arXiv subjects

Joonhyung Lee

Publications and source records attributed to Joonhyung Lee.

10 recordsLinked to original sources

arXiv2026

SelfJudge: Faster Speculative Decoding via Self-Supervised Judge Verification

Speculative decoding accelerates LLM inference by verifying candidate tokens from a draft model against a larger target model. Recent judge decoding boosts this process by relaxing verification criteria by accepting draft tokens that may exhibit minor discrepancies from target model output, but existing methods are restricted by their reliance on human annotations or tasks with verifiable ground truths, limiting generalizability across diverse NLP tasks. We propose SelfJudge, which trains judge verifiers via self-supervision of the target model. Our method measures semantic preservation by assessing whether token-substituted responses preserve the meaning of original responses, enabling automatic verifier training across diverse NLP tasks. Our experiments show SelfJudge achieves superior inference-accuracy trade-offs than judge decoding baselines, offering a broadly applicable solution for faster LLM inference.

cs.CL↗

arXiv2026

Affine-Scaled Attention: Towards Flexible and Stable Transformer Attention

Transformer attention is typically implemented using softmax normalization, which enforces attention weights with unit sum normalization. While effective in many settings, this constraint can limit flexibility in controlling attention magnitudes and may contribute to overly concentrated or unstable attention patterns during training. Prior work has explored modifications such as attention sinks or gating mechanisms, but these approaches provide only limited or indirect control over attention reweighting. We propose Affine-Scaled Attention, a simple extension to standard attention that introduces input-dependent scaling and a corresponding bias term applied to softmax-normalized attention weights. This design relaxes the strict normalization constraint while maintaining aggregation of value representations, allowing the model to adjust both the relative distribution and the scale of attention in a controlled manner. We empirically evaluate Affine-Scaled Attention in large-scale language model pretraining across multiple model sizes. Experimental results show consistent improvements in training stability, optimization behavior, and downstream task performance compared to standard softmax attention and attention sink baselines. These findings suggest that modest reweighting of attention outputs provides a practical and effective way to improve attention behavior in Transformer models.

cs.CL↗

arXiv2025

An Inquiry into Datacenter TCO for LLM Inference with FP8

As large language models (LLMs) continue to scale, the high power consumption of AI accelerators in datacenters presents significant challenges, substantially increasing the total cost of ownership (TCO) for cloud service providers (CSPs) that provide LLM inference. In this work, we analyze the computational characteristics of LLM inference from a TCO perspective and present a generalizable framework to compare AI accelerators across diverse operational requirements. Using this model, we investigate key workload characteristics influencing TCO for AI accelerators from Intel (Gaudi 2 & 3) and NVIDIA (H100 & H200), especially thin GEMM utilization and FP8 quantization. In particular, as FP8 emerges as the baseline precision for next-generation LLMs, understanding how different architectures implement and benefit from low-precision computation is increasingly critical. Throughput on thin GEMMs has a greater impact on TCO than theoretical hardware peak throughput because the memory-bound decode phase is dominated by GEMV-like computations. We find that Gaudi HPUs achieve superior utilization on thin GEMMs compared to their counterparts, especially in FP8-quantized models. Our result underscores the importance of empirical, workload-level analysis in evaluating accelerator performance, rather than relying solely on theoretical hardware specifications. By studying the interaction between power consumption, quantization strategies, and hardware architecture, we provide insights to support informed deployment decisions and guide future accelerator designs aimed at improving the TCO of LLM inference workloads.

cs.LG↗

arXiv2025

To FP8 and Back Again: Quantifying Reduced Precision Effects on LLM Training Stability

The massive computational costs associated with large language model (LLM) pretraining have spurred great interest in reduced-precision floating-point representations to accelerate the process. As a result, the BrainFloat16 (BF16) precision has become the de facto standard for LLM training, with hardware support included in recent generations of accelerators. This trend has gone even further in the latest processors, where FP8 has recently been introduced. However, prior experience with FP16, which was found to be less stable than BF16, raises concerns as to whether FP8, with even fewer bits than FP16, can be a cost-effective option for LLM training. We argue that reduced-precision training schemes must have similar training stability and hyperparameter sensitivities to their higher-precision counterparts in order to be cost-effective. However, we find that currently available methods for FP8 training are not robust enough to allow their use as economical replacements. This prompts us to investigate the stability of reduced-precision LLM training in terms of robustness across random seeds, learning rates, and datasets. To this end, we propose new evaluation techniques and a new metric for quantifying loss landscape sharpness in autoregressive language models. By simulating incremental bit reductions in floating-point representations, we analyze the relationship between representational power and training stability with the intent of aiding future research into the field.

cs.LG↗

arXiv2025

Debunking the CUDA Myth Towards GPU-based AI Systems

This paper presents a comprehensive evaluation of Intel Gaudi NPUs as an alternative to NVIDIA GPUs, which is currently the de facto standard in AI system design. First, we create a suite of microbenchmarks to compare Intel Gaudi-2 with NVIDIA A100, showing that Gaudi-2 achieves competitive performance not only in primitive AI compute, memory, and communication operations but also in executing several important AI workloads end-to-end. We then assess Gaudi NPU's programmability by discussing several software-level optimization strategies to employ for implementing critical FBGEMM operators and vLLM, evaluating their efficiency against GPU-optimized counterparts. Results indicate that Gaudi-2 achieves energy efficiency comparable to A100, though there are notable areas for improvement in terms of software maturity. Overall, we conclude that, with effective integration into high-level AI frameworks, Gaudi NPUs could challenge NVIDIA GPU's dominance in the AI server market, though further improvements are necessary to fully compete with NVIDIA's robust software ecosystem.

