arXiv subjects

Mingu Kang

Publications and source records attributed to Mingu Kang.

At least 37 records · Page 2Linked to original sources

arXiv2025

Time-domain decoding of unconventional charge order mechanisms in nonmagnetic and magnetic kagome metals

In kagome lattice materials, quantum interplay between charge, spin, orbital, and lattice degrees of freedom gives rise to a remarkably rich set of emergent phenomena, ranging from unconventional charge order and superconductivity to topological magnetism. While the exact nature of these exotic orders is often challenging to comprehend in static experiments, time-resolved techniques can offer critical insights by disentangling coupled degrees of freedom on the time-axis. In this work, we demonstrate that the nature of charge orders in two representative kagome metals - nonmagnetic ScV6Sn6 and magnetic FeGe - which has been highly controversial in static studies, can be directly deciphered in the time-domain through their fundamentally distinct order parameter dynamics measured via time-resolved X-ray scattering at an X-ray free electron laser. In nonmagnetic ScV6Sn6, the dynamics are characterized by ultrafast melting and coherent amplitudon oscillations, typical of a phonon-coupled charge order. In stark contrast, magnetic FeGe exhibits resilient metastable charge order dynamics, hitherto unobserved in any other charge-ordered system - this unique time-domain behavior directly signifies an unconventional magnetism-interlocked charge order state realized in this kagome magnet. Our results not only provide a model case where unconventional nature of electronic order, hidden in equilibrium, is directly unraveled in the time-domain, but also pave the way for future out-of-equilibrium engineering of novel quantum orders in kagome lattice platforms.

cond-mat.str-el↗

arXiv2025

Hybrid SLC-MLC RRAM Mixed-Signal Processing-in-Memory Architecture for Transformer Acceleration via Gradient Redistribution

Transformers, while revolutionary, face challenges due to their demanding computational cost and large data movement. To address this, we propose HyFlexPIM, a novel mixed-signal processing-in-memory (PIM) accelerator for inference that flexibly utilizes both single-level cell (SLC) and multi-level cell (MLC) RRAM technologies to trade-off accuracy and efficiency. HyFlexPIM achieves efficient dual-mode operation by utilizing digital PIM for high-precision and write-intensive operations while analog PIM for high parallel and low-precision computations. The analog PIM further distributes tasks between SLC and MLC PIM operations, where a single analog PIM module can be reconfigured to switch between two operations (SLC/MLC) with minimal overhead (<1% for area & energy). Critical weights are allocated to SLC RRAM for high accuracy, while less critical weights are assigned to MLC RRAM to maximize capacity, power, and latency efficiency. However, despite employing such a hybrid mechanism, brute-force mapping on hardware fails to deliver significant benefits due to the limited proportion of weights accelerated by the MLC and the noticeable degradation in accuracy. To maximize the potential of our hybrid hardware architecture, we propose an algorithm co-optimization technique, called gradient redistribution, which uses Singular Value Decomposition (SVD) to decompose and truncate matrices based on their importance, then fine-tune them to concentrate significance into a small subset of weights. By doing so, only 5-10% of the weights have dominantly large gradients, making it favorable for HyFlexPIM by minimizing the use of expensive SLC RRAM while maximizing the efficient MLC RRAM. Our evaluation shows that HyFlexPIM significantly enhances computational throughput and energy efficiency, achieving maximum 1.86X and 1.45X higher than state-of-the-art methods.

cs.AR↗

arXiv2025

Phonon fluctuation diagnostics: Origin of charge order in AV$_3$Sb$_5$ kagome metals

The microsopic origin of the charge-density wave (CDW) in AV$_3$Sb$_5$ (A = K, Rb, Cs) kagome metals remains a longstanding question, often revolving around electron-phonon coupling and purely electronic mechanisms involving Van Hove scenarios, nesting, and sublattice interference. To reveal the processes driving the CDW transition, we combine ab-initio calculations analysis of the phonon self-energy and angle-resolved photoemission spectroscopy (ARPES). Our momentum-resolved study, supported by ARPES data, reveals that lattice instabilities in the V-135 family of kagome metals appear to also be driven by electronic states far from high-symmetry points, where these states exhibit the strongest coupling with the phonon modes responsible for the CDW distortion. Footing on an interpretation scheme based on phonon fluctuation diagnostics, our work challenges and revises theories that so far have exclusively attributed CDW formation to nesting effects close to the Fermi level.

