三星电子

Clustered semiconductor fabrication (CFAB) has recently been introduced in South Korea, connecting multiple FABs to meet the high demands of large-scale integration chip products. To make these FABs connected effectively, CFAB utilizes various types of interface modes (IMs) such as lifters and conveyors to transfer wafers between FABs. However, due to the operational algorithm limitations of IMs and the frequent wafer transfers between FABs to meet production requirements, traffic congestion among IMs often occurs unexpectedly. These instances of congestion can cause unpredictable traffic conditions in other IMs, making it more challenging to manage traffic in a CFAB. In this paper, we propose three analysis methods based on data mining and graph network analysis approaches to explore reciprocal relationships between IMs regarding traffic congestion and describe the reasons behind traffic congestion in each of them. To demonstrate the effectiveness of the proposed method, we conduct numerical experiments using a dataset collected from the real-world CFAB in South Korea. The findings through experiment results suggest that most traffic congestion occurs in only a few specific IMs in the CFAB, indicating that managing these IMs can reduce overall traffic congestion. Moreover, we provide visual insights into managerial implications to support decision-making for enhancing IM operation performance and mitigating traffic congestion bottlenecks in complex CFAB environments.

We present a high-precision measurement procedure for lens mold asphericity with high dynamic range, incorporating a novel calibration strategy based on radial shearing interferometry (RSI). The RSI system was first modeled using ZEMAX® ray-tracing software with a bulk-optics configuration and experimentally validated. A numerical approach, referred to as iterative reverse optimization, was introduced to adapt the initial lens design model to measured wavefront data. In this study, RSI was employed to rapidly minimize measurement departures, enabling calibration with a spherical target using multiple axial displacements at non-confocal positions combined with reverse optimization. The ZEMAX® model was optimized to simultaneously fit all non-confocal experimental wavefront departures, thereby revealing complete system parameters, including assembly deviations and misalignments. The calibrated system was subsequently applied to measure the asphericity of a lens mold, with the aspheric surface parameters retrieved through the proposed reverse optimization procedure. The extracted asphericity showed excellent agreement with independent measurements performed by the Form Talysurf contact method. This approach demonstrates a flexible and practical route for measuring asphericities of varying degrees in non-null testing conditions.

Recently, TabPFN has gained attention as a foundation model for tabular data. However, it struggles to integrate heterogeneous modalities such as images and text, which are common in domains like healthcare and marketing, thereby limiting its applicability. To address this, we present the Multi-Modal Prior-data Fitted Network (MMPFN), which extends TabPFN to handle tabular and non-tabular modalities in a unified manner. MMPFN comprises per-modality encoders, modality projectors, and pre-trained foundation models. The modality projectors serve as the critical bridge, transforming non-tabular embeddings into tabular-compatible tokens for unified processing. To this end, we introduce a multi-head gated MLP and a cross-attention sampler that extract richer context from non-tabular inputs while mitigates attention imbalance issue in multimodal learning. Extensive experiments on medical and general-purpose multimodal datasets demonstrate that MMPFN consistently outperforms competitive state-of-the-art methods and effectively exploits non-tabular modalities alongside tabular features. These results highlight the promise of extending prior-data fitted networks to the multimodal setting, offering a scalable and effective framework for heterogeneous data learning.

Brushing is a common interaction technique in 2D scatterplots, allowing users to select clustered points within a continuous, enclosed region for further analysis or filtering. However, applying conventional brushing to 2D representations of multidimensional (MD) data, i.e., Multidimensional Projections (MDPs), can lead to unreliable cluster analysis due to MDP-induced distortions that inaccurately represent the cluster structure of the original MD data. To alleviate this problem, we introduce a novel brushing technique for MDPs called Distortion-aware brushing. As users perform brushing, Distortion-aware brushing correct distortions around the currently brushed points by dynamically relocating points in the projection, pulling data points close to the brushed points in MD space while pushing distant ones apart. This dynamic adjustment helps users brush MD clusters more accurately, leading to more reliable cluster analysis. Our user studies with 24 participants show that Distortion-aware brushing significantly outperforms previous brushing techniques for MDPs in accurately separating clusters in the MD space and remains robust against distortions. We further demonstrate the effectiveness of our technique through two use cases: (1) conducting cluster analysis of geospatial data and (2) interactively labeling MD clusters.

Wearable foundation models have the potential to transform digital health by learning transferable representations from large-scale biosignals collected in everyday settings. While recent progress has been made in large-scale pretraining, most approaches overlook the spectral structure of photoplethysmography (PPG) signals, wherein physiological rhythms unfold across multiple frequency bands. Motivated by the insight that many downstream health-related tasks depend on multi-resolution features spanning fine-grained waveform morphology to global rhythmic dynamics, we introduce Masked Multiscale Reconstruction (MMR) for PPG representation learning -- a self-supervised pretraining framework that explicitly learns from hierarchical time–frequency scales of PPG data. The pretraining task is designed to reconstruct randomly masked out coefficients obtained from a wavelet-based multiresolution decomposition of PPG signals, forcing the transformer encoder to integrate information across temporal and spectral scales. We pretrain our model with MMR using ~17 million unlabeled 10-second PPG segments collected from over ~32000 smartwatch users largely in naturalistic field settings, ensuring high variability and ecological validity. On 11 of 13 diverse health-related tasks, MMR trained on large-scale wearable PPG data outperforms or matches state-of-the-art open-source PPG foundation models, time-series foundation models and other self-supervised baselines. Extensive analysis of our learned embeddings and systematic ablations underscore the value of wavelet-based representations, showing that they capture robust and physiologically-grounded features. Together, these results highlight the potential of MMR as a step toward generalizable PPG foundation models.