Searcharxiv⌕ Search

arXiv subjects

Youngeun Kim

Publications and source records attributed to Youngeun Kim.

At least 19 recordsLinked to original sources
arXiv2026

Deferred Audio Pruning with Local Audio-Visual Dynamics for Omni-LLMs

Omni-modal LLMs jointly process audio, video, and text, but long multimodal sequences incur substantial prefill and KV-cache costs. Existing omni-modal compression methods primarily focus on pre-LLM token reduction, leaving modality-specific compression across the LLM boundary underexplored. We propose A-PACK, a two-stage framework that defers audio pruning until query-conditioned multimodal interactions emerge. Our analysis shows that audio exhibits higher task-relevant information density and representational diversity per token than video. We further find that local audio-visual dynamics provide a more effective cue for visual selection than token-wise matching. We therefore preserve audio and compress video with local dynamics before the LLM, then progressively prune low-relevance audio and visual tokens and their KV-cache entries inside the LLM. Across four benchmarks on Qwen2.5-Omni-7B/3B, A-PACK achieves the strongest average performance among the evaluated prior methods while reducing prefill FLOPs by up to 78% and improving decoding throughput by up to 2.21x.

cs.AI↗
arXiv2026

SkillRet: A Large-Scale Benchmark for Skill Retrieval in LLM Agents

As LLM agents are increasingly deployed with large libraries of reusable skills, selecting the right skill for a user request has become a critical systems challenge. In small libraries, users may invoke skills explicitly by name, but this assumption breaks down as skill ecosystems grow under tight context and latency budgets. Despite its practical importance, skill retrieval remains underexplored, with limited benchmarks and little understanding of retrieval behavior on realistic skill libraries. To address this gap, we introduce SkillRet, a large-scale benchmark for skill retrieval in LLM agents. SkillRet contains 16,129 public agent skills, organized with structured semantic tags and a two-level taxonomy spanning 6 major categories and 18 sub-categories. It provides 63,259 training samples and 4,392 evaluation queries with disjoint skill pools, enabling both benchmarking and retrieval-oriented training. Across a diverse set of retrievers, we find that skill retrieval remains far from solved: off-the-shelf models struggle on realistic large-scale skill libraries, and prior skill-retrieval models still leave substantial headroom. Task-specific fine-tuning on SkillRet improves NDCG@10 by 12.9 points over the strongest prior retriever and by 16.2 points over the strongest off-the-shelf retriever. Our analysis further suggests that these gains arise because fine-tuned models better focus on the small skill-relevant signals within long and noisy queries. These results establish SkillRet as a strong benchmark and foundation for future research on retrieval in large-scale agent systems. We publicly release the benchmark (https://huggingface.co/datasets/ThakiCloud/SKILLRET), code (https://github.com/ThakiCloud/SKILLRET), and model checkpoints (0.6B: https://huggingface.co/ThakiCloud/SKILLRET-Embedding-0.6B; 8B: https://huggingface.co/ThakiCloud/SKILLRET-Embedding-8B).

cs.AI↗
arXiv2026

Real-Time Visual Attribution Streaming in Thinking Model

We present an amortized framework for real-time visual attribution streaming in multimodal thinking models. When these models generate code from a screenshot or solve math problems from images, their long reasoning traces should be grounded in visual evidence. However, verifying this reliance is challenging: faithful causal methods require costly repeated backward passes or perturbations, while raw attention maps offer instant access, they lack causal validity. To resolve this, we introduce an amortized approach that learns to estimate the causal effects of semantic regions directly from the rich signals encoded in attention features. Across five diverse benchmarks and four thinking models, our approach achieves faithfulness comparable to exhaustive causal methods while enabling visual attribution streaming, where users observe grounding evidence as the model reasons, not after. Our results demonstrate that real-time, faithful attribution in multimodal thinking models is achievable through lightweight learning, not brute-force computation.

cs.CV↗
arXiv2026

VisRef: Visual Refocusing while Thinking Improves Test-Time Scaling in Multi-Modal Large Reasoning Models

