arXiv subjects

Anjana Arunkumar

Publications and source records attributed to Anjana Arunkumar.

At least 19 recordsLinked to original sources

arXiv2026

Visualizing in the Mind's Eye: Icon Design Shapes Mental Imagery of Fire Risks

We introduce mental imagery, or seeing images in the "mind's eye," as a cognitive process that can be shaped by data visualization design and in turn impact decisions. We found in a preregistered study (n = 400) that abstract geometric icons visualizing fire risk data evoked more mental images than concrete house-on-fire icons, and also produced more diverse and personalized mental images. Mediation analysis showed that increased mental imagery subsequently led to risk-averse decisions through evoking negative affect. These findings reveal a nuanced mechanism through which visualization concreteness influences decisions: concrete designs may actually suppress affect-driven behavior by restricting mental imagery.

cs.HC↗

arXiv2025

An Analysis of Text Functions in Information Visualization

Text is an integral but understudied component of visualization design. Although recent studies have examined how text elements (e.g., titles and annotations) influence comprehension, preferences, and predictions, many questions remain about textual design and use in practice. This paper introduces a framework for understanding text functions in information visualizations, building on and filling gaps in prior classifications and taxonomies. Through an analysis of 120 real-world visualizations and 804 text elements, we identified ten distinct text functions, ranging from identifying data mappings to presenting valenced subtext. We further identify patterns in text usage and conduct a factor analysis, revealing four overarching text-informed design strategies: Attribution and Variables, Annotation-Centric Design, Visual Embellishments, and Narrative Framing. In addition to these factors, we explore features of title rhetoric and text multifunctionality, while also uncovering previously unexamined text functions, such as text replacing visual elements. Our findings highlight the flexibility of text, demonstrating how different text elements in a given design can combine to communicate, synthesize, and frame visual information. This framework adds important nuance and detail to existing frameworks that analyze the diverse roles of text in visualization.

cs.HC↗

arXiv2025

Lost in Translation: How Does Bilingualism Shape Reader Preferences for Annotated Charts?

Visualizations are powerful tools for conveying information but often rely on accompanying text for essential context and guidance. This study investigates the impact of annotation patterns on reader preferences and comprehension accuracy among multilingual populations, addressing a gap in visualization research. We conducted experiments with two groups fluent in English and either Tamil (n = 557) or Arabic (n = 539) across six visualization types, each varying in annotation volume and semantic content. Full-text annotations yielded the highest comprehension accuracy across all languages, while preferences diverged: English readers favored highly annotated charts, whereas Tamil/Arabic readers preferred full-text or minimally annotated versions. Semantic variations in annotations (L1-L4) did not significantly affect comprehension, demonstrating the robustness of text comprehension across languages. English annotations were generally preferred, with a tendency to think technically in English linked to greater aversion to non-English annotations, though this diminished among participants who regularly switched languages internally. Non-English annotations incorporating visual or external knowledge were less favored, particularly in titles. Our findings highlight cultural and educational factors influencing perceptions of visual information, underscoring the need for inclusive annotation practices for diverse linguistic audiences. All data and materials are available at: https://osf.io/ckdb4/.

cs.HC↗

arXiv2024

Mind Drifts, Data Shifts: Utilizing Mind Wandering to Track the Evolution of User Experience with Data Visualizations

User experience in data visualization is typically assessed through post-viewing self-reports, but these overlook the dynamic cognitive processes during interaction. This study explores the use of mind wandering -- a phenomenon where attention spontaneously shifts from a primary task to internal, task-related thoughts or unrelated distractions -- as a dynamic measure during visualization exploration. Participants reported mind wandering while viewing visualizations from a pre-labeled visualization database and then provided quantitative ratings of trust, engagement, and design quality, along with qualitative descriptions and short-term/long-term recall assessments. Results show that mind wandering negatively affects short-term visualization recall and various post-viewing measures, particularly for visualizations with little text annotation. Further, the type of mind wandering impacts engagement and emotional response. Mind wandering also functions as an intermediate process linking visualization design elements to post-viewing measures, influencing how viewers engage with and interpret visual information over time. Overall, this research underscores the importance of incorporating mind wandering as a dynamic measure in visualization design and evaluation, offering novel avenues for enhancing user engagement and comprehension.

