Bum Chul Kwon

Bum Chul Kwon, Jungsoo Lee, Chaeyeon Chung et al. · ibm-research

Image classification models often learn to predict a class based on irrelevant co-occurrences between input features and an output class in training data. We call the unwanted correlations "data biases," and the visual features causing data biases "bias factors." It is challenging to identify and mitigate biases automatically without human intervention. Therefore, we conducted a design study to find a human-in-the-loop solution. First, we identified user tasks that capture the bias mitigation process for image classification models with three experts. Then, to support the tasks, we developed a visual analytics system called DASH that allows users to visually identify bias factors, to iteratively generate synthetic images using a state-of-the-art image-to-image translation model, and to supervise the model training process for improving the classification accuracy. Our quantitative evaluation and qualitative study with ten participants demonstrate the usefulness of DASH and provide lessons for future work.

Hongyang Li, Sanjoy Dey, Bum Chul Kwon et al.

Large language models (LLMs) trained on text demonstrated remarkable results on natural language processing (NLP) tasks. These models have been adapted to decipher the language of DNA, where sequences of nucleotides act as "words" that encode genomic functions. However, the genome differs fundamentally from natural language, as it lacks clearly defined words or a consistent grammar. Although DNA language models (DNALMs) such as DNABERT, GENA-LM have achieved high level of performance on genome-related biological tasks, these models do not encode biological functions in the presence of sequence variations. To address this problem, we pre-train foundation models that effectively integrate sequence variations, in particular Single Nucleotide Polymorphisms (SNPs), as they underlie important biological functions. Specifically, we use ModernBERT to pre-train two different Biomedical Foundation Models (BMFM), namely, BMFM-DNA-REF in which the model is trained with sequences of varying lengths along with their reverse complements derived from the reference genome and BMFM-DNA-SNP in which the model is trained with sequences created using a novel representation scheme that encodes sequence variations. Our findings indicate that integrating sequence variations into DNALMs helps capture the biological functions as seen in improvements on all fine-tuning tasks. To explore the model's practical utility, we experimented with various strategies for SNP imputation on promoter detection task introduced in DNABERT-2. However, we acknowledge that the current benchmarks are limited in their ability to fully evaluate these models. To enable more comprehensive assessment in the future and encourage community contributions, we release our models through HuggingFace and the code to reproduce the results at https://github.com/BiomedSciAI/biomed-multi-omic

Bum Chul Kwon, Nandana Mihindukulasooriya

Pre-trained transformer-based language models are becoming increasingly popular due to their exceptional performance on various benchmarks. However, concerns persist regarding the presence of hidden biases within these models, which can lead to discriminatory outcomes and reinforce harmful stereotypes. To address this issue, we propose Finspector, a human-centered visual inspection tool designed to detect biases in different categories through log-likelihood scores generated by language models. The goal of the tool is to enable researchers to easily identify potential biases using visual analytics, ultimately contributing to a fairer and more just deployment of these models in both academic and industrial settings. Finspector is available at https://github.com/IBM/finspector.

Bum Chul Kwon, Ben Shapira, Moshiko Raboh et al.

The chemical space of drug-like molecules is vast, motivating the development of generative models that must learn broad chemical distributions, enable conditional generation by capturing structure-property representations, and provide fast molecular generation. Meeting the objectives depends on modeling choices, including the probabilistic modeling approach, the conditional generative formulation, the architecture, and the molecular input representation. To address the challenges, we present STAR-VAE (Selfies-encoded, Transformer-based, AutoRegressive Variational Auto Encoder), a scalable latent-variable framework with a Transformer encoder and an autoregressive Transformer decoder. It is trained on 79 million drug-like molecules from PubChem, using SELFIES to guarantee syntactic validity. The latent-variable formulation enables conditional generation: a property predictor supplies a conditioning signal that is applied consistently to the latent prior, the inference network, and the decoder. Our contributions are: (i) a Transformer-based latent-variable encoder-decoder model trained on SELFIES representations; (ii) a principled conditional latent-variable formulation for property-guided generation; and (iii) efficient finetuning with low-rank adapters (LoRA) in both encoder and decoder, enabling fast adaptation with limited property and activity data. On the GuacaMol and MOSES benchmarks, our approach matches or exceeds baselines, and latent-space analyses reveal smooth, semantically structured representations that support both unconditional exploration and property-aware generation. On the Tartarus benchmarks, the conditional model shifts docking-score distributions toward stronger predicted binding. These results suggest that a modernized, scale-appropriate VAE remains competitive for molecular generation when paired with principled conditioning and parameter-efficient finetuning.