Advances in large reasoning models have shown strong performance on complex reasoning tasks by scaling test-time compute through extended reasoning. However, recent studies observe that in vision-dependent tasks, extended textual reasoning at inference time can degrade performance as models progressively lose attention to visual tokens and increasingly rely on textual priors alone. To address this, prior works use reinforcement learning (RL)-based fine-tuning to route visual tokens or employ refocusing mechanisms during reasoning. While effective, these methods are computationally expensive, requiring large-scale data generation and policy optimization. To leverage the benefits of test-time compute without additional RL fine-tuning, we propose VisRef, a visually grounded test-time scaling framework. Our key idea is to actively guide the reasoning process by re-injecting a coreset of visual tokens that are semantically relevant to the reasoning context while remaining diverse and globally representative of the image, enabling more grounded multi-modal reasoning. Experiments on three visual reasoning benchmarks with state-of-the-art multi-modal large reasoning models demonstrate that, under fixed test-time compute budgets, VisRef consistently outperforms existing test-time scaling approaches by up to 6.4%.

cs.CV↗
arXiv2026

MC-GRPO: Median-Centered Group Relative Policy Optimization for Small-Rollout Reinforcement Learning

Group-relative policy optimization methods train language models by generating multiple rollouts per prompt and normalizing rewards with a shared mean reward baseline. In resource-constrained settings where the rollout budget is small, accuracy often degrades. We find that noise in the shared baseline induces advantage sign flips, where some rollouts receive an incorrect advantage sign, and the update direction is reversed. To address this, we propose Median-Centered Group Relative Policy Optimization (MC-GRPO), a simple and effective solution for small-rollout training. Our main idea is to replace the mean baseline with a median baseline: the median is far less sensitive to outlier rewards than the mean, mitigating the sign flips under small rollout size (G). We generate one additional rollout for median reference (G+1), and compute advantages by using the group median. With an odd-sized group, exactly one completion is the median and receives zero advantage, we exclude this pivot rollout from backpropagation so the number of gradient-contributing samples per prompt remains G, preserving the core update cost of standard G-rollout training. Across various GRPO-family methods and a wide range of models and scales, this median-centered training consistently improves stability and final accuracy in the low-rollout regime, reducing the gap between G=2 and G=8 to within 1%. Code is available at https://github.com/lotusroot-kim/MC-GRPO

cs.LG↗
arXiv2025

MD-SNN: Membrane Potential-aware Distillation on Quantized Spiking Neural Network

Spiking Neural Networks (SNNs) offer a promising and energy-efficient alternative to conventional neural networks, thanks to their sparse binary activation. However, they face challenges regarding memory and computation overhead due to complex spatio-temporal dynamics and the necessity for multiple backpropagation computations across timesteps during training. To mitigate this overhead, compression techniques such as quantization are applied to SNNs. Yet, naively applying quantization to SNNs introduces a mismatch in membrane potential, a crucial factor for the firing of spikes, resulting in accuracy degradation. In this paper, we introduce Membrane-aware Distillation on quantized Spiking Neural Network (MD-SNN), which leverages membrane potential to mitigate discrepancies after weight, membrane potential, and batch normalization quantization. To our knowledge, this study represents the first application of membrane potential knowledge distillation in SNNs. We validate our approach on various datasets, including CIFAR10, CIFAR100, N-Caltech101, and TinyImageNet, demonstrating its effectiveness for both static and dynamic data scenarios. Furthermore, for hardware efficiency, we evaluate the MD-SNN with SpikeSim platform, finding that MD-SNNs achieve 14.85X lower energy-delay-area product (EDAP), 2.64X higher TOPS/W, and 6.19X higher TOPS/mm2 compared to floating point SNNs at iso-accuracy on N-Caltech101 dataset.

cs.NE↗
arXiv2025

ZOO-Prune: Training-Free Token Pruning via Zeroth-Order Gradient Estimation in Vision-Language Models

Large Vision-Language Models (VLMs) enable strong multimodal reasoning but incur heavy inference costs from redundant visual tokens. Token pruning alleviates this issue, yet existing approaches face limitations. Attention-based methods rely on raw attention scores, which are often unstable across layers and heads and can lead to redundant selections. Diversity-based methods improve robustness by selecting tokens far apart in feature space, but risk dropping regions needed for accurate prediction. We propose ZOO-Prune, a training-free framework built on the intuition that highly sensitive tokens have a stronger influence on the model's output and capture complementary visual cues rather than redundant ones. To achieve this, we estimate token sensitivity using zeroth-order perturbations at the lightweight projection layer. This measures how small random perturbations affect the projected features and enables efficient approximation of each token's influence without backpropagation. Extensive experiments across multiple VLMs and benchmarks show that ZOO-Prune consistently outperforms prior methods while pruning up to 94.4% of tokens without sacrificing accuracy. Our method also improves efficiency, reaching up to 2.30x faster end-to-end inference compared to the baseline.