cs.HC↗

arXiv2023

Image or Information? Examining the Nature and Impact of Visualization Perceptual Classification

How do people internalize visualizations: as images or information? In this study, we investigate the nature of internalization for visualizations (i.e., how the mind encodes visualizations in memory) and how memory encoding affects its retrieval. This exploratory work examines the influence of various design elements on a user's perception of a chart. Specifically, which design elements lead to perceptions of visualization as an image or as information? Understanding how design elements contribute to viewers perceiving a visualization more as an image or information will help designers decide which elements to include to achieve their communication goals. For this study, we annotated 500 visualizations and analyzed the responses of 250 online participants, who rated the visualizations on a bilinear scale as image or information. We then conducted an in-person study (n = 101) using a free recall task to examine how the image/information ratings and design elements impact memory. The results revealed several interesting findings: Image-rated visualizations were perceived as more aesthetically appealing, enjoyable, and pleasing. Information-rated visualizations were perceived as less difficult to understand and more aesthetically likable and nice, though participants expressed higher positive sentiment when viewing image-rated visualizations and felt less guided to a conclusion. We also found different patterns among participants that were older. Importantly, we show that visualizations internalized as images are less effective in conveying trends and messages, though they elicit a more positive emotional judgment, while informative visualizations exhibit annotation focused recall and elicit a more positive design judgment. We discuss the implications of this dissociation between aesthetic pleasure and perceived ease of use in visualization design.

cs.HC↗

arXiv2023

LINGO : Visually Debiasing Natural Language Instructions to Support Task Diversity

Cross-task generalization is a significant outcome that defines mastery in natural language understanding. Humans show a remarkable aptitude for this, and can solve many different types of tasks, given definitions in the form of textual instructions and a small set of examples. Recent work with pre-trained language models mimics this learning style: users can define and exemplify a task for the model to attempt as a series of natural language prompts or instructions. While prompting approaches have led to higher cross-task generalization compared to traditional supervised learning, analyzing 'bias' in the task instructions given to the model is a difficult problem, and has thus been relatively unexplored. For instance, are we truly modeling a task, or are we modeling a user's instructions? To help investigate this, we develop LINGO, a novel visual analytics interface that supports an effective, task-driven workflow to (1) help identify bias in natural language task instructions, (2) alter (or create) task instructions to reduce bias, and (3) evaluate pre-trained model performance on debiased task instructions. To robustly evaluate LINGO, we conduct a user study with both novice and expert instruction creators, over a dataset of 1,616 linguistic tasks and their natural language instructions, spanning 55 different languages. For both user groups, LINGO promotes the creation of more difficult tasks for pre-trained models, that contain higher linguistic diversity and lower instruction bias. We additionally discuss how the insights learned in developing and evaluating LINGO can aid in the design of future dashboards that aim to minimize the effort involved in prompt creation across multiple domains.

cs.HC↗

arXiv2023

PromptAid: Prompt Exploration, Perturbation, Testing and Iteration using Visual Analytics for Large Language Models

