classical scattering and the on-shell amplitude, from which we obtain all-order exact impulses of classical observables.

hep-th↗

arXiv2023

Poincaré generators at second post-Minkowskian order

We verify the global Poincaré invariance of the Hamiltonian mechanics of gravitating binary dynamics at the second post Minkowskian (2PM) order. For spinless point particles, based on the known 2PM Hamiltonian in the center of momentum frame, we compute the general 2PM Hamiltonian valid in an arbitrary reference frame. An off-shell extension of the 1PM Hamiltonian, which contributes at the 2PM order through an iteration process, plays a crucial role. We then construct the 2PM boost generator that uniquely satisfies all the conditions imposed by the Poincaré algebra.

hep-th↗

arXiv2023

Data-driven Nonlinear Parametric Model Order Reduction Framework using Deep Hierarchical Variational Autoencoder

A data-driven parametric model order reduction (MOR) method using a deep artificial neural network is proposed. The present network, which is the least-squares hierarchical variational autoencoder (LSH-VAE), is capable of performing nonlinear MOR for the parametric interpolation of a nonlinear dynamic system with a significant number of degrees of freedom. LSH-VAE exploits two major changes to the existing networks: a hierarchical deep structure and a hybrid weighted, probabilistic loss function. The enhancements result in a significantly improved accuracy and stability compared against the conventional nonlinear MOR methods, autoencoder, and variational autoencoder. Upon LSH-VAE, a parametric MOR framework is presented based on the spherically linear interpolation of the latent manifold. The present framework is validated and evaluated on three nonlinear and multiphysics dynamic systems. First, the present framework is evaluated on the fluid-structure interaction benchmark problem to assess its efficiency and accuracy. Then, a highly nonlinear aeroelastic phenomenon, limit cycle oscillation, is analyzed. Finally, the present framework is applied to a three-dimensional fluid flow to demonstrate its capability of efficiently analyzing a significantly large number of degrees of freedom. The performance of LSH-VAE is emphasized by comparing its results against that of the widely used nonlinear MOR methods, convolutional autoencoder, and $β$-VAE. The present framework exhibits a significantly enhanced accuracy to the conventional methods while still exhibiting a large speed-up factor.

cs.LG↗

arXiv2023

Training Debiased Subnetworks with Contrastive Weight Pruning

Neural networks are often biased to spuriously correlated features that provide misleading statistical evidence that does not generalize. This raises an interesting question: ``Does an optimal unbiased functional subnetwork exist in a severely biased network? If so, how to extract such subnetwork?" While empirical evidence has been accumulated about the existence of such unbiased subnetworks, these observations are mainly based on the guidance of ground-truth unbiased samples. Thus, it is unexplored how to discover the optimal subnetworks with biased training datasets in practice. To address this, here we first present our theoretical insight that alerts potential limitations of existing algorithms in exploring unbiased subnetworks in the presence of strong spurious correlations. We then further elucidate the importance of bias-conflicting samples on structure learning. Motivated by these observations, we propose a Debiased Contrastive Weight Pruning (DCWP) algorithm, which probes unbiased subnetworks without expensive group annotations. Experimental results demonstrate that our approach significantly outperforms state-of-the-art debiasing methods despite its considerable reduction in the number of parameters.

cs.LG↗

arXiv2023

Classical observables from partial wave amplitudes

We study the formalism of Kosower-Maybee-O'Connell (KMOC) to extract classical impulse from quantum amplitude in the context of the partial wave expansion of a 2-to-2 elastic scattering. We take two complementary approaches to establish the connection. The first one takes advantage of Clebsch-Gordan relations for the base amplitudes of the partial wave expansion. The second one is a novel adaptation of the traditional saddle point approximation in the semi-classical limit. In the former, an interference between the S-matrix and its conjugate leads to a large degree of cancellation such that the saddle point approximation to handle a rapidly oscillating integral is no longer needed. As an example with a non-orbital angular momentum, we apply our methods to the charge-monopole scattering problem in the probe limit and reproduce both of the two angles characterizing the classical scattering. A spinor basis for the partial wave expansion, a non-relativistic avatar of the spinor-helicity variables, plays a crucial role throughout our computations.

hep-th↗

arXiv2023

Poincaré invariance of spinning binary dynamics in the post-Minkowskian Hamiltonian approach