cs.CV↗
arXiv2025

Backpropagation-Free Test-Time Adaptation via Probabilistic Gaussian Alignment

Test-time adaptation (TTA) enhances the zero-shot robustness under distribution shifts by leveraging unlabeled test data during inference. Despite notable advances, several challenges still limit its broader applicability. First, most methods rely on backpropagation or iterative optimization, which limits scalability and hinders real-time deployment. Second, they lack explicit modeling of class-conditional feature distributions. This modeling is crucial for producing reliable decision boundaries and calibrated predictions, but it remains underexplored due to the lack of both source data and supervision at test time. In this paper, we propose ADAPT, an Advanced Distribution-Aware and backPropagation-free Test-time adaptation method. We reframe TTA as a Gaussian probabilistic inference task by modeling class-conditional likelihoods using gradually updated class means and a shared covariance matrix. This enables closed-form, training-free inference. To correct potential likelihood bias, we introduce lightweight regularization guided by CLIP priors and a historical knowledge bank. ADAPT requires no source data, no gradient updates, and no full access to target data, supporting both online and transductive settings. Extensive experiments across diverse benchmarks demonstrate that our method achieves state-of-the-art performance under a wide range of distribution shifts with superior scalability and robustness.

cs.CV↗
arXiv2025

Task Vector Quantization for Memory-Efficient Model Merging

Model merging enables efficient multi-task models by combining task-specific fine-tuned checkpoints. However, storing multiple task-specific checkpoints requires significant memory, limiting scalability and restricting model merging to larger models and diverse tasks. In this paper, we propose quantizing task vectors (i.e., the difference between pre-trained and fine-tuned checkpoints) instead of quantizing fine-tuned checkpoints. We observe that task vectors exhibit a narrow weight range, enabling low precision quantization (e.g., 4 bit) within existing task vector merging frameworks. To further mitigate quantization errors within ultra-low bit precision (e.g., 2 bit), we introduce Residual Task Vector Quantization, which decomposes the task vector into a base vector and offset component. We allocate bits based on quantization sensitivity, ensuring precision while minimizing error within a memory budget. Experiments on image classification and dense prediction show our method maintains or improves model merging performance while using only 8% of the memory required for full-precision checkpoints.

cs.LG↗
arXiv2024

Spiking Transformer with Spatial-Temporal Attention

Spike-based Transformer presents a compelling and energy-efficient alternative to traditional Artificial Neural Network (ANN)-based Transformers, achieving impressive results through sparse binary computations. However, existing spike-based transformers predominantly focus on spatial attention while neglecting crucial temporal dependencies inherent in spike-based processing, leading to suboptimal feature representation and limited performance. To address this limitation, we propose Spiking Transformer with Spatial-Temporal Attention (STAtten), a simple and straightforward architecture that efficiently integrates both spatial and temporal information in the self-attention mechanism. STAtten introduces a block-wise computation strategy that processes information in spatial-temporal chunks, enabling comprehensive feature capture while maintaining the same computational complexity as previous spatial-only approaches. Our method can be seamlessly integrated into existing spike-based transformers without architectural overhaul. Extensive experiments demonstrate that STAtten significantly improves the performance of existing spike-based transformers across both static and neuromorphic datasets, including CIFAR10/100, ImageNet, CIFAR10-DVS, and N-Caltech101. The code is available at https://github.com/Intelligent-Computing-Lab-Yale/STAtten

cs.NE↗
arXiv2024

Open-World Dynamic Prompt and Continual Visual Representation Learning

The open world is inherently dynamic, characterized by ever-evolving concepts and distributions. Continual learning (CL) in this dynamic open-world environment presents a significant challenge in effectively generalizing to unseen test-time classes. To address this challenge, we introduce a new practical CL setting tailored for open-world visual representation learning. In this setting, subsequent data streams systematically introduce novel classes that are disjoint from those seen in previous training phases, while also remaining distinct from the unseen test classes. In response, we present Dynamic Prompt and Representation Learner (DPaRL), a simple yet effective Prompt-based CL (PCL) method. Our DPaRL learns to generate dynamic prompts for inference, as opposed to relying on a static prompt pool in previous PCL methods. In addition, DPaRL jointly learns dynamic prompt generation and discriminative representation at each training stage whereas prior PCL methods only refine the prompt learning throughout the process. Our experimental results demonstrate the superiority of our approach, surpassing state-of-the-art methods on well-established open-world image retrieval benchmarks by an average of 4.7% improvement in Recall@1 performance.