Large Language Models (LLMs) have gained widespread popularity due to their ability to perform ad-hoc Natural Language Processing (NLP) tasks with a simple natural language prompt. Part of the appeal for LLMs is their approachability to the general public, including individuals with no prior technical experience in NLP techniques. However, natural language prompts can vary significantly in terms of their linguistic structure, context, and other semantics. Modifying one or more of these aspects can result in significant differences in task performance. Non-expert users may find it challenging to identify the changes needed to improve a prompt, especially when they lack domain-specific knowledge and lack appropriate feedback. To address this challenge, we present PromptAid, a visual analytics system designed to interactively create, refine, and test prompts through exploration, perturbation, testing, and iteration. PromptAid uses multiple, coordinated visualizations which allow users to improve prompts by using the three strategies: keyword perturbations, paraphrasing perturbations, and obtaining the best set of in-context few-shot examples. PromptAid was designed through an iterative prototyping process involving NLP experts and was evaluated through quantitative and qualitative assessments for LLMs. Our findings indicate that PromptAid helps users to iterate over prompt template alterations with less cognitive overhead, generate diverse prompts with help of recommendations, and analyze the performance of the generated prompts while surpassing existing state-of-the-art prompting interfaces in performance.

cs.HC↗

arXiv2023

Real-Time Visual Feedback to Guide Benchmark Creation: A Human-and-Metric-in-the-Loop Workflow

Recent research has shown that language models exploit `artifacts' in benchmarks to solve tasks, rather than truly learning them, leading to inflated model performance. In pursuit of creating better benchmarks, we propose VAIDA, a novel benchmark creation paradigm for NLP, that focuses on guiding crowdworkers, an under-explored facet of addressing benchmark idiosyncrasies. VAIDA facilitates sample correction by providing realtime visual feedback and recommendations to improve sample quality. Our approach is domain, model, task, and metric agnostic, and constitutes a paradigm shift for robust, validated, and dynamic benchmark creation via human-and-metric-in-the-loop workflows. We evaluate via expert review and a user study with NASA TLX. We find that VAIDA decreases effort, frustration, mental, and temporal demands of crowdworkers and analysts, simultaneously increasing the performance of both user groups with a 45.8% decrease in the level of artifacts in created samples. As a by product of our user study, we observe that created samples are adversarial across models, leading to decreases of 31.3% (BERT), 22.5% (RoBERTa), 14.98% (GPT-3 fewshot) in performance.

cs.CL↗

arXiv2022

Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks

How well can NLP models generalize to a variety of unseen tasks when provided with task instructions? To address this question, we first introduce Super-NaturalInstructions, a benchmark of 1,616 diverse NLP tasks and their expert-written instructions. Our collection covers 76 distinct task types, including but not limited to classification, extraction, infilling, sequence tagging, text rewriting, and text composition. This large and diverse collection of tasks enables rigorous benchmarking of cross-task generalization under instructions -- training models to follow instructions on a subset of tasks and evaluating them on the remaining unseen ones. Furthermore, we build Tk-Instruct, a transformer model trained to follow a variety of in-context instructions (plain language task definitions or k-shot examples). Our experiments show that Tk-Instruct outperforms existing instruction-following models such as InstructGPT by over 9% on our benchmark despite being an order of magnitude smaller. We further analyze generalization as a function of various scaling parameters, such as the number of observed tasks, the number of instances per task, and model sizes. We hope our dataset and model facilitate future progress towards more general-purpose NLP models.

cs.CL↗

arXiv2022

A Survey of Parameters Associated with the Quality of Benchmarks in NLP

Several benchmarks have been built with heavy investment in resources to track our progress in NLP. Thousands of papers published in response to those benchmarks have competed to top leaderboards, with models often surpassing human performance. However, recent studies have shown that models triumph over several popular benchmarks just by overfitting on spurious biases, without truly learning the desired task. Despite this finding, benchmarking, while trying to tackle bias, still relies on workarounds, which do not fully utilize the resources invested in benchmark creation, due to the discarding of low quality data, and cover limited sets of bias. A potential solution to these issues -- a metric quantifying quality -- remains underexplored. Inspired by successful quality indices in several domains such as power, food, and water, we take the first step towards a metric by identifying certain language properties that can represent various possible interactions leading to biases in a benchmark. We look for bias related parameters which can potentially help pave our way towards the metric. We survey existing works and identify parameters capturing various properties of bias, their origins, types and impact on performance, generalization, and robustness. Our analysis spans over datasets and a hierarchy of tasks ranging from NLI to Summarization, ensuring that our parameters are generic and are not overfitted towards a specific task or dataset. We also develop certain parameters in this process.

