arXiv subjects

Mingu Kang

Publications and source records attributed to Mingu Kang.

At least 19 recordsLinked to original sources

arXiv2026

D-NOVA: In-Storage Retrieval Accelerator via Dual-Bound 3D NAND-Optimized Similarity Search with Vector Adaptation

Retrieval-Augmented Generation (RAG) enhances the factual grounding of large language model (LLM) inference by retrieving relevant information from external knowledge bases. However, its dense vector retrieval introduces significant latency and energy overhead, becoming the primary performance bottleneck. Although recent in-storage accelerators aim to reduce data movement, they still rely on host or embedded processors outside the memory, where nearly 70% of the total retrieval time is spent. As a result, they cannot fully overcome the bandwidth limitations, leading to yet another memory bottleneck. To tackle these limitations, we present D-NOVA, a hardware-software co-designed in-storage retrieval accelerator. D-NOVA executes an inverted file (IVF)-based hierarchical retrieval pipeline by deeply embedding the search functionality directly into the NAND memory array. This is achieved by incorporating a new distance metric, Dual-Bound Tight Similarity Sensing (DTS), which is specifically tailored for searching within the NAND string. In addition, we introduce a lightweight contrastive adapter that maps embedding vectors into a DTS-friendly domain, recovering near-software recall while improving performance and energy efficiency. D-NOVA is up to 41.7x faster and 71x more energy-efficient than a CPU baseline, and achieves 12.13x higher throughput while being up to 1.26x more energy-efficient than state-of-the-art in-storage RAG accelerators, demonstrating the potential of fully in-storage vector search for scalable RAG acceleration.

cs.AR↗

arXiv2026

Cross-Domain Acceleration of Open Modification Search: From Commodity Platforms to Emerging Memory and Storage Devices

Open modification search (OMS) in mass spectrometry (MS) is a data-intensive workload whose performance is dominantly limited by reference data movement rather than computation. Prior OMS accelerators have largely been evaluated in isolation, making it difficult to understand system-level trade-offs across platforms. This paper presents the first workload-driven, cross-platform survey of accelerators for MS search by studying not only commodity platforms, but also emerging memory- and storage-centric architectures, including GPUs, near-storage FPGAs, DRAM near-memory processing, ReRAM/PCM in-memory processing, and 3D NAND/FeNAND in-storage processing, under consistent algorithmic and accuracy assumptions. Leveraging a binary hyperdimensional computing (HDC)-based OMS formulation that reduces similarity evaluation to lightweight bitwise primitives and tolerates device-level non-idealities, we enable a robust execution on memory-centric architectures despite device-level non-idealities and limited computing capability. Overall, this study identifies memory- and storage-centric architectures as a key architectural breakthrough for large-scale, high-speed search acceleration, delivering up to >100x speedup and >40,000x improvement in energy efficiency.

cs.AR↗

arXiv2026

STAR-KV: Low-Rank KV Cache Compression via Soft Thresholding for Adaptive Rank Control

Low-rank projection has emerged as a promising approach for compressing the KV cache by exploiting hidden-dimension redundancy. However, prior methods rely on fixed or heuristic rank selection and struggle to achieve aggressive compression with minimal accuracy degradation. We propose STAR-KV, an adaptive low-rank KV cache compression framework with fine-grained rank control. STAR-KV encompasses 1) a differentiable thresholding mechanism that enables optimal rank selection at both attention-head and block levels, 2) a hybrid decomposition strategy that applies different low-rank factorizations according to the sensitivity of key and value projections, and 3) a low-rank-aware mixed precision quantization that leverages data statistics for near lossless low-bit quantization. Evaluated across multiple LLMs and benchmarks, STAR-KV achieves up to 75% KV cache compression and up to 20x overall KV cache reduction when combined with quantization. Enabled by custom Triton-based GPU kernels, STAR-KV delivers up to 6.9x speedup for the attention module and 3.1x end-to-end generation throughput. Our code is publicly available at: https://github.com/PriyanshBhatnagar/STAR-KV.

cs.LG↗

arXiv2026

Interaction-Breaking Adversarial Learning Framework for Robust Multi-Agent Reinforcement Learning