cs.CV↗
arXiv2024

ReSpike: Residual Frames-based Hybrid Spiking Neural Networks for Efficient Action Recognition

Spiking Neural Networks (SNNs) have emerged as a compelling, energy-efficient alternative to traditional Artificial Neural Networks (ANNs) for static image tasks such as image classification and segmentation. However, in the more complex video classification domain, SNN-based methods fall considerably short of ANN-based benchmarks due to the challenges in processing dense frame sequences. To bridge this gap, we propose ReSpike, a hybrid framework that synergizes the strengths of ANNs and SNNs to tackle action recognition tasks with high accuracy and low energy cost. By decomposing film clips into spatial and temporal components, i.e., RGB image Key Frames and event-like Residual Frames, ReSpike leverages ANN for learning spatial information and SNN for learning temporal information. In addition, we propose a multi-scale cross-attention mechanism for effective feature fusion. Compared to state-of-the-art SNN baselines, our ReSpike hybrid architecture demonstrates significant performance improvements (e.g., >30% absolute accuracy improvement on HMDB-51, UCF-101, and Kinetics-400). Furthermore, ReSpike achieves comparable performance with prior ANN approaches while bringing better accuracy-energy tradeoff.

cs.CV↗
arXiv2024

TReX- Reusing Vision Transformer's Attention for Efficient Xbar-based Computing

Due to the high computation overhead of Vision Transformers (ViTs), In-memory Computing architectures are being researched towards energy-efficient deployment in edge-computing scenarios. Prior works have proposed efficient algorithm-hardware co-design and IMC-architectural improvements to improve the energy-efficiency of IMC-implemented ViTs. However, all prior works have neglected the overhead and co-depencence of attention blocks on the accuracy-energy-delay-area of IMC-implemented ViTs. To this end, we propose TReX- an attention-reuse-driven ViT optimization framework that effectively performs attention reuse in ViT models to achieve optimal accuracy-energy-delay-area tradeoffs. TReX optimally chooses the transformer encoders for attention reuse to achieve near iso-accuracy performance while meeting the user-specified delay requirement. Based on our analysis on the Imagenet-1k dataset, we find that TReX achieves 2.3x (2.19x) EDAP reduction and 1.86x (1.79x) TOPS/mm2 improvement with ~1% accuracy drop in case of DeiT-S (LV-ViT-S) ViT models. Additionally, TReX achieves high accuracy at high EDAP reduction compared to state-of-the-art token pruning and weight sharing approaches. On NLP tasks such as CoLA, TReX leads to 2% higher non-ideal accuracy compared to baseline at 1.6x lower EDAP.

cs.AI↗
arXiv2024

When In-memory Computing Meets Spiking Neural Networks -- A Perspective on Device-Circuit-System-and-Algorithm Co-design

This review explores the intersection of bio-plausible artificial intelligence in the form of Spiking Neural Networks (SNNs) with the analog In-Memory Computing (IMC) domain, highlighting their collective potential for low-power edge computing environments. Through detailed investigation at the device, circuit, and system levels, we highlight the pivotal synergies between SNNs and IMC architectures. Additionally, we emphasize the critical need for comprehensive system-level analyses, considering the inter-dependencies between algorithms, devices, circuit & system parameters, crucial for optimal performance. An in-depth analysis leads to identification of key system-level bottlenecks arising from device limitations which can be addressed using SNN-specific algorithm-hardware co-design techniques. This review underscores the imperative for holistic device to system design space co-exploration, highlighting the critical aspects of hardware and algorithm research endeavors for low-power neuromorphic solutions.

