Conversational search, powered by Large Language Models (LLMs), has rapidly become a dominant mode of information seeking. While LLMs reduce the effort of information seeking, they also introduce the risk of distorted information deceptively embedded in responses. Prior work has sought technical mitigations, but such distortion cannot be fully eliminated. We therefore shift the focus to the user level, supporting users in deliberately engaging with information when reading LLM responses. We conducted a user study with frequent conversational search users (N=16), comparing a baseline with two probes—LLM-box (LLM-as-a-judge feedback) and Thinking-box (checkpoints from hallucination patterns)—to examine how these probes influenced users’ recognition of distorted information and their experience of guidance. Our findings indicate that even indirect suggestions significantly improved users’ ability to filter distorted information, while also revealing that guidance must be selective to prevent cognitive overload. These insights point to design implications that enable more deliberate user engagement with LLM responses.

The field of Multimodal Large Language Models (MLLMs) has made remarkable progress in visual understanding tasks, presenting a vast opportunity to predict the perceptual and emotional impact of charts. However, it also raises concerns, as many applications of LLMs are based on overgeneralized assumptions from a few examples, lacking sufficient validation of their performance and effectiveness. We introduce Chart-to-Experience, a benchmark dataset comprising 36 charts, evaluated by crowdsourced workers for their impact on seven experiential factors. Using the dataset as ground truth, we evaluated capabilities of state-of-the-art MLLMs on two tasks: direct prediction and pairwise comparison of charts. Our findings imply that MLLMs are not as sensitive as human evaluators when assessing individual charts, but are accurate and reliable in pairwise comparisons.

The integration of generative AI in education is expanding, yet empirical analyses of large-scale and real-world interactions between students and AI systems still remain limited. Addressing this gap, we present RECIPE4U (RECIPE for University), a dataset sourced from a semester-long experiment with 212 college students in English as Foreign Language (EFL) writing courses. During the study, students engaged in dialogues with ChatGPT to revise their essays. RECIPE4U includes comprehensive records of these interactions, including conversation logs, students’ intent, students’ self-rated satisfaction, and students’ essay edit histories. In particular, we annotate the students’ utterances in RECIPE4U with 13 intention labels based on our coding schemes. We establish baseline results for two subtasks in task-oriented dialogue systems within educational contexts: intent detection and satisfaction estimation. As a foundational step, we explore student-ChatGPT interaction patterns through RECIPE4U and analyze them by focusing on students’ dialogue, essay data statistics, and students’ essay edits. We further illustrate potential applications of RECIPE4U dataset for enhancing the incorporation of LLMs in educational frameworks. RECIPE4U is publicly available at https://zeunie.github.io/RECIPE4U/.

Visualization recommendation systems make understanding data more accessible to users of all skill levels by automatically generating visualizations for users to explore. However, most existing visualization recommendation systems focus on ranking all possible visualizations based on the attributes or encodings, which makes it difficult to find the most interesting or relevant insights. We therefore introduce a novel class of visualization recommendation systems that automatically rank and recommend both groups of related insights and the most important insights within each group. Our approach combines results across different learning-based methods to discover insights automatically and generalizes to a variety of attribute types (e.g., categorical, numerical, and temporal), including non-trivial combinations of these attribute types. We then implemented a new insight-centric visualization recommendation system, SpotLight, which ranks annotated visualizations in visual insight groups. Finally, we conducted a user study which showed that users are able to quickly understand and find relevant insights in unfamiliar data.

The integration of generative AI in the field of education is actively being explored. In particular, ChatGPT has garnered significant interest, offering an opportunity to examine its effectiveness in English as a foreign language (EFL) education. To address this need, we present a novel learning platform called RECIPE (Revising an Essay with ChatGPT on an Interactive Platform for EFL learners). Our platform features two types of prompts that facilitate conversations between ChatGPT and students: (1) a hidden prompt for ChatGPT to take an EFL teacher role and (2) an open prompt for students to initiate a dialogue with a self-written summary of what they have learned. We deployed this platform for 213 undergraduate and graduate students enrolled in EFL writing courses and seven instructors. For this study, we collect students' interaction data from RECIPE, including students' perceptions and usage of the platform, and user scenarios are examined with the data. We also conduct a focus group interview with six students and an individual interview with one EFL instructor to explore design opportunities for leveraging generative AI models in the field of EFL education.