Cooperation is central to multi-agent reinforcement learning (MARL), yet learned coordination can be fragile when external perturbations disrupt inter-agent interactions. Prior robust MARL methods have primarily considered value-oriented attacks, leaving a gap in robustness when interaction structures themselves are corrupted. In this paper, we propose an interaction-breaking adversarial learning (IBAL) framework that takes an information-theoretic view to construct attacks that impede coordination by perturbing agents' observations and actions, and trains agents to perform reliably under such disruptions. Empirically, our approach improves robustness over existing robust MARL baselines across diverse attack settings and yields stronger performance even under agent-missing scenarios. Our code is available at https://sunwoolee0504.github.io/IBAL.

cs.LG↗

arXiv2026

Shaping Zero-Shot Coordination via State Blocking

Zero-shot coordination (ZSC) aims to enable agents to cooperate with independently trained partners without prior interaction, a key requirement for real-world multi-agent systems and human-AI collaboration. Existing approaches have largely emphasized increasing partner diversity during training, yet such strategies often fall short of achieving reliable generalization to unseen partners. We introduce State-Blocked Coordination (SBC), a simple yet effective framework that improves ZSC by inducing diverse interaction scenarios without direct environment modification. Specifically, SBC generates a family of virtual environments through state blocking, allowing agents to experience a wide range of suboptimal partner policies. Across multiple benchmarks, SBC demonstrates superior performance in zero-shot coordination, including strong generalization to human partners.

cs.LG↗

arXiv2026

TripleSumm: Adaptive Triple-Modality Fusion for Video Summarization

The exponential growth of video content necessitates effective video summarization to efficiently extract key information from long videos. However, current approaches struggle to fully comprehend complex videos, primarily because they employ static or modality-agnostic fusion strategies. These methods fail to account for the dynamic, frame-dependent variations in modality saliency inherent in video data. To overcome these limitations, we propose TripleSumm, a novel architecture that adaptively weights and fuses the contributions of visual, text, and audio modalities at the frame level. Furthermore, a significant bottleneck for research into multimodal video summarization has been the lack of comprehensive benchmarks. Addressing this bottleneck, we introduce MoSu (Most Replayed Multimodal Video Summarization), the first large-scale benchmark that provides all three modalities. Extensive experiments demonstrate that TripleSumm achieves state-of-the-art performance, outperforming existing methods by a significant margin on four benchmarks, including MoSu. Our code and dataset are available at https://github.com/smkim37/TripleSumm.

cs.CV↗

arXiv2026

VERSA: Verified Event Data Format for Reliable Soccer Analytics

Event stream data is a critical resource for fine-grained analysis across various domains, including financial transactions, system operations, and sports. In sports, it is actively used for fine-grained analyses such as quantifying player contributions and identifying tactical patterns. However, the reliability of these models is fundamentally limited by inherent data quality issues that cause logical inconsistencies (e.g., incorrect event ordering or missing events). To this end, this study proposes VERSA (Verified Event Data Format for Reliable Soccer Analytics), a systematic verification framework that ensures the integrity of event stream data within the soccer domain. VERSA is based on a state-transition model that defines valid event sequences, thereby enabling the automatic detection and correction of anomalous patterns within the event stream data. Notably, our examination of event data from the K League 1 (2024 season), provided by Bepro, detected that 18.81% of all recorded events exhibited logical inconsistencies. Addressing such integrity issues, our experiments demonstrate that VERSA significantly enhances cross-provider consistency, ensuring stable and unified data representation across heterogeneous sources. Furthermore, we demonstrate that data refined by VERSA significantly improves the robustness and performance of a downstream task called VAEP, which evaluates player contributions. These results highlight that the verification process is highly effective in increasing the reliability of data-driven analysis.

cs.AI↗

arXiv2025

FSL-HDnn: A 40 nm Few-shot On-Device Learning Accelerator with Integrated Feature Extraction and Hyperdimensional Computing