Jinbin Huang, Chen Chen, Aditi Mishra et al.

Generative image models have emerged as a promising technology to produce realistic images. Despite potential benefits, concerns grow about its misuse, particularly in generating deceptive images that could raise significant ethical, legal, and societal issues. Consequently, there is growing demand to empower users to effectively discern and comprehend patterns of AI-generated images. To this end, we developed ASAP, an interactive visualization system that automatically extracts distinct patterns of AI-generated images and allows users to interactively explore them via various views. To uncover fake patterns, ASAP introduces a novel image encoder, adapted from CLIP, which transforms images into compact "distilled" representations, enriched with information for differentiating authentic and fake images. These representations generate gradients that propagate back to the attention maps of CLIP's transformer block. This process quantifies the relative importance of each pixel to image authenticity or fakeness, exposing key deceptive patterns. ASAP enables the at scale interactive analysis of these patterns through multiple, coordinated visualizations. This includes a representation overview with innovative cell glyphs to aid in the exploration and qualitative evaluation of fake patterns across a vast array of images, as well as a pattern view that displays authenticity-indicating patterns in images and quantifies their impact. ASAP supports the analysis of cutting-edge generative models with the latest architectures, including GAN-based models like proGAN and diffusion models like the latent diffusion model. We demonstrate ASAP's usefulness through two usage scenarios using multiple fake image detection benchmark datasets, revealing its ability to identify and understand hidden patterns in AI-generated images, especially in detecting fake human faces produced by diffusion-based techniques.

Grace Guo, Lifu Deng, Animesh Tandon et al.

The recent prevalence of publicly accessible, large medical imaging datasets has led to a proliferation of artificial intelligence (AI) models for cardiovascular image classification and analysis. At the same time, the potentially significant impacts of these models have motivated the development of a range of explainable AI (XAI) methods that aim to explain model predictions given certain image inputs. However, many of these methods are not developed or evaluated with domain experts, and explanations are not contextualized in terms of medical expertise or domain knowledge. In this paper, we propose a novel framework and python library, MiMICRI, that provides domain-centered counterfactual explanations of cardiovascular image classification models. MiMICRI helps users interactively select and replace segments of medical images that correspond to morphological structures. From the counterfactuals generated, users can then assess the influence of each segment on model predictions, and validate the model against known medical facts. We evaluate this library with two medical experts. Our evaluation demonstrates that a domain-centered XAI approach can enhance the interpretability of model explanations, and help experts reason about models in terms of relevant domain knowledge. However, concerns were also surfaced about the clinical plausibility of the counterfactuals generated. We conclude with a discussion on the generalizability and trustworthiness of the MiMICRI framework, as well as the implications of our findings on the development of domain-centered XAI methods for model interpretability in healthcare contexts.

Parthasarathy Suryanarayanan, Yunguang Qiu, Shreyans Sethi et al. · ibm-research

Quality molecular representations are key to foundation model development in bio-medical research. Previous efforts have typically focused on a single representation or molecular view, which may have strengths or weaknesses on a given task. We develop Multi-view Molecular Embedding with Late Fusion (MMELON), an approach that integrates graph, image and text views in a foundation model setting and may be readily extended to additional representations. Single-view foundation models are each pre-trained on a dataset of up to 200M molecules. The multi-view model performs robustly, matching the performance of the highest-ranked single-view. It is validated on over 120 tasks, including molecular solubility, ADME properties, and activity against G Protein-Coupled receptors (GPCRs). We identify 33 GPCRs that are related to Alzheimer's disease and employ the multi-view model to select strong binders from a compound screen. Predictions are validated through structure-based modeling and identification of key binding motifs.

