arXiv subjects

Hyeonsoo Lee

Publications and source records attributed to Hyeonsoo Lee.

8 recordsLinked to original sources

arXiv2026

Hayden--Preskill recovery at finite temperature on a quantum processor: dynamics and initial state from the SYK model

In the original Hayden--Preskill recovery, the post-injection scrambler and initial state are {\it not related}. We extend this setup in two ways: by using a SWAP gate so that the scrambler and initial state are {\it related}, and by considering recovery at {\it finite} temperature. For this modified protocol, we show that the information is successfully recovered in the sense that the postselection probability is non-negligible and the conditional fidelity is large. We find that both the postselection probability and the conditional fidelity are proportional to temperature, reflecting the reduced entanglement of the initial state at lower temperatures. We also derive their late-time analytic estimates under the assumption of uniform operator spreading and show that they agree well with the numerical results. This demonstrates that strong scrambling is important for successful information recovery. Implementing the protocol on an IBM superconducting processor using a binary sparse SYK Hamiltonian with $N = 8$ Majoranas, we observe that the data retain the qualitative recovery dynamics and that a SWAP-based error-mitigation scheme improves both the postselection probability and the conditional fidelity.

hep-th↗

arXiv2026

Quantum simulation of traversable-wormhole-inspired quantum teleportation in a chaotic binary sparse SYK model

We report the experimental observation of holographically motivated quantum teleportation on a quantum processor, driven by the highly entangled, chaotic dynamics of a many-body system. Specifically, we implement the traversable-wormhole (TW) protocol utilizing a \textit{chaotic} binary sparse $N = 8$ Sachdev--Ye--Kitaev (SYK) model. This optimized approach dramatically reduces circuit depth for noisy intermediate-scale quantum (NISQ) hardware while rigorously preserving the spectral chaos required for gravitational duality. Diagnosing the teleportation signal via mutual information, we find that while inherent noise in NISQ hardware precludes perfect quantitative agreement with exact numerical simulations, our experimental results clearly demonstrate the essential qualitative signature: a sign-dependent asymmetry. This work establishes a practical, scalable framework for holographic quantum simulations, offering a novel empirical testbed for exploring holographic quantum gravity.

hep-th↗

arXiv2026

DFLOP: A Data-driven Framework for Multimodal LLM Training Pipeline Optimization

Multimodal Large Language Models (MLLMs) have achieved remarkable advances by integrating text, image, and audio understanding within a unified architecture. However, existing distributed training frameworks remain fundamentally data-blind: they parallelize computation without accounting for variations in input data characteristics. This data unawareness leads to severe computation skew across stages and microbatches, where heterogeneous multimodal inputs incur different processing costs. Consequently, GPU resources are unevenly utilized, synchronization delays accumulate, and overall training efficiency degrades. To address this limitation, we present DFLOP, a data-driven framework for multimodal LLM training pipeline optimization. DFLOP continuously profiles runtime behavior to capture data-induced computation variance and employs predictive scheduling to balance workloads across stages and microbatches. By coupling data characteristics with execution planning, DFLOP substantially improves GPU utilization and throughput. Extensive experiments on large-scale multimodal benchmarks show that DFLOP achieves up to 3.6x faster training compared to state-of-the-art distributed training frameworks.

cs.DC↗

arXiv2025

Dual Mixture-of-Experts Framework for Discrete-Time Survival Analysis

Survival analysis is a task to model the time until an event of interest occurs, widely used in clinical and biomedical research. A key challenge is to model patient heterogeneity while also adapting risk predictions to both individual characteristics and temporal dynamics. We propose a dual mixture-of-experts (MoE) framework for discrete-time survival analysis. Our approach combines a feature-encoder MoE for subgroup-aware representation learning with a hazard MoE that leverages patient features and time embeddings to capture temporal dynamics. This dual-MoE design flexibly integrates with existing deep learning based survival pipelines. On METABRIC and GBSG breast cancer datasets, our method consistently improves performance, boosting the time-dependent C-index up to 0.04 on the test sets, and yields further gains when incorporated into the Consurv framework.