cs.DC↗

arXiv2025

Faster Inference of LLMs using FP8 on the Intel Gaudi

Low-precision data types are essential in modern neural networks during both training and inference as they enhance throughput and computational capacity by better exploiting available hardware resources. Despite the incorporation of FP8 in commercially available neural network accelerators, a comprehensive exposition of its underlying mechanisms, along with rigorous performance and accuracy evaluations, is still lacking. In this work, we contribute in three significant ways. First, we analyze the implementation details and quantization options associated with FP8 for inference on the Intel Gaudi AI accelerator. Second, we empirically quantify the throughput improvements afforded by the use of FP8 at both the operator level and in end-to-end scenarios. Third, we assess the accuracy impact of various FP8 quantization methods. Our experimental results indicate that the Intel Gaudi 2 accelerator consistently achieves high computational unit utilization, frequently exceeding 90% MFU, while incurring an accuracy degradation of less than 1%.

cs.AR↗

arXiv2024

CLARA: Classifying and Disambiguating User Commands for Reliable Interactive Robotic Agents

In this paper, we focus on inferring whether the given user command is clear, ambiguous, or infeasible in the context of interactive robotic agents utilizing large language models (LLMs). To tackle this problem, we first present an uncertainty estimation method for LLMs to classify whether the command is certain (i.e., clear) or not (i.e., ambiguous or infeasible). Once the command is classified as uncertain, we further distinguish it between ambiguous or infeasible commands leveraging LLMs with situational aware context in a zero-shot manner. For ambiguous commands, we disambiguate the command by interacting with users via question generation with LLMs. We believe that proper recognition of the given commands could lead to a decrease in malfunction and undesired actions of the robot, enhancing the reliability of interactive robot agents. We present a dataset for robotic situational awareness, consisting pair of high-level commands, scene descriptions, and labels of command type (i.e., clear, ambiguous, or infeasible). We validate the proposed method on the collected dataset, pick-and-place tabletop simulation. Finally, we demonstrate the proposed approach in real-world human-robot interaction experiments, i.e., handover scenarios.

cs.RO↗

arXiv2024

Visual Preference Inference: An Image Sequence-Based Preference Reasoning in Tabletop Object Manipulation

In robotic object manipulation, human preferences can often be influenced by the visual attributes of objects, such as color and shape. These properties play a crucial role in operating a robot to interact with objects and align with human intention. In this paper, we focus on the problem of inferring underlying human preferences from a sequence of raw visual observations in tabletop manipulation environments with a variety of object types, named Visual Preference Inference (VPI). To facilitate visual reasoning in the context of manipulation, we introduce the Chain-of-Visual-Residuals (CoVR) method. CoVR employs a prompting mechanism that describes the difference between the consecutive images (i.e., visual residuals) and incorporates such texts with a sequence of images to infer the user's preference. This approach significantly enhances the ability to understand and adapt to dynamic changes in its visual environment during manipulation tasks. Furthermore, we incorporate such texts along with a sequence of images to infer the user's preferences. Our method outperforms baseline methods in terms of extracting human preferences from visual sequences in both simulation and real-world environments. Code and videos are available at: \href{https://joonhyung-lee.github.io/vpi/}{https://joonhyung-lee.github.io/vpi/}

cs.RO↗

arXiv2023

SPOTS: Stable Placement of Objects with Reasoning in Semi-Autonomous Teleoperation Systems

Pick-and-place is one of the fundamental tasks in robotics research. However, the attention has been mostly focused on the ``pick'' task, leaving the ``place'' task relatively unexplored. In this paper, we address the problem of placing objects in the context of a teleoperation framework. Particularly, we focus on two aspects of the place task: stability robustness and contextual reasonableness of object placements. Our proposed method combines simulation-driven physical stability verification via real-to-sim and the semantic reasoning capability of large language models. In other words, given place context information (e.g., user preferences, object to place, and current scene information), our proposed method outputs a probability distribution over the possible placement candidates, considering the robustness and reasonableness of the place task. Our proposed method is extensively evaluated in two simulation and one real world environments and we show that our method can greatly increase the physical plausibility of the placement as well as contextual soundness while considering user preferences.

cs.RO↗

arXiv2022

Patch-wise Deep Metric Learning for Unsupervised Low-Dose CT Denoising

The acquisition conditions for low-dose and high-dose CT images are usually different, so that the shifts in the CT numbers often occur. Accordingly, unsupervised deep learning-based approaches, which learn the target image distribution, often introduce CT number distortions and result in detrimental effects in diagnostic performance. To address this, here we propose a novel unsupervised learning approach for lowdose CT reconstruction using patch-wise deep metric learning. The key idea is to learn embedding space by pulling the positive pairs of image patches which shares the same anatomical structure, and pushing the negative pairs which have same noise level each other. Thereby, the network is trained to suppress the noise level, while retaining the original global CT number distributions even after the image translation. Experimental results confirm that our deep metric learning plays a critical role in producing high quality denoised images without CT number shift.

eess.IV↗