cond-mat.str-el↗

arXiv2024

Measurements of the quantum geometric tensor in solids

Understanding the geometric properties of quantum states and their implications in fundamental physical phenomena is at the core of modern physics. The Quantum Geometric Tensor (QGT) is a central physical object in this regard, encoding complete information about the geometry of the quantum state. The imaginary part of the QGT is the well-known Berry curvature, which plays a fundamental role in the topological magnetoelectric and optoelectronic phenomena. The real part of the QGT is the quantum metric, whose importance has come to prominence very recently, giving rise to a new set of quantum geometric phenomena, such as anomalous Landau levels, flat band superfluidity, excitonic Lamb shifts, and nonlinear Hall effect. Despite the central importance of the QGT, its experimental measurements have been restricted only to artificial two-level systems. In this work, we develop a framework to measure the QGT (both quantum metric and Berry curvature) in crystalline solids using polarization-, spin-, and angle-resolved photoemission spectroscopy. Using this framework, we demonstrate the effective reconstruction of the QGT in solids in the archetype kagome metal CoSn, which hosts topological flat bands. The key idea is to introduce another geometrical tensor, the quasi-QGT, whose components, the band Drude weight and orbital angular momentum, are experimentally accessible and can be used for extracting the QGT. Establishing such a momentum- and energy-resolved spectroscopic probe of the QGT is poised to significantly advance our understanding of quantum geometric responses in a wide range of crystalline systems.

cond-mat.mtrl-sci↗

arXiv2024

SoftLMs: Efficient Adaptive Low-Rank Approximation of Language Models using Soft-Thresholding Mechanism

Extensive efforts have been made to boost the performance in the domain of language models by introducing various attention-based transformers. However, the inclusion of linear layers with large dimensions contributes to significant computational and memory overheads. The escalating computational demands of these models necessitate the development of various compression techniques to ensure their deployment on devices, particularly in resource-constrained environments. In this paper, we propose a novel compression methodology that dynamically determines the rank of each layer using a soft thresholding mechanism, which clips the singular values with a small magnitude in a differentiable form. This approach automates the decision-making process to identify the optimal degree of compression for each layer. We have successfully applied the proposed technique to attention-based architectures, including BERT for discriminative tasks and GPT2 and TinyLlama for generative tasks. Additionally, we have validated our method on Mamba, a recently proposed state-space model. Our experiments demonstrate that the proposed technique achieves a speed-up of 1.33X to 1.72X in the encoder/ decoder with a 50% reduction in total parameters.

cs.LG↗

arXiv2024

MaD-Scientist: AI-based Scientist solving Convection-Diffusion-Reaction Equations Using Massive PINN-Based Prior Data

Large language models (LLMs), like ChatGPT, have shown that even trained with noisy prior data, they can generalize effectively to new tasks through in-context learning (ICL) and pre-training techniques. Motivated by this, we explore whether a similar approach can be applied to scientific foundation models (SFMs). Our methodology is structured as follows: (i) we collect low-cost physics-informed neural network (PINN)-based approximated prior data in the form of solutions to partial differential equations (PDEs) constructed through an arbitrary linear combination of mathematical dictionaries; (ii) we utilize Transformer architectures with self and cross-attention mechanisms to predict PDE solutions without knowledge of the governing equations in a zero-shot setting; (iii) we provide experimental evidence on the one-dimensional convection-diffusion-reaction equation, which demonstrate that pre-training remains robust even with approximated prior data, with only marginal impacts on test accuracy. Notably, this finding opens the path to pre-training SFMs with realistic, low-cost data instead of (or in conjunction with) numerical high-cost data. These results support the conjecture that SFMs can improve in a manner similar to LLMs, where fully cleaning the vast set of sentences crawled from the Internet is nearly impossible.

cs.LG↗

arXiv2024

An Analog and Digital Hybrid Attention Accelerator for Transformers with Charge-based In-memory Computing

The attention mechanism is a key computing kernel of Transformers, calculating pairwise correlations across the entire input sequence. The computing complexity and frequent memory access in computing self-attention put a huge burden on the system especially when the sequence length increases. This paper presents an analog and digital hybrid processor to accelerate the attention mechanism for transformers in 65nm CMOS technology. We propose an analog computing-in-memory (CIM) core, which prunes ~75% of low-score tokens on average during runtime at ultra-low power and delay. Additionally, a digital processor performs precise computations only for ~25% unpruned tokens selected by the analog CIM core, preventing accuracy degradation. Measured results show peak energy efficiency of 14.8 and 1.65 TOPS/W, and peak area efficiency of 976.6 and 79.4 GOPS/mm$^\mathrm{2}$ in the analog core and the system-on-chip (SoC), respectively.

cs.AR↗

arXiv2024

FSL-HDnn: A 5.7 TOPS/W End-to-end Few-shot Learning Classifier Accelerator with Feature Extraction and Hyperdimensional Computing