cs.NE↗
arXiv2024

LoAS: Fully Temporal-Parallel Dataflow for Dual-Sparse Spiking Neural Networks

Spiking Neural Networks (SNNs) have gained significant research attention in the last decade due to their potential to drive resource-constrained edge devices. Though existing SNN accelerators offer high efficiency in processing sparse spikes with dense weights, opportunities are less explored in SNNs with sparse weights, i.e., dual-sparsity. In this work, we study the acceleration of dual-sparse SNNs, focusing on their core operation, sparse-matrix-sparse-matrix multiplication (spMspM). We observe that naively running a dual-sparse SNN on existing spMspM accelerators designed for dual-sparse Artificial Neural Networks (ANNs) exhibits sub-optimal efficiency. The main challenge is that processing timesteps, a natural property of SNNs, introduces an extra loop to ANN spMspM, leading to longer latency and more memory traffic. To address the problem, we propose a fully temporal-parallel (FTP) dataflow, which minimizes both data movement across timesteps and the end-to-end latency of dual-sparse SNNs. To maximize the efficiency of FTP dataflow, we propose an FTP-friendly spike compression mechanism that efficiently compresses single-bit spikes and ensures contiguous memory access. We further propose an FTP-friendly inner-join circuit that can lower the cost of the expensive prefix-sum circuits with almost no throughput penalty. All the above techniques for FTP dataflow are encapsulated in LoAS, a Low-latency inference Accelerator for dual-sparse SNNs. With FTP dataflow, compression, and inner-join, running dual-sparse SNN workloads on LoAS demonstrates significant speedup (up to $8.51\times$) and energy reduction (up to $3.68\times$) compared to running it on prior dual-sparse accelerators.

cs.AR↗
arXiv2024

Do We Really Need a Large Number of Visual Prompts?

Due to increasing interest in adapting models on resource-constrained edges, parameter-efficient transfer learning has been widely explored. Among various methods, Visual Prompt Tuning (VPT), prepending learnable prompts to input space, shows competitive fine-tuning performance compared to training of full network parameters. However, VPT increases the number of input tokens, resulting in additional computational overhead. In this paper, we analyze the impact of the number of prompts on fine-tuning performance and self-attention operation in a vision transformer architecture. Through theoretical and empirical analysis we show that adding more prompts does not lead to linear performance improvement. Further, we propose a Prompt Condensation (PC) technique that aims to prevent performance degradation from using a small number of prompts. We validate our methods on FGVC and VTAB-1k tasks and show that our approach reduces the number of prompts by ~70% while maintaining accuracy.

cs.CV↗
arXiv2024

Workload-Balanced Pruning for Sparse Spiking Neural Networks

Pruning for Spiking Neural Networks (SNNs) has emerged as a fundamental methodology for deploying deep SNNs on resource-constrained edge devices. Though the existing pruning methods can provide extremely high weight sparsity for deep SNNs, the high weight sparsity brings a workload imbalance problem. Specifically, the workload imbalance happens when a different number of non-zero weights are assigned to hardware units running in parallel. This results in low hardware utilization and thus imposes longer latency and higher energy costs. In preliminary experiments, we show that sparse SNNs (~98% weight sparsity) can suffer as low as ~59% utilization. To alleviate the workload imbalance problem, we propose u-Ticket, where we monitor and adjust the weight connections of the SNN during Lottery Ticket Hypothesis (LTH) based pruning, thus guaranteeing the final ticket gets optimal utilization when deployed onto the hardware. Experiments indicate that our u-Ticket can guarantee up to 100% hardware utilization, thus reducing up to 76.9% latency and 63.8% energy cost compared to the non-utilization-aware LTH method.

cs.NE↗
arXiv2024

One-stage Prompt-based Continual Learning

Prompt-based Continual Learning (PCL) has gained considerable attention as a promising continual learning solution as it achieves state-of-the-art performance while preventing privacy violation and memory overhead issues. Nonetheless, existing PCL approaches face significant computational burdens because of two Vision Transformer (ViT) feed-forward stages; one is for the query ViT that generates a prompt query to select prompts inside a prompt pool; the other one is a backbone ViT that mixes information between selected prompts and image tokens. To address this, we introduce a one-stage PCL framework by directly using the intermediate layer's token embedding as a prompt query. This design removes the need for an additional feed-forward stage for query ViT, resulting in ~50% computational cost reduction for both training and inference with marginal accuracy drop < 1%. We further introduce a Query-Pool Regularization (QR) loss that regulates the relationship between the prompt query and the prompt pool to improve representation power. The QR loss is only applied during training time, so there is no computational overhead at inference from the QR loss. With the QR loss, our approach maintains ~ 50% computational cost reduction during inference as well as outperforms the prior two-stage PCL methods by ~1.4% on public class-incremental continual learning benchmarks including CIFAR-100, ImageNet-R, and DomainNet.

cs.CV↗