This paper introduces FSL-HDnn, an energy-efficient accelerator that implements the end-to-end pipeline of feature extraction and on-device few-shot learning (FSL). The accelerator addresses fundamental challenges of on-device learning (ODL) for resource-constrained edge applications through two synergistic modules: a parameter-efficient feature extractor employing weight clustering and an FSL classifier based on hyperdimensional computing (HDC). The feature extractor exploits the weight clustering mechanism to reduce computational complexity, while the HDC-based FSL classifier eliminates gradient-based back propagation operations, enabling single-pass training with substantially reduced latency. Additionally, FSL-HDnn enables low-latency ODL and inference via two proposed optimization strategies, including an early-exit mechanism with branch feature extraction and batched single-pass training that improves hardware utilization. Measurement results demonstrate that our chip fabricated in a 40 nm CMOS process delivers superior training energy efficiency of 6 mJ/image and end-to-end training throughput of 28 images/s on a 10-way 5-shot FSL task. The end-to-end training latency is also reduced by 2x to 20.9x compared to state-of-the-art ODL chips.

cs.AR↗

arXiv2025

Sangam: Chiplet-Based DRAM-PIM Accelerator with CXL Integration for LLM Inferencing

Large Language Models (LLMs) are becoming increasingly data-intensive due to growing model sizes, and they are becoming memory-bound as the context length and, consequently, the key-value (KV) cache size increase. Inference, particularly the decoding phase, is dominated by memory-bound GEMV or flat GEMM operations with low operational intensity (OI), making it well-suited for processing-in-memory (PIM) approaches. However, existing in/near-memory solutions face critical limitations such as reduced memory capacity due to the high area cost of integrating processing elements (PEs) within DRAM chips, and limited PE capability due to the constraints of DRAM fabrication technology. This work presents a chiplet-based memory module that addresses these limitations by decoupling logic and memory into chiplets fabricated in heterogeneous technology nodes and connected via an interposer. The logic chiplets sustain high bandwidth access to the DRAM chiplets, which house the memory banks, and enable the integration of advanced processing components such as systolic arrays and SRAM-based buffers to accelerate memory-bound GEMM kernels, capabilities that were not feasible in prior PIM architectures. We propose Sangam, a CXL-attached PIM-chiplet based memory module that can either act as a drop-in replacement for GPUs or co-executes along side the GPUs. Sangam achieves speedup of 3.93, 4.22, 2.82x speedup in end-to-end query latency, 10.3, 9.5, 6.36x greater decoding throughput, and order of magnitude energy savings compared to an H100 GPU for varying input size, output length, and batch size on LLaMA 2-7B, Mistral-7B, and LLaMA 3-70B, respectively.

cs.AR↗

arXiv2025

Role of Phase Fluctuation in Dynamic Competition Between Charge Order and Superconductivity in Cuprates

Phase fluctuations are a key factor distinguishing nonthermal (ultrafast) and thermal phase transitions. Charge order in cuprates is characterized by short-range coherence while competing with superconductivity, and as such, it provides a representative case to study the role of phase fluctuation in coupled order parameter dynamics. In this work, we investigated the intertwined evolution of charge order and superconductivity in cuprate/manganite heterostructures using time-resolved resonant X-ray scattering. The resulting dynamics are analyzed within a space- and time-dependent nonperturbative model capturing both amplitude and phase dynamics. At low fluence, photo-induced suppression of superconductivity results in a nonthermal enhancement of charge order, underscoring the dynamic competition between charge order and superconductivity. With increasing fluence, the slowing down of melting and recovery dynamics is observed, indicating a critical role of phase fluctuations. At high fluence, both charge order and superconductivity remain suppressed for an extended time window due to decoupling between amplitude and phase dynamics and the delayed recovery of phase coherence. Our work underscores the importance of phase fluctuation for understanding the dynamic competition between order parameters in cuprates.

cond-mat.str-el↗

arXiv2025

FeNOMS: Enhancing Open Modification Spectral Library Search with In-Storage Processing on Ferroelectric NAND (FeNAND) Flash

The rapid expansion of mass spectrometry (MS) data, now exceeding hundreds of terabytes, poses significant challenges for efficient, large-scale library search - a critical component for drug discovery. Traditional processors struggle to handle this data volume efficiently, making in-storage computing (ISP) a promising alternative. This work introduces an ISP architecture leveraging a 3D Ferroelectric NAND (FeNAND) structure, providing significantly higher density, faster speeds, and lower voltage requirements compared to traditional NAND flash. Despite its superior density, the NAND structure has not been widely utilized in ISP applications due to limited throughput associated with row-by-row reads from serially connected cells. To overcome these limitations, we integrate hyperdimensional computing (HDC), a brain-inspired paradigm that enables highly parallel processing with simple operations and strong error tolerance. By combining HDC with the proposed dual-bound approximate matching (D-BAM) distance metric, tailored to the FeNAND structure, we parallelize vector computations to enable efficient MS spectral library search, achieving 43x speedup and 21x higher energy efficiency over state-of-the-art 3D NAND methods, while maintaining comparable accuracy.

