arXiv subjects

Jiwon Jeong

Publications and source records attributed to Jiwon Jeong.

17 recordsLinked to original sources

arXiv2026

RaLMPH: Reliability-aware Learning for Multi-Pathologist Harmonization in Whole-Slide Image Classification

Multiple Instance Learning (MIL) is a standard paradigm for Whole-Slide Image (WSI) analysis and has achieved strong results in computational pathology. However, most MIL pipelines assume a single "gold" label per slide, which conflicts with clinical practice where substantial inter-pathologist variability is common. Existing multi-annotator learning and label-refinement methods typically estimate global annotator reliability or rely on single-instance assumptions, making them poorly suited to MIL and to localized diagnostic contexts where experts disagree. We propose RaLMPH (Reliability-aware Learning for Multi-Pathologist Harmonization), a MIL-based label reconciliation framework for WSIs annotated by multiple pathologists. RaLMPH introduces a reliability field that jointly models (i) local neighborhood structure in WSI feature space and (ii) expert uncertainty (entropy), enabling per-sample identification of trustworthy reference neighborhoods. Leveraging this field, RaLMPH performs sample-wise local annotator ranking to select reliable opinions per slide and applies an adaptive gating mechanism to fuse labels conditioned on local reliability. Experiments on a clinical WSI dataset with labels from six pathologists, as well as controlled simulated benchmarks, show that RaLMPH consistently outperforms existing approaches. Further analyses clarify how our reliability-aware mechanism improves label reconciliation and downstream MIL performance.

cs.CV↗

arXiv2026

One Sequential Recommendation Model Pretrained from Synthetic Priors Predicts Multiple Datasets

Existing sequential recommendation models rely on dataset-specific training, where the learned parameters are fitted to the item catalog and the observed interaction distribution of the training data. This limits generalization to new domains, typically requiring retraining from scratch. In this work, we propose SRPFN, a Prior-data Fitted Network for sequential recommendation -- predicting the next item in a single forward pass without any gradient-based parameter updates in the target domain. SRPFN is pretrained offline on 25.6M sequences sampled from a synthetic prior that spans diverse item-to-item transition patterns, learning to produce posterior predictive next-item distributions. At inference time, SRPFN generates recommendations by conditioning on a support set of item-item transition examples from the target domain, adapting to domain-specific patterns without retraining. Extensive experiments on five benchmarks across 10 baselines show that SRPFN achieves the best or second-best performance across nearly all metrics and datasets, while being substantially more computationally efficient than trained baselines. These results establish that a single model pretrained on synthetic priors can generalize across diverse real-world domains, offering a framework for update-free sequential recommendation.

cs.IR↗

arXiv2026

Every Preference Has Its Strength: Injecting Ordinal Semantics into LLM-Based Recommenders

Recent work has shown that large language models (LLMs) can enhance recommender systems by integrating collaborative filtering (CF) signals through hybrid prompting. However, most existing CF-LLM frameworks collapse explicit ratings into implicit or positive-only feedback, discarding the ordinal structure that conveys fine-grained preference strength. As a result, these models struggle to exploit graded semantics and nuanced preference distinctions. We propose Ordinal Semantic Anchoring (OSA), a hybrid CF-LLM framework that explicitly incorporates preference strength by modeling interaction-level user feedback. OSA represents ordinal preference levels as numeric textual tokens and uses their token embeddings as semantic anchors to align user-item interaction representations in the LLM latent space. Through strength-aware alignment across ordinal levels, OSA preserves preference semantics when integrating collaborative signals with LLMs. Experiments on multiple real-world datasets demonstrate that OSA consistently outperforms existing baselines, particularly in pairwise preference evaluation, highlighting its effectiveness in modeling fine-grained user preferences over prior CF-LLM methods.

cs.IR↗

arXiv2026

Scaling the Queue: Reinforcement Learning for Equitable Call Classification Capacity in NYC Municipal Complaint Systems