cs.CL↗

arXiv2022

Hardness of Samples Need to be Quantified for a Reliable Evaluation System: Exploring Potential Opportunities with a New Task

Evaluation of models on benchmarks is unreliable without knowing the degree of sample hardness; this subsequently overestimates the capability of AI systems and limits their adoption in real world applications. We propose a Data Scoring task that requires assignment of each unannotated sample in a benchmark a score between 0 to 1, where 0 signifies easy and 1 signifies hard. Use of unannotated samples in our task design is inspired from humans who can determine a question difficulty without knowing its correct answer. This also rules out the use of methods involving model based supervision (since they require sample annotations to get trained), eliminating potential biases associated with models in deciding sample difficulty. We propose a method based on Semantic Textual Similarity (STS) for this task; we validate our method by showing that existing models are more accurate with respect to the easier sample-chunks than with respect to the harder sample-chunks. Finally we demonstrate five novel applications.

cs.CL↗

arXiv2022

Investigating the Failure Modes of the AUC metric and Exploring Alternatives for Evaluating Systems in Safety Critical Applications

With the increasing importance of safety requirements associated with the use of black box models, evaluation of selective answering capability of models has been critical. Area under the curve (AUC) is used as a metric for this purpose. We find limitations in AUC; e.g., a model having higher AUC is not always better in performing selective answering. We propose three alternate metrics that fix the identified limitations. On experimenting with ten models, our results using the new metrics show that newer and larger pre-trained models do not necessarily show better performance in selective answering. We hope our insights will help develop better models tailored for safety-critical applications.

cs.CL↗

arXiv2022

PMU Tracker: A Visualization Platform for Epicentric Event Propagation Analysis in the Power Grid

The electrical power grid is a critical infrastructure, with disruptions in transmission having severe repercussions on daily activities, across multiple sectors. To identify, prevent, and mitigate such events, power grids are being refurbished as 'smart' systems that include the widespread deployment of GPS-enabled phasor measurement units (PMUs). PMUs provide fast, precise, and time-synchronized measurements of voltage and current, enabling real-time wide-area monitoring and control. However, the potential benefits of PMUs, for analyzing grid events like abnormal power oscillations and load fluctuations, are hindered by the fact that these sensors produce large, concurrent volumes of noisy data. In this paper, we describe working with power grid engineers to investigate how this problem can be addressed from a visual analytics perspective. As a result, we have developed PMU Tracker, an event localization tool that supports power grid operators in visually analyzing and identifying power grid events and tracking their propagation through the power grid's network. As a part of the PMU Tracker interface, we develop a novel visualization technique which we term an epicentric cluster dendrogram, which allows operators to analyze the effects of an event as it propagates outwards from a source location. We robustly validate PMU Tracker with: (1) a usage scenario demonstrating how PMU Tracker can be used to analyze anomalous grid events, and (2) case studies with power grid operators using a real-world interconnection dataset. Our results indicate that PMU Tracker effectively supports the analysis of power grid events; we also demonstrate and discuss how PMU Tracker's visual analytics approach can be generalized to other domains composed of time-varying networks with epicentric event characteristics.

cs.HC↗

arXiv2022

A Proposal to Study "Is High Quality Data All We Need?"