cs.AR↗

arXiv2025

Stratum: System-Hardware Co-Design with Tiered Monolithic 3D-Stackable DRAM for Efficient MoE Serving

As Large Language Models (LLMs) continue to evolve, Mixture of Experts (MoE) architecture has emerged as a prevailing design for achieving state-of-the-art performance across a wide range of tasks. MoE models use sparse gating to activate only a handful of expert sub-networks per input, achieving billion-parameter capacity with inference costs akin to much smaller models. However, such models often pose challenges for hardware deployment due to the massive data volume introduced by the MoE layers. To address the challenges of serving MoE models, we propose Stratum, a system-hardware co-design approach that combines the novel memory technology Monolithic 3D-Stackable DRAM (Mono3D DRAM), near-memory processing (NMP), and GPU acceleration. The logic and Mono3D DRAM dies are connected through hybrid bonding, whereas the Mono3D DRAM stack and GPU are interconnected via silicon interposer. Mono3D DRAM offers higher internal bandwidth than HBM thanks to the dense vertical interconnect pitch enabled by its monolithic structure, which supports implementations of higher-performance near-memory processing. Furthermore, we tackle the latency differences introduced by aggressive vertical scaling of Mono3D DRAM along the z-dimension by constructing internal memory tiers and assigning data across layers based on access likelihood, guided by topic-based expert usage prediction to boost NMP throughput. The Stratum system achieves up to 8.29x improvement in decoding throughput and 7.66x better energy efficiency across various benchmarks compared to GPU baselines.

cs.AR↗

arXiv2025

Clo-HDnn: A 4.66 TFLOPS/W and 3.78 TOPS/W Continual On-Device Learning Accelerator with Energy-efficient Hyperdimensional Computing via Progressive Search

Clo-HDnn is an on-device learning (ODL) accelerator designed for emerging continual learning (CL) tasks. Clo-HDnn integrates hyperdimensional computing (HDC) along with low-cost Kronecker HD Encoder and weight clustering feature extraction (WCFE) to optimize accuracy and efficiency. Clo-HDnn adopts gradient-free CL to efficiently update and store the learned knowledge in the form of class hypervectors. Its dual-mode operation enables bypassing costly feature extraction for simpler datasets, while progressive search reduces complexity by up to 61% by encoding and comparing only partial query hypervectors. Achieving 4.66 TFLOPS/W (FE) and 3.78 TOPS/W (classifier), Clo-HDnn delivers 7.77x and 4.85x higher energy efficiency compared to SOTA ODL accelerators.

cs.AR↗

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

DisCoRD: Discrete Tokens to Continuous Motion via Rectified Flow Decoding

Human motion is inherently continuous and dynamic, posing significant challenges for generative models. While discrete generation methods are widely used, they suffer from limited expressiveness and frame-wise noise artifacts. In contrast, continuous approaches produce smoother, more natural motion but often struggle to adhere to conditioning signals due to high-dimensional complexity and limited training data. To resolve this 'discord' between discrete and continuous representations we introduce DisCoRD: Discrete Tokens to Continuous Motion via Rectified Flow Decoding, a novel method that leverages rectified flow to decode discrete motion tokens in the continuous, raw motion space. Our core idea is to frame token decoding as a conditional generation task, ensuring that DisCoRD captures fine-grained dynamics and achieves smoother, more natural motions. Compatible with any discrete-based framework, our method enhances naturalness without compromising faithfulness to the conditioning signals on diverse settings. Extensive evaluations demonstrate that DisCoRD achieves state-of-the-art performance, with FID of 0.032 on HumanML3D and 0.169 on KIT-ML. These results establish DisCoRD as a robust solution for bridging the divide between discrete efficiency and continuous realism. Project website: https://whwjdqls.github.io/discord-motion/

cs.CV↗