This paper introduces FSL-HDnn, an energy-efficient accelerator that implements the end-to-end pipeline of feature extraction, classification, and on-chip few-shot learning (FSL) through gradient-free learning techniques in a 40 nm CMOS process. At its core, FSL-HDnn integrates two low-power modules: Weight clustering feature extractor and Hyperdimensional Computing (HDC). Feature extractor utilizes advanced weight clustering and pattern reuse strategies for optimized CNN-based feature extraction. Meanwhile, HDC emerges as a novel approach for lightweight FSL classifier, employing hyperdimensional vectors to improve training accuracy significantly compared to traditional distance-based approaches. This dual-module synergy not only simplifies the learning process by eliminating the need for complex gradients but also dramatically enhances energy efficiency and performance. Specifically, FSL-HDnn achieves an Intensity unprecedented energy efficiency of 5.7 TOPS/W for feature 1 extraction and 0.78 TOPS/W for classification and learning Training Intensity phases, achieving improvements of 2.6X and 6.6X, respectively, Storage over current state-of-the-art CNN and FSL processors.

cs.AR↗

arXiv2024

Collective nature of orbital excitations in layered cuprates in the absence of apical oxygens

We have investigated the 3d orbital excitations in CaCuO2 (CCO), Nd2CuO4 (NCO), and La2CuO4 (LCO) using high-resolution resonant inelastic x-ray scattering. In LCO they behave as well-localized excitations, similarly to several other cuprates. On the contrary, in CCO and NCO the dxy orbital clearly disperse, pointing to a collective character of this excitation (orbiton) in compounds without apical oxygen. We ascribe the origin of the dispersion as stemming from a substantial next-nearest-neighbor (NNN) orbital superexchange. Such an exchange leads to the liberation of orbiton from its coupling to magnons, which is associated with the orbiton hopping between nearest neighbor copper sites. We show that the exceptionally large NNN orbital superexchange can be traced back to the absence of apical oxygens suppressing the charge transfer energy.

cond-mat.str-el↗

arXiv2024

Nodal fermions in a strongly spin-orbit coupled frustrated pyrochlore superconductor

The pyrochlore lattice, a three-dimensional network of corner-sharing tetrahedra, is a promising material playground for correlated topological phases arising from the interplay between spin-orbit coupling (SOC) and electron-electron interactions. Due to its geometrically frustrated lattice structure, exotic correlated states on the pyrochlore lattice have been extensively studied using various spin Hamiltonians in the localized limit. On the other hand, the topological properties of the electronic structure in the pyrochlore lattice have rarely been explored, due to the scarcity of pyrochlore materials in the itinerant paramagnetic limit. Here, we explore the topological electronic band structure of pyrochlore superconductor RbBi$_{2}$ using angle-resolved photoemission spectroscopy. Thanks to the strong SOC of the Bi pyrochlore network, we experimentally confirm the existence of three-dimensional (3D) massless Dirac fermions enforced by nonsymmorphic symmetry, as well as a 3D quadratic band crossing protected by cubic crystalline symmetry. Furthermore, we identify an additional 3D linear Dirac dispersion associated with band inversion protected by threefold rotation symmetry. These observations reveal the rich non-trivial band topology of itinerant pyrochlore lattice systems in the strong SOC regime. Through manipulation of electron correlations and SOC of the frustrated pyrochlore lattices, this material platform is a natural host for exotic phases of matter, including the fractionalized quantum spin Hall effect in the topological Mott insulator phase, as well as axion electrodynamics in the axion insulator phase.

cond-mat.mtrl-sci↗

arXiv2024

SPADE: Sparse Pillar-based 3D Object Detection Accelerator for Autonomous Driving

3D object detection using point cloud (PC) data is essential for perception pipelines of autonomous driving, where efficient encoding is key to meeting stringent resource and latency requirements. PointPillars, a widely adopted bird's-eye view (BEV) encoding, aggregates 3D point cloud data into 2D pillars for fast and accurate 3D object detection. However, the state-of-the-art methods employing PointPillars overlook the inherent sparsity of pillar encoding where only a valid pillar is encoded with a vector of channel elements, missing opportunities for significant computational reduction. Meanwhile, current sparse convolution accelerators are designed to handle only element-wise activation sparsity and do not effectively address the vector sparsity imposed by pillar encoding. In this paper, we propose SPADE, an algorithm-hardware co-design strategy to maximize vector sparsity in pillar-based 3D object detection and accelerate vector-sparse convolution commensurate with the improved sparsity. SPADE consists of three components: (1) a dynamic vector pruning algorithm balancing accuracy and computation savings from vector sparsity, (2) a sparse coordinate management hardware transforming 2D systolic array into a vector-sparse convolution accelerator, and (3) sparsity-aware dataflow optimization tailoring sparse convolution schedules for hardware efficiency. Taped-out with a commercial technology, SPADE saves the amount of computation by 36.3--89.2\% for representative 3D object detection networks and benchmarks, leading to 1.3--10.9$\times$ speedup and 1.5--12.6$\times$ energy savings compared to the ideal dense accelerator design. These sparsity-proportional performance gains equate to 4.1--28.8$\times$ speedup and 90.2--372.3$\times$ energy savings compared to the counterpart server and edge platforms.