Municipal 311 call centers and complaint intake systems face a structural mismatch between incoming volume and classification capacity. The staff and heuristics available to triage, route, and prioritize complaints cannot scale with demand. This bottleneck produces differential service quality that follows income and racial lines (\cite{liu2024sla}). We develop an equity-centered reinforcement learning (RL) framework that augments call classification capacity across six New York City Department of Buildings (DOB) operational domains: boiler safety, crane and derrick oversight, heat and hot water complaints, housing complaint triage, scaffold safety, and Natural Area District (SNAD) protection. Rather than replacing human classifiers, our agents act as intelligent intake routers: learning to assign incoming complaints to action categories: escalate, batch, defer, inspect now. The proposed technique is designed to maximize throughput, minimize misclassification cost, and actively narrow historical equity gaps in service delivery. We formalize each domain as a Markov Decision Process (MDP) in which equitable classification coverage is a first-class reward objective. Post-hoc SHAP attribution reveals that complaint recurrence and neighborhood-level statistics are stronger predictors of actionable violations than raw complaint volume. This finding has direct implications for complaint routing given the demographic correlates of those features.

econ.EM↗

arXiv2026

MechELK: A Mechanistic Interpretability Framework for Eliciting Latent Knowledge in Large Language Models

Large language models (LLMs) frequently encode factual and reasoning knowledge in their internal representations that is not faithfully reflected in their surface-level outputs -- a phenomenon known as \emph{latent knowledge}. Existing approaches to eliciting latent knowledge, such as Contrastive Consistency Search (CCS), rely on contrastive activation patterns and struggle with complex multi-step reasoning tasks, while mechanistic interpretability tools have primarily been used to \emph{understand} model behavior rather than to \emph{extract} hidden knowledge. We present \textbf{MechELK}, a unified three-stage framework that bridges mechanistic interpretability and latent knowledge elicitation. MechELK operates through: (1) \textbf{Locate} -- using Sparse Autoencoder (SAE) feature analysis and activation patching to identify knowledge-bearing representations; (2) \textbf{Verify} -- employing causal probing to distinguish genuine latent knowledge from spurious correlations; and (3) \textbf{Elicit} -- applying representation engineering to surface hidden knowledge without modifying model weights. Evaluated on TruthfulQA, a curated Deceptive Alignment benchmark, and the Quirky LM dataset, MechELK achieves an average elicitation accuracy of 84.7\%, outperforming CCS by 6.2\% and direct linear probing by 9.1\%. Crucially, MechELK successfully identifies latent knowledge in 78.3\% of cases where the model's surface output is incorrect or evasive, demonstrating its utility for AI safety applications including deceptive alignment detection.

cs.CL↗

arXiv2026

Every Error has Its Magnitude: Asymmetric Mistake Severity Training for Multiclass Multiple Instance Learning

Multiple Instance Learning (MIL) has emerged as a promising paradigm for Whole Slide Image (WSI) diagnosis, offering effective learning with limited annotations. However, existing MIL frameworks overlook diagnostic priorities and fail to differentiate the severity of misclassifications in multiclass, leaving clinically critical errors unaddressed. We propose a mistake-severity-aware training strategy that organizes diagnostic classes into a hierarchical structure, with each level optimized using a severity-weighted cross-entropy loss that penalizes high-severity misclassifications more strongly. Additionally, hierarchical consistency is enforced through probabilistic alignment, a semantic feature remix applied to the instance bag to robustly train class priority and accommodate clinical cases involving multiple symptoms. An asymmetric Mikel's Wheel-based metric is also introduced to quantify the severity of errors specific to medical fields. Experiments on challenging public and real-world in-house datasets demonstrate that our approach significantly mitigates critical errors in MIL diagnosis compared to existing methods. We present additional experimental results on natural domain data to demonstrate the generalizability of our proposed method beyond medical contexts.

cs.CV↗

arXiv2025

Diagnose Like A REAL Pathologist: An Uncertainty-Focused Approach for Trustworthy Multi-Resolution Multiple Instance Learning