We initiate the construction of the global Poincaré algebra generators in the context of the post-Minkowskian Hamiltonian formulation of gravitating binary dynamics in isotropic coordinates that is partly inspired by scattering amplitudes. At the first post-Minkowskian (1PM) order, we write down the Hamiltonian in a form valid in an arbitrary inertial frame. Then we construct the boost generator at the same order which uniquely solves all the equations required by the Poincaré algebra. Our results are linear in Newton's constant but exact in velocities and spins, including all spin multiple moments. We also compute the generators of canonical transformations that proves the equivalence between our new generators and the corresponding generators in the ADM coordinates up to the second post-Newtonian (2PN) order.

gr-qc↗

arXiv2023

Data Topology-Dependent Upper Bounds of Neural Network Widths

This paper investigates the relationship between the universal approximation property of deep neural networks and topological characteristics of datasets. Our primary contribution is to introduce data topology-dependent upper bounds on the network width. Specifically, we first show that a three-layer neural network, applying a ReLU activation function and max pooling, can be designed to approximate an indicator function over a compact set, one that is encompassed by a tight convex polytope. This is then extended to a simplicial complex, deriving width upper bounds based on its topological structure. Further, we calculate upper bounds in relation to the Betti numbers of select topological spaces. Finally, we prove the universal approximation property of three-layer ReLU networks using our topological approach. We also verify that gradient descent converges to the network structure proposed in our study.

cs.LG↗

arXiv2022

Tuning orbital-selective phase transitions in a two-dimensional Hund's correlated system

Hund's rule coupling ($\textit{J}$) has attracted much attention recently for its role in the description of the novel quantum phases of multi orbital materials. Depending on the orbital occupancy, $\textit{J}$ can lead to various intriguing phases. However, experimental confirmation of the orbital occupancy dependency has been difficult as controlling the orbital degrees of freedom normally accompanies chemical inhomogeneities. Here, we demonstrate a method to investigate the role of orbital occupancy in $\textit{J}$ related phenomena without inducing inhomogeneities. By growing SrRuO$_3$ monolayers on various substrates with symmetry-preserving interlayers, we gradually tune the crystal field splitting and thus the orbital degeneracy of the Ru $\textit{t$_2$$_g$}$ orbitals. It effectively varies the orbital occupancies of two-dimensional (2D) ruthenates. Via in-situ angle-resolved photoemission spectroscopy, we observe a progressive metal-insulator transition (MIT). It is found that the MIT occurs with orbital differentiation: concurrent opening of a band insulating gap in the $\textit{d$_x$$_y$}$ band and a Mott gap in the $\textit{d$_x$$_z$$_/$$_y$$_z$}$ bands. Our study provides an effective experimental method for investigation of orbital-selective phenomena in multi-orbital materials.

cond-mat.str-el↗

arXiv2022

Magnitude and Angle Dynamics in Training Single ReLU Neurons

To understand learning the dynamics of deep ReLU networks, we investigate the dynamic system of gradient flow $w(t)$ by decomposing it to magnitude $w(t)$ and angle $ϕ(t):= π- θ(t) $ components. In particular, for multi-layer single ReLU neurons with spherically symmetric data distribution and the square loss function, we provide upper and lower bounds for magnitude and angle components to describe the dynamics of gradient flow. Using the obtained bounds, we conclude that small scale initialization induces slow convergence speed for deep single ReLU neurons. Finally, by exploiting the relation of gradient flow and gradient descent, we extend our results to the gradient descent approach. All theoretical results are verified by experiments.

cs.LG↗

arXiv2022

Support Vectors and Gradient Dynamics of Single-Neuron ReLU Networks

Understanding implicit bias of gradient descent for generalization capability of ReLU networks has been an important research topic in machine learning research. Unfortunately, even for a single ReLU neuron trained with the square loss, it was recently shown impossible to characterize the implicit regularization in terms of a norm of model parameters (Vardi & Shamir, 2021). In order to close the gap toward understanding intriguing generalization behavior of ReLU networks, here we examine the gradient flow dynamics in the parameter space when training single-neuron ReLU networks. Specifically, we discover an implicit bias in terms of support vectors, which plays a key role in why and how ReLU networks generalize well. Moreover, we analyze gradient flows with respect to the magnitude of the norm of initialization, and show that the norm of the learned weight strictly increases through the gradient flow. Lastly, we prove the global convergence of single ReLU neuron for $d = 2$ case.