cs.AR↗

arXiv2023

Photoinduced dynamics of flat bands in the kagome metal CoSn

CoSn is a prototypical kagome compound showing lattice-born flat bands with suppressed bandwidth over large parts of the Brillouin zone. Here, by means of time- and angle-resolved photoelectron spectroscopy, we provide direct evidence of the response to photoexcitation of the flat bands, that underlie information about localization in real space. In particular, we detect a sudden shift and broadening of the flat bands, while after one picosecond only a broadening survives. We ascribe both these effects to an ultrafast disruption of electron localization, which renormalizes the effective electron-electron interaction and affects the flat band dispersion. Since both variations are in the order of a few meV, our measurements suggest that the flat bands are resilient to near-infrared photoexcitation.

cond-mat.str-el↗

arXiv2023

Nature of charge density wave in kagome metal ScV6Sn6

Kagome lattice materials offer a fertile ground to discover novel quantum phases of matter, ranging from unconventional superconductivity and quantum spin liquids to charge orders of various profiles. However, understanding the genuine origin of the quantum phases in kagome materials is often challenging, owing to the intertwined atomic, electronic, and structural degrees of freedom. Here, we combine angle-resolved photoemission spectroscopy, phonon mode calculation, and chemical doping to elucidate the driving mechanism of the root3*root3 charge order in a newly discovered kagome metal ScV6Sn6. In contrast to the case of the archetype kagome system AV3Sb5 (A= K, Rb, Cs), the van Hove singularities in ScV6Sn6 remain intact across the charge order transition, indicating a marginal role of the electronic instability from the V kagome lattice. Instead, we identified a three-dimensional band with dominant planar Sn character opening a large charge order gap of 260 meV and strongly reconstructing the Fermi surface. Our complementary phonon dispersion calculations further emphasize the role of the structural components other than the V kagome lattice by revealing the unstable planar Sn and Sc phonon modes associated to the root3*root3 phase. Finally, in the constructed phase diagram of Sc(V1-xCrx)6Sn6, the charge order remains robust in a wide doping range x = 0 ~ 0.10 against the Fermi level shift up to ~ 120 meV, further making the electronic scenarios such as Fermi surface or saddle point nesting unlikely. Our multimodal investigations demonstrate that the physics of ScV6Sn6 is fundamentally different from the canonical kagome metal AV3Sb5, uncovering a new mechanism to induce symmetry-breaking phase transition in kagome lattice materials.

cond-mat.str-el↗

arXiv2023

Benchmarking Self-Supervised Learning on Diverse Pathology Datasets

Computational pathology can lead to saving human lives, but models are annotation hungry and pathology images are notoriously expensive to annotate. Self-supervised learning has shown to be an effective method for utilizing unlabeled data, and its application to pathology could greatly benefit its downstream tasks. Yet, there are no principled studies that compare SSL methods and discuss how to adapt them for pathology. To address this need, we execute the largest-scale study of SSL pre-training on pathology image data, to date. Our study is conducted using 4 representative SSL methods on diverse downstream tasks. We establish that large-scale domain-aligned pre-training in pathology consistently out-performs ImageNet pre-training in standard SSL settings such as linear and fine-tuning evaluations, as well as in low-label regimes. Moreover, we propose a set of domain-specific techniques that we experimentally show leads to a performance boost. Lastly, for the first time, we apply SSL to the challenging task of nuclei instance segmentation and show large and consistent performance improvements under diverse settings.

cs.CV↗

arXiv2023

Small Fermi pockets intertwined with charge stripes and pair density wave order in a kagome superconductor