Bum Chul Kwon, Peter Achenbach, Jessica L. Dunne et al.

Analyzing disease progression patterns can provide useful insights into the disease processes of many chronic conditions. These analyses may help inform recruitment for prevention trials or the development and personalization of treatments for those affected. We learn disease progression patterns using Hidden Markov Models (HMM) and distill them into distinct trajectories using visualization methods. We apply it to the domain of Type 1 Diabetes (T1D) using large longitudinal observational data from the T1DI study group. Our method discovers distinct disease progression trajectories that corroborate with recently published findings. In this paper, we describe the iterative process of developing the model. These methods may also be applied to other chronic conditions that evolve over time.

Bum Chul Kwon

A goal of clinical researchers is to understand the progression of a disease through a set of biomarkers. Researchers often conduct observational studies, where they collect numerous samples from selected subjects throughout multiple years. Hidden Markov Models (HMMs) can be applied to discover latent states and their transition probabilities over time. However, it is challenging for clinical researchers to interpret the outcomes and to gain insights about the disease. Thus, this demo introduces an interactive visualization system called DPVis, which was designed to help researchers to interactively explore HMM outcomes. The demo provides guidelines of how to implement the clinician-in-the-loop approach for analyzing longitudinal, observational health data with visual analytics.

Chunggi Lee, Sanghoon Kim, Dongyun Han et al.

Users may face challenges while designing graphical user interfaces, due to a lack of relevant experience and guidance. This paper aims to investigate the issues that users with no experience face during the design process, and how to resolve them. To this end, we conducted semi-structured interviews, based on which we built a GUI prototyping assistance tool called GUIComp. This tool can be connected to GUI design software as an extension, and it provides real-time, multi-faceted feedback on a user's current design. Additionally, we conducted two user studies, in which we asked participants to create mobile GUIs with or without GUIComp, and requested online workers to assess the created GUIs. The experimental results show that GUIComp facilitated iterative design and the participants with GUIComp had better a user experience and produced more acceptable designs than those who did not.

Bahador Saket, Subhajit Das, Bum Chul Kwon et al.

Biologists often perform clustering analysis to derive meaningful patterns, relationships, and structures from data instances and attributes. Though clustering plays a pivotal role in biologists' data exploration, it takes non-trivial efforts for biologists to find the best grouping in their data using existing tools. Visual cluster analysis is currently performed either programmatically or through menus and dialogues in many tools, which require parameter adjustments over several steps of trial-and-error. In this paper, we introduce Geono-Cluster, a novel visual analysis tool designed to support cluster analysis for biologists who do not have formal data science training. Geono-Cluster enables biologists to apply their domain expertise into clustering results by visually demonstrating how their expected clustering outputs should look like with a small sample of data instances. The system then predicts users' intentions and generates potential clustering results. Our study follows the design study protocol to derive biologists' tasks and requirements, design the system, and evaluate the system with experts on their own dataset. Results of our study with six biologists provide initial evidence that Geono-Cluster enables biologists to create, refine, and evaluate clustering results to effectively analyze their data and gain data-driven insights. At the end, we discuss lessons learned and the implications of our study.

Cheonbok Park, Inyoup Na, Yongjang Jo et al.

Attention networks, a deep neural network architecture inspired by humans' attention mechanism, have seen significant success in image captioning, machine translation, and many other applications. Recently, they have been further evolved into an advanced approach called multi-head self-attention networks, which can encode a set of input vectors, e.g., word vectors in a sentence, into another set of vectors. Such encoding aims at simultaneously capturing diverse syntactic and semantic features within a set, each of which corresponds to a particular attention head, forming altogether multi-head attention. Meanwhile, the increased model complexity prevents users from easily understanding and manipulating the inner workings of models. To tackle the challenges, we present a visual analytics system called SANVis, which helps users understand the behaviors and the characteristics of multi-head self-attention networks. Using a state-of-the-art self-attention model called Transformer, we demonstrate usage scenarios of SANVis in machine translation tasks. Our system is available at http://short.sanvis.org

Hwiyeon Kim, Juyoung Oh, Yunha Han et al.