With the increasing demand for histopathological specimen examination and diagnostic reporting, Multiple Instance Learning (MIL) has received heightened research focus as a viable solution for AI-centric diagnostic aid. Recently, to improve its performance and make it work more like a pathologist, several MIL approaches based on the use of multiple-resolution images have been proposed, delivering often higher performance than those that use single-resolution images. Despite impressive recent developments of multiple-resolution MIL, previous approaches only focus on improving performance, thereby lacking research on well-calibrated MIL that clinical experts can rely on for trustworthy diagnostic results. In this study, we propose Uncertainty-Focused Calibrated MIL (UFC-MIL), which more closely mimics the pathologists' examination behaviors while providing calibrated diagnostic predictions, using multiple images with different resolutions. UFC-MIL includes a novel patch-wise loss that learns the latent patterns of instances and expresses their uncertainty for classification. Also, the attention-based architecture with a neighbor patch aggregation module collects features for the classifier. In addition, aggregated predictions are calibrated through patch-level uncertainty without requiring multiple iterative inferences, which is a key practical advantage. Against challenging public datasets, UFC-MIL shows superior performance in model calibration while achieving classification accuracy comparable to that of state-of-the-art methods.

cs.CV↗

arXiv2025

Majority Vote Compressed Sensing

We consider the problem of non-coherent over-the-air computation (AirComp), where $n$ devices carry high-dimensional data vectors $\mathbf{x}_i\in\mathbb{R}^d$ of sparsity $\lVert\mathbf{x}_i\rVert_0\leq k$ whose sum has to be computed at a receiver. Previous results on non-coherent AirComp require more than $d$ channel uses to compute functions of $\mathbf{x}_i$, where the extra redundancy is used to combat non-coherent signal aggregation. However, if the data vectors are sparse, sparsity can be exploited to offer significantly cheaper communication. In this paper, we propose to use random transforms to transmit lower-dimensional projections $\mathbf{s}_i\in\mathbb{R}^T$ of the data vectors. These projected vectors are communicated to the receiver using a majority vote (MV)-AirComp scheme, which estimates the bit-vector corresponding to the signs of the aggregated projections, i.e., $\mathbf{y} = \text{sign}(\sum_i\mathbf{s}_i)$. By leveraging 1-bit compressed sensing (1bCS) at the receiver, the real-valued and high-dimensional aggregate $\sum_i\mathbf{x}_i$ can be recovered from $\mathbf{y}$. We prove analytically that the proposed MVCS scheme estimates the aggregated data vector $\sum_i \mathbf{x}_i$ with $\ell_2$-norm error $ε$ in $T=\mathcal{O}(kn\log(d)/ε^2)$ channel uses. Moreover, we specify algorithms that leverage MVCS for histogram estimation and distributed machine learning. Finally, we provide numerical evaluations that reveal the advantage of MVCS compared to the state-of-the-art.

eess.SP↗

arXiv2025

Information-computation trade-offs in non-linear transforms

In this work, we explore the interplay between information and computation in non-linear transform-based compression for broad classes of modern information-processing tasks. We first investigate two emerging nonlinear data transformation frameworks for image compression: Implicit Neural Representations (INRs) and 2D Gaussian Splatting (GS). We analyze their representational properties, behavior under lossy compression, and convergence dynamics. Our results highlight key trade-offs between INR's compact, resolution-flexible neural field representations and GS's highly parallelizable, spatially interpretable fitting, providing insights for future hybrid and compression-aware frameworks. Next, we introduce the textual transform that enables efficient compression at ultra-low bitrate regimes and simultaneously enhances human perceptual satisfaction. When combined with the concept of denoising via lossy compression, the textual transform becomes a powerful tool for denoising tasks. Finally, we present a Lempel-Ziv (LZ78) "transform", a universal method that, when applied to any member of a broad compressor family, produces new compressors that retain the asymptotic universality guarantees of the LZ78 algorithm. Collectively, these three transforms illuminate the fundamental trade-offs between coding efficiency and computational cost. We discuss how these insights extend beyond compression to tasks such as classification, denoising, and generative AI, suggesting new pathways for using non-linear transformations to balance resource constraints and performance.

cs.IT↗

arXiv2025

Large Language Models Are Better Logical Fallacy Reasoners with Counterargument, Explanation, and Goal-Aware Prompt Formulation