cs.LG↗

arXiv2022

Quantum trace map for 3-manifolds and a 'length conjecture'

We introduce a quantum trace map for an ideally triangulated hyperbolic knot complement $S^3\backslash \mathcal{K}$. The map assigns a quantum operator to each element of Kauffmann Skein module of the 3-manifold. The quantum operator lives in a module generated by products of quantized edge parameters of the ideal triangulation modulo some equivalence relations determined by gluing equations. Combining the quantum map with a state-integral model of $SL(2,\mathbb{C})$ Chern-Simons theory, one can define perturbative invariants of knot $K$ in the knot complement whose leading part is determined by its complex hyperbolic length. We then conjecture that the perturbative invariants determine an asymptotic expansion of the Jones polynomial for a link composed of $\mathcal{K}$ and $K$. We propose the explicit quantum trace map for figure-eight knot complement and confirm the length conjecture up to the second order in the asymptotic expansion both numerically and analytically.

hep-th↗

arXiv2022

Heteroepitaxial control of Fermi liquid, Hund metal, and Mott insulator phases in the single-atomic-layer limit

Interfaces between dissimilar correlated oxides can offer devices with versatile functionalities. In that respect, manipulating and measuring novel physical properties of oxide heterointerfaces are highly desired. Yet, despite extensive studies, obtaining direct information on their momentum-resolved electronic structure remains a great challenge. This is because most correlated interfacial phenomena appear within a few atomic layers from the interface, thus limiting the application of available experimental probes. Here, we utilize atomic-scale epitaxy and photoemission spectroscopy to demonstrate the interface control of correlated electronic phases in atomic-scale ruthenate--titanate heterostructures. While bulk SrRuO$_3$ is a ferromagnetic metal, the heterointerfaces exclusively realize three distinct correlated phases in the single-atomic-layer limit. Our theory reveals that atomic-scale structural proximity effects lead to the emergence of Fermi liquid, Hund metal, and Mott insulator phases in the quantum-confined SrRuO$_3$. These results highlight the extensive interfacial tunability of electronic phases, hitherto hidden in the atomically thin correlated heterostructure.

cond-mat.str-el↗

arXiv2022

OpenStreetMap-based LiDAR Global Localization in Urban Environment without a Prior LiDAR Map

Using publicly accessible maps, we propose a novel vehicle localization method that can be applied without using prior light detection and ranging (LiDAR) maps. Our method generates OSM descriptors by calculating the distances to buildings from a location in OpenStreetMap at a regular angle, and LiDAR descriptors by calculating the shortest distances to building points from the current location at a regular angle. Comparing the OSM descriptors and LiDAR descriptors yields a highly accurate vehicle localization result. Compared to methods that use prior LiDAR maps, our method presents two main advantages: (1) vehicle localization is not limited to only places with previously acquired LiDAR maps, and (2) our method is comparable to LiDAR map-based methods, and especially outperforms the other methods with respect to the top one candidate at KITTI dataset sequence 00.

cs.RO↗

arXiv2021

Graph Picture of Linear Quantum Networks and Entanglement

The indistinguishability of quantum particles is widely used as a resource for the generation of entanglement. Linear quantum networks (LQNs), in which identical particles linearly evolve to arrive at multimode detectors, exploit the indistinguishability to generate various multipartite entangled states by the proper control of transformation operators. However, it is challenging to devise a suitable LQN that carries a specific entangled state or compute the possible entangled state in a given LQN as the particle and mode number increase. This research presents a mapping process of arbitrary LQNs to graphs, which provides a powerful tool for analyzing and designing LQNs to generate multipartite entanglement. We also introduce the perfect matching diagram (PM diagram), which is a refined directed graph that includes all the essential information on the entanglement generation by an LQN. The PM diagram furnishes rigorous criteria for the entanglement of an LQN and solid guidelines for designing suitable LQNs for the genuine entanglement. Based on the structure of PM diagrams, we compose LQNs for fundamental $N$-partite genuinely entangled states.

quant-ph↗