cs.LG↗

arXiv2024

SelectiveKD: A semi-supervised framework for cancer detection in DBT through Knowledge Distillation and Pseudo-labeling

When developing Computer Aided Detection (CAD) systems for Digital Breast Tomosynthesis (DBT), the complexity arising from the volumetric nature of the modality poses significant technical challenges for obtaining large-scale accurate annotations. Without access to large-scale annotations, the resulting model may not generalize to different domains. Given the costly nature of obtaining DBT annotations, how to effectively increase the amount of data used for training DBT CAD systems remains an open challenge. In this paper, we present SelectiveKD, a semi-supervised learning framework for building cancer detection models for DBT, which only requires a limited number of annotated slices to reach high performance. We achieve this by utilizing unlabeled slices available in a DBT stack through a knowledge distillation framework in which the teacher model provides a supervisory signal to the student model for all slices in the DBT volume. Our framework mitigates the potential noise in the supervisory signal from a sub-optimal teacher by implementing a selective dataset expansion strategy using pseudo labels. We evaluate our approach with a large-scale real-world dataset of over 10,000 DBT exams collected from multiple device manufacturers and locations. The resulting SelectiveKD process effectively utilizes unannotated slices from a DBT stack, leading to significantly improved cancer classification performance (AUC) and generalization performance.

cs.CV↗

arXiv2023

Enhancing Breast Cancer Risk Prediction by Incorporating Prior Images

Recently, deep learning models have shown the potential to predict breast cancer risk and enable targeted screening strategies, but current models do not consider the change in the breast over time. In this paper, we present a new method, PRIME+, for breast cancer risk prediction that leverages prior mammograms using a transformer decoder, outperforming a state-of-the-art risk prediction method that only uses mammograms from a single time point. We validate our approach on a dataset with 16,113 exams and further demonstrate that it effectively captures patterns of changes from prior mammograms, such as changes in breast density, resulting in improved short-term and long-term breast cancer risk prediction. Experimental results show that our model achieves a statistically significant improvement in performance over the state-of-the-art based model, with a C-index increase from 0.68 to 0.73 (p < 0.05) on held-out test sets.

cs.CV↗

arXiv2023

Bayesian Optimization Meets Self-Distillation

Bayesian optimization (BO) has contributed greatly to improving model performance by suggesting promising hyperparameter configurations iteratively based on observations from multiple training trials. However, only partial knowledge (i.e., the measured performances of trained models and their hyperparameter configurations) from previous trials is transferred. On the other hand, Self-Distillation (SD) only transfers partial knowledge learned by the task model itself. To fully leverage the various knowledge gained from all training trials, we propose the BOSS framework, which combines BO and SD. BOSS suggests promising hyperparameter configurations through BO and carefully selects pre-trained models from previous trials for SD, which are otherwise abandoned in the conventional BO process. BOSS achieves significantly better performance than both BO and SD in a wide range of tasks including general image classification, learning with noisy labels, semi-supervised learning, and medical image analysis tasks.

cs.CV↗

arXiv2020

Scribble2Label: Scribble-Supervised Cell Segmentation via Self-Generating Pseudo-Labels with Consistency

Segmentation is a fundamental process in microscopic cell image analysis. With the advent of recent advances in deep learning, more accurate and high-throughput cell segmentation has become feasible. However, most existing deep learning-based cell segmentation algorithms require fully annotated ground-truth cell labels, which are time-consuming and labor-intensive to generate. In this paper, we introduce Scribble2Label, a novel weakly-supervised cell segmentation framework that exploits only a handful of scribble annotations without full segmentation labels. The core idea is to combine pseudo-labeling and label filtering to generate reliable labels from weak supervision. For this, we leverage the consistency of predictions by iteratively averaging the predictions to improve pseudo labels. We demonstrate the performance of Scribble2Label by comparing it to several state-of-the-art cell segmentation methods with various cell image modalities, including bright-field, fluorescence, and electron microscopy. We also show that our method performs robustly across different levels of scribble details, which confirms that only a few scribble annotations are required in real-use cases.

cs.CV↗