When people browse online news, small thumbnail images accompanying links to articles attract their attention and help them to decide which articles to read. As an increasing proportion of online news can be construed as data journalism, we have witnessed a corresponding increase in the incorporation of visualization in article thumbnails. However, there is little research to support alternative design choices for visualization thumbnails, which include resizing, cropping, simplifying, and embellishing charts appearing within the body of the associated article. We therefore sought to better understand these design choices and determine what makes a visualization thumbnail inviting and interpretable. This paper presents our findings from a survey of visualization thumbnails collected online and from conversations with data journalists and news graphics designers. Our study reveals that there exists an uncharted design space, one that is in need of further empirical study. Our work can thus be seen as a first step toward providing structured guidance on how to design thumbnails for data stories.

Bum Chul Kwon, Vibha Anand, Kristen A Severson et al.

Clinical researchers use disease progression models to understand patient status and characterize progression patterns from longitudinal health records. One approach for disease progression modeling is to describe patient status using a small number of states that represent distinctive distributions over a set of observed measures. Hidden Markov models (HMMs) and its variants are a class of models that both discover these states and make inferences of health states for patients. Despite the advantages of using the algorithms for discovering interesting patterns, it still remains challenging for medical experts to interpret model outputs, understand complex modeling parameters, and clinically make sense of the patterns. To tackle these problems, we conducted a design study with clinical scientists, statisticians, and visualization experts, with the goal to investigate disease progression pathways of chronic diseases, namely type 1 diabetes (T1D), Huntington's disease, Parkinson's disease, and chronic obstructive pulmonary disease (COPD). As a result, we introduce DPVis which seamlessly integrates model parameters and outcomes of HMMs into interpretable and interactive visualizations. In this study, we demonstrate that DPVis is successful in evaluating disease progression models, visually summarizing disease states, interactively exploring disease progression patterns, and building, analyzing, and comparing clinically relevant patient subgroups.

Bum Chul Kwon, Min-Je Choi, Joanne Taery Kim et al.

We have recently seen many successful applications of recurrent neural networks (RNNs) on electronic medical records (EMRs), which contain histories of patients' diagnoses, medications, and other various events, in order to predict the current and future states of patients. Despite the strong performance of RNNs, it is often challenging for users to understand why the model makes a particular prediction. Such black-box nature of RNNs can impede its wide adoption in clinical practice. Furthermore, we have no established methods to interactively leverage users' domain expertise and prior knowledge as inputs for steering the model. Therefore, our design study aims to provide a visual analytics solution to increase interpretability and interactivity of RNNs via a joint effort of medical experts, artificial intelligence scientists, and visual analytics researchers. Following the iterative design process between the experts, we design, implement, and evaluate a visual analytics tool called RetainVis, which couples a newly improved, interpretable and interactive RNN-based model called RetainEX and visualizations for users' exploration of EMR data in the context of prediction tasks. Our study shows the effective use of RetainVis for gaining insights into how individual medical codes contribute to making risk predictions, using EMRs of patients with heart failure and cataract symptoms. Our study also demonstrates how we made substantial changes to the state-of-the-art RNN model called RETAIN in order to make use of temporal information and increase interactivity. This study will provide a useful guideline for researchers that aim to design an interpretable and interactive visual analytics tool for RNNs.

Jina Huh, Bum Chul Kwon, Jaegul Choo et al.

As online health communities (OHCs) grow, users find it challenging to properly search, read, and contribute to the community because of its overwhelming content. Our goal is to understand OHC users' needs and requirements for better delivering large-scale OHC content. We interviewed 14 OHC users with interests in diabetes to investigate their attitudes and needs towards using OHCs and 2 OHC administrators to assess our findings. Four personas -Coddlers, Scientists, Adventurers, and Opportunists- emerged, which inform users' interaction behavior and attitudes with OHCs. An individual can possess the characteristics of multiple personas, which can also change over time. Our personas uniquely describe users' OHC participation intertwined with illness contexts compared to existing social types in general online communities. We discuss broader implications back to the literature and how our findings apply to other illness contexts in OHCs. We end with requirements for personalized delivery of large-scale OHC content.