The advancement of Large Language Models (LLMs) has greatly improved our ability to process complex language. However, accurately detecting logical fallacies remains a significant challenge. This study presents a novel and effective prompt formulation approach for logical fallacy detection, applicable in both supervised (fine-tuned) and unsupervised (zero-shot) settings. Our method enriches input text incorporating implicit contextual information -- counterarguments, explanations, and goals -- which we query for validity within the context of the argument. We then rank these queries based on confidence scores to inform classification. We evaluate our approach across multiple datasets from 5 domains, covering 29 distinct fallacy types, using models from the GPT and LLaMA series. The results show substantial improvements over state-of-the-art models, with F1 score increases of up to 0.60 in zero-shot settings and up to 0.45 in fine-tuned models. Extensive analyses further illustrate why and how our method excels.

cs.AI↗

arXiv2025

Universal Discrete Filtering with Lookahead or Delay

We consider the universal discrete filtering problem, where an input sequence generated by an unknown source passes through a discrete memoryless channel, and the goal is to estimate its components based on the output sequence with limited lookahead or delay. We propose and establish the universality of a family of schemes for this setting. These schemes are induced by universal Sequential Probability Assignments (SPAs), and inherit their computational properties. We show that the schemes induced by LZ78 are practically implementable and well-suited for scenarios with limited computational resources and latency constraints. In passing, we use some of the intermediate results to obtain upper and lower bounds that appear to be new, in the purely Bayesian setting, on the optimal filtering performance in terms, respectively, of the mutual information between the noise-free and noisy sequence, and the entropy of the noise-free sequence causally conditioned on the noisy one.

eess.SP↗

arXiv2024

Fully Geometric Panoramic Localization

We introduce a lightweight and accurate localization method that only utilizes the geometry of 2D-3D lines. Given a pre-captured 3D map, our approach localizes a panorama image, taking advantage of the holistic 360 view. The system mitigates potential privacy breaches or domain discrepancies by avoiding trained or hand-crafted visual descriptors. However, as lines alone can be ambiguous, we express distinctive yet compact spatial contexts from relationships between lines, namely the dominant directions of parallel lines and the intersection between non-parallel lines. The resulting representations are efficient in processing time and memory compared to conventional visual descriptor-based methods. Given the groups of dominant line directions and their intersections, we accelerate the search process to test thousands of pose candidates in less than a millisecond without sacrificing accuracy. We empirically show that the proposed 2D-3D matching can localize panoramas for challenging scenes with similar structures, dramatic domain shifts or illumination changes. Our fully geometric approach does not involve extensive parameter tuning or neural network training, making it a practical algorithm that can be readily deployed in the real world. Project page including the code is available through this link: https://82magnolia.github.io/fgpl/.

cs.CV↗

arXiv2022

Controlled doping of electrocatalysts through engineering impurities

Fuel cells recombine water from H2 and O2 thereby powering e.g. cars or houses with no direct carbon emission. In anion-exchange membrane fuel cells (AEMFCs), to reach high power densities, operating at high pH is an alternative to using large volumes of noble metals catalysts at the cathode, where the oxygen-reduction reaction occurs. However, the sluggish kinetics of the hydrogen-oxidation reaction (HOR) hinders upscaling despite promising catalysts. Here, we observe an unexpected ingress of B into Pd nano-catalysts synthesised by wet-chemistry, gain control over this B-doping, and report on its influence on the HOR activity in alkaline conditions. We rationalize our findings using ab-initio calculations of both H- and OH-adsorption on B-doped Pd. Using this "impurity engineering" approach, we thus design Pt-free catalysts as required in electrochemical energy conversion devices, e.g. next generations of AEMFCs, that satisfy the economic and environmental constraints, i.e. reasonable operating costs and long-term stability, to enable the hydrogen economy.

cond-mat.mtrl-sci↗

arXiv2021

Free, flexible and fast: orientation mapping using the multi-core and GPU-accelerated template matching capabilities in the python-based open source 4D-STEM analysis toolbox Pyxem