Even though deep neural models have achieved superhuman performance on many popular benchmarks, they have failed to generalize to OOD or adversarial datasets. Conventional approaches aimed at increasing robustness include developing increasingly large models and augmentation with large scale datasets. However, orthogonal to these trends, we hypothesize that a smaller, high quality dataset is what we need. Our hypothesis is based on the fact that deep neural networks are data driven models, and data is what leads/misleads models. In this work, we propose an empirical study that examines how to select a subset of and/or create high quality benchmark data, for a model to learn effectively. We seek to answer if big datasets are truly needed to learn a task, and whether a smaller subset of high quality data can replace big datasets. We plan to investigate both data pruning and data creation paradigms to generate high quality datasets.

cs.LG↗

arXiv2021

Bayesian Modelling of Alluvial Diagram Complexity

Alluvial diagrams are a popular technique for visualizing flow and relational data. However, successfully reading and interpreting the data shown in an alluvial diagram is likely influenced by factors such as data volume, complexity, and chart layout. To understand how alluvial diagram consumption is impacted by its visual features, we conduct two crowdsourced user studies with a set of alluvial diagrams of varying complexity, and examine (i) participant performance on analysis tasks, and (ii) the perceived complexity of the charts. Using the study results, we employ Bayesian modelling to predict participant classification of diagram complexity. We find that, while multiple visual features are important in contributing to alluvial diagram complexity, interestingly the importance of features seems to depend on the type of complexity being modeled, i.e. task complexity vs. perceived complexity.

cs.HC↗

arXiv2021

How Robust are Model Rankings: A Leaderboard Customization Approach for Equitable Evaluation

Models that top leaderboards often perform unsatisfactorily when deployed in real world applications; this has necessitated rigorous and expensive pre-deployment model testing. A hitherto unexplored facet of model performance is: Are our leaderboards doing equitable evaluation? In this paper, we introduce a task-agnostic method to probe leaderboards by weighting samples based on their `difficulty' level. We find that leaderboards can be adversarially attacked and top performing models may not always be the best models. We subsequently propose alternate evaluation metrics. Our experiments on 10 models show changes in model ranking and an overall reduction in previously reported performance -- thus rectifying the overestimation of AI systems' capabilities. Inspired by behavioral testing principles, we further develop a prototype of a visual analytics tool that enables leaderboard revamping through customization, based on an end user's focus area. This helps users analyze models' strengths and weaknesses, and guides them in the selection of a model best suited for their application scenario. In a user study, members of various commercial product development teams, covering 5 focus areas, find that our prototype reduces pre-deployment development and testing effort by 41% on average.

cs.CL↗

arXiv2021

Front Contribution instead of Back Propagation

Deep Learning's outstanding track record across several domains has stemmed from the use of error backpropagation (BP). Several studies, however, have shown that it is impossible to execute BP in a real brain. Also, BP still serves as an important and unsolved bottleneck for memory usage and speed. We propose a simple, novel algorithm, the Front-Contribution algorithm, as a compact alternative to BP. The contributions of all weights with respect to the final layer weights are calculated before training commences and all the contributions are appended to weights of the final layer, i.e., the effective final layer weights are a non-linear function of themselves. Our algorithm then essentially collapses the network, precluding the necessity for weight updation of all weights not in the final layer. This reduction in parameters results in lower memory usage and higher training speed. We show that our algorithm produces the exact same output as BP, in contrast to several recently proposed algorithms approximating BP. Our preliminary experiments demonstrate the efficacy of the proposed algorithm. Our work provides a foundation to effectively utilize these presently under-explored "front contributions", and serves to inspire the next generation of training algorithms.

cs.LG↗

arXiv2020

DQI: A Guide to Benchmark Evaluation

A `state of the art' model A surpasses humans in a benchmark B, but fails on similar benchmarks C, D, and E. What does B have that the other benchmarks do not? Recent research provides the answer: spurious bias. However, developing A to solve benchmarks B through E does not guarantee that it will solve future benchmarks. To progress towards a model that `truly learns' an underlying task, we need to quantify the differences between successive benchmarks, as opposed to existing binary and black-box approaches. We propose a novel approach to solve this underexplored task of quantifying benchmark quality by debuting a data quality metric: DQI.

cs.CL↗