The kagome superconductor family AV3Sb5 (A=Cs, K, Rb) emerged as an exciting platform to study exotic Fermi surface instabilities. Here we use spectroscopic-imaging scanning tunneling microscopy (SI-STM) and angle-resolved photoemission spectroscopy (ARPES) to reveal how the surprising cascade of higher and lower-dimensional density waves in CsV3Sb5 is intimately tied to a set of small reconstructed Fermi pockets. ARPES measurements visualize the formation of these pockets generated by a 3D charge density wave transition. The pockets are connected by dispersive q* wave vectors observed in Fourier transforms of STM differential conductance maps. As the additional 1D charge order emerges at a lower temperature, q* wave vectors become substantially renormalized, signaling further reconstruction of the Fermi pockets. Remarkably, in the superconducting state, the superconducting gap modulations give rise to an in-plane Cooper pair-density-wave at the same q* wave vectors. Our work demonstrates the intrinsic origin of the charge-stripes and the pair-density-wave in CsV3Sb5 and their relationship to the Fermi pockets. These experiments uncover a unique scenario of how Fermi pockets generated by a parent charge density wave state can provide a favorable platform for the emergence of additional density waves.

cond-mat.supr-con↗

arXiv2022

Charge order landscape and competition with superconductivity in kagome metals

In kagome metals AV3Sb5 (A = K, Rb, Cs), three-dimensional charge order (3D-CO) is the primary instability that sets the stage for other collective orders to emerge, including unidirectional stripe order, orbital flux order, electronic nematicity, and superconductivity. Here, we use high-resolution angle-resolved photoemission spectroscopy to determine the microscopic structure of three-dimensional charge order (3D-CO) in AV3Sb5 and its interplay with superconductivity. Our approach is based on identifying an unusual splitting of kagome bands induced by 3D-CO, which provides a sensitive way to refine the spatial charge patterns in neighboring kagome planes. We found a marked dependence of the 3D-CO structure on composition and doping. The observed difference between CsV3Sb5 and the other compounds potentially underpins the double-dome superconductivity in CsV3(Sb,Sn)5 and the suppression of Tc in KV3Sb5 and RbV3Sb5. Our results provide fresh insights into the rich phase diagram of AV3Sb5.

cond-mat.str-el↗

arXiv2022

Sparse Attention Acceleration with Synergistic In-Memory Pruning and On-Chip Recomputation

As its core computation, a self-attention mechanism gauges pairwise correlations across the entire input sequence. Despite favorable performance, calculating pairwise correlations is prohibitively costly. While recent work has shown the benefits of runtime pruning of elements with low attention scores, the quadratic complexity of self-attention mechanisms and their on-chip memory capacity demands are overlooked. This work addresses these constraints by architecting an accelerator, called SPRINT, which leverages the inherent parallelism of ReRAM crossbar arrays to compute attention scores in an approximate manner. Our design prunes the low attention scores using a lightweight analog thresholding circuitry within ReRAM, enabling SPRINT to fetch only a small subset of relevant data to on-chip memory. To mitigate potential negative repercussions for model accuracy, SPRINT re-computes the attention scores for the few fetched data in digital. The combined in-memory pruning and on-chip recompute of the relevant attention scores enables SPRINT to transform quadratic complexity to a merely linear one. In addition, we identify and leverage a dynamic spatial locality between the adjacent attention operations even after pruning, which eliminates costly yet redundant data fetches. We evaluate our proposed technique on a wide range of state-of-the-art transformer models. On average, SPRINT yields 7.5x speedup and 19.6x energy reduction when total 16KB on-chip memory is used, while virtually on par with iso-accuracy of the baseline models (on average 0.36% degradation).

cs.LG↗

arXiv2022

Accelerating Attention through Gradient-Based Learned Runtime Pruning

Self-attention is a key enabler of state-of-art accuracy for various transformer-based Natural Language Processing models. This attention mechanism calculates a correlation score for each word with respect to the other words in a sentence. Commonly, only a small subset of words highly correlates with the word under attention, which is only determined at runtime. As such, a significant amount of computation is inconsequential due to low attention scores and can potentially be pruned. The main challenge is finding the threshold for the scores below which subsequent computation will be inconsequential. Although such a threshold is discrete, this paper formulates its search through a soft differentiable regularizer integrated into the loss function of the training. This formulation piggy backs on the back-propagation training to analytically co-optimize the threshold and the weights simultaneously, striking a formally optimal balance between accuracy and computation pruning. To best utilize this mathematical innovation, we devise a bit-serial architecture, dubbed LeOPArd, for transformer language models with bit-level early termination microarchitectural mechanism. We evaluate our design across 43 back-end tasks for MemN2N, BERT, ALBERT, GPT-2, and Vision transformer models. Post-layout results show that, on average, LeOPArd yields 1.9x and 3.9x speedup and energy reduction, respectively, while keeping the average accuracy virtually intact (<0.2% degradation)

cs.CL↗