This work presents the new template matching capabilities implemented in Pyxem, an open source Python library for analyzing four-dimensional scanning transmission electron microscopy (4D-STEM) data. Template matching is a brute force approach for deriving local crystal orientations. It works by comparing a library of simulated diffraction patterns to experimental patterns collected with nano-beam and precession electron diffraction (NBED and PED). This is a computationally demanding task, therefore the implementation combines efficiency and scalability by utilizing multiple CPU cores or a graphical processing unit (GPU). The code is built on top of the scientific python ecosystem, and is designed to support custom and reproducible workflows that combine the image processing, template library generation, indexation and visualisation all in one environment. The tools are agnostic to file size and format, which is significant in light of the increased adoption of pixelated detectors from different manufacturers. This paper details the implementation, validation, and benchmarking results of the method. The method is illustrated by calculating orientation maps of nanocrystalline materials and precipitates embedded in a crystalline matrix. The combination of speed and flexibility opens the door for automated parameter studies and real-time on-line orientation mapping inside the TEM.

cond-mat.mtrl-sci↗

arXiv2021

Understanding alkali contamination in colloidal nanomaterials to unlock grain boundary impurity engineering

Metal nano-aerogels combine a large surface area, a high structural stability, and a high catalytic activity towards a variety of chemical reactions. The performance of such nanostructures is underpinned by the atomic-level distribution of their constituents. Yet monitoring their sub-nanoscale structure and composition to guide property optimization remains extremely challenging. Here, we synthesized Pd nano-aerogels from a K2PdCl4 precursor and two different NaBH4 reductant concentrations in distilled water. Atom probe tomography reveals that the aerogel is poly-crystalline and that impurities (Na, K) are integrated from the solution into grain boundaries. Ab initio calculations indicate that these impurities preferentially bound to the Pd-metal surface and are ultimately found in grain boundaries forming as the particles coalesce during synthesis, with Na atoms thermodynamically equilibrating with the surrounding solution and K atoms remaining between growing grains. If controlled, impurity integration, i.e. grain boundary decoration, may offer opportunities for designing new nano-aerogels.

cond-mat.mtrl-sci↗

arXiv2021

Investigation of the orientation relationship between nano-sized G-phase precipitates and austenite with scanning nano-beam electron diffraction using a pixelated detector

Scanning nano-beam electron diffraction with a pixelated detector was employed to investigate the orientation relationship of nanometer sized, irradiation induced G-phase (M$_6$Ni$_{16}$Si$_{7}$) precipitates in an austenite matrix. Using this detector, the faint diffraction spots originating from the small G-phase particles could be resolved simultaneously as the intense matrix reflections. The diffraction patterns were analyzed using a two-stage template matching scheme, whereby the matrix is indexed first and the precipitates are indexed second after subtraction of the matrix contribution to the diffraction patterns. The results show that G-phase forms with orientation relationships relative to austenite that are characteristic of face-centered cubic (FCC) to body-centered cubic (BCC) transformations. This work demonstrates that nano-beam electron diffraction with a pixelated detector is a promising technique to investigate orientation relationships of nano-sized precipitates with complex crystal structures in other material systems with relative ease.

cond-mat.mtrl-sci↗

arXiv2021

Automated crystal orientation mapping by precession electron diffraction assisted four-dimensional scanning transmission electron microscopy (4D-STEM) using a scintillator based CMOS detector

The recent development of electron sensitive and pixelated detectors has attracted the use of four-dimensional scanning transmission electron microscopy (4D-STEM). Here, we present a precession electron diffraction assisted 4D-STEM technique for automated orientation mapping using diffraction spot patterns directly captured by an in-column scintillator based complementary metal-oxide-semiconductor (CMOS) detector. We compare the results to a conventional approach, which utilizes a fluorescent screen filmed by an external CCD camera. The high dynamic range and signal-to-noise characteristics of the detector largely improve the image quality of the diffraction patterns, especially the visibility of diffraction spots at high scattering angles. In the orientation maps reconstructed via the template matching process, the CMOS data yields a significant reduction of false indexing and higher reliability compared to the conventional approach. The angular resolution of misorientation measurement could also be improved by masking reflections close to the direct beam. This is because the orientation sensitive, weak and small diffraction spots at high scattering angle are more significant. The results show that fine details such as nanograins, nanotwins and sub-grain boundaries can be resolved with a sub-degree angular resolution which is comparable to orientation mapping using Kikuchi diffraction patterns.

physics.ins-det↗