Although we have seen a proliferation of algorithms for recommending visualizations, these algorithms are rarely compared with one another, making it difficult to ascertain which algorithm is best for a given visual analysis scenario. Though several formal frameworks have been proposed in response, we believe this issue persists because visualization recommendation algorithms are inadequately specified from an evaluation perspective. In this paper, we propose an evaluation-focused framework to contextualize and compare a broad range of visualization recommendation algorithms. We present the structure of our framework, where algorithms are specified using three components: (1) a graph representing the full space of possible visualization designs, (2) the method used to traverse the graph for potential candidates for recommendation, and (3) an oracle used to rank candidate designs. To demonstrate how our framework guides the formal comparison of algorithmic performance, we not only theoretically compare five existing representative recommendation algorithms, but also empirically compare four new algorithms generated based on our findings from the theoretical comparison. Our results show that these algorithms behave similarly in terms of user performance, highlighting the need for more rigorous formal comparisons of recommendation algorithms to further clarify their benefits in various analysis scenarios.

Visualization recommendation work has focused solely on scoring visualizations based on the underlying dataset, and not the actualuserand their past visualization feedback. These systems recommend the same visualizations for every user, despite that the underlying user interests, intent, and visualization preferences are likely to be fundamentally different, yet vitally important. In this work, we formally introduce the problem ofpersonalized visualization recommendationand present a generic learning framework for solving it. In particular, we focus on recommending visualizations personalized for each individual user based on their past visualization interactions (e.g., viewed, clicked, manually created) along with the data from those visualizations. More importantly, the framework can learn from visualizations relevant to other users, even if the visualizations are generated from completely different datasets. Experiments demonstrate the effectiveness of the approach as it leads to higher quality visualization recommendations tailored to the specific user intent and preferences. To support research on this new problem, we release our user-centric visualization corpus consisting of 17.4k users exploring 94k datasets with 2.3 million attributes and 32k user-generated visualizations.

Visualization recommendation is important for exploratory analysis and making sense of the data quickly by automatically recommending relevant visualizations to the user. In this work, we propose the first end-to-end ML-based visualization recommendation system that leverages a large corpus of datasets and their relevant visualizations to learn a visualization recommendation model automatically. Then, given a new unseen dataset from an arbitrary user, the model automatically generates visualizations for that new dataset, derives scores for the visualizations, and outputs a list of recommended visualizations to the user ordered by effectiveness. We also describe an evaluation framework to quantitatively evaluate visualization recommendation models learned from a large corpus of visualizations and datasets. Through quantitative experiments, a user study, and qualitative analysis, we show that our end-to-end ML-based system recommends more effective and useful visualizations compared to existing state-of-the-art rule-based systems.

Column annotation, the process of annotating tabular columns with labels, plays a fundamental role in digital marketing data governance. It has a direct impact on how customers manage their data and facilitates compliance with regulations, restrictions, and policies applicable to data use. Despite substantial gains in accuracy brought by recent deep learning-driven column annotation methods, their incapability of explaining why columns are matched with particular target labels has drawn concern, due to the black-box nature of deep neural networks. Such explainability is of particular importance in industrial marketing scenarios, where data stewards need to quickly verify and calibrate the annotation results to ascertain the correctness of downstream applications. This work sheds new light on the explainable column annotation problem, the first of its kind column annotation task. To achieve this, we propose a new approach called EXACTA, which conducts multi-hop knowledge graph reasoning using inverse reinforcement learning to find a path from a column to a potential target label while ensuring both annotation performance and explainability. We experiment on four benchmarks, both publicly available and real-world ones, and undertake a comprehensive analysis on the explainability. The results suggest that our method not only provides competitive annotation performance compared with existing deep learning-based models, but more importantly, produces faithfully explainable paths for annotated columns to facilitate human examination.

Scientific-style figures are commonly used on the web to present numerical information. Captions that tell accurate figure information and sound natural would significantly improve figure accessibility. In this paper, we present promising results on machine figure captioning. A recent corpus analysis of real-world captions reveals that machine figure captioning systems should start by generating accurate caption units. We formulate the caption unit generation problem as a controlled captioning problem. Given a caption unit type as a control signal, a model generates an accurate caption unit of that type. As a proof-of-concept on single bar charts, we propose a model, FigJAM, that achieves this goal through utilizing metadata information and a joint static and dynamic dictionary. Quantitative evaluations with two datasets from the figure question answering task show that our model can generate more accurate caption units than competitive baseline models. A user study with ten human experts confirms the value of machine-generated caption units in their standalone accuracy and naturalness. Finally, a post-editing simulation study demonstrates the potential for models to paraphrase and stitch together single-type caption units into multi-type captions by learning from data.

Programming-by-Example (PBE) enables users to create programs without writing a line of code. However, there is little research on people's ability to accomplish complex tasks by providing examples, which is the key to successful PBE solutions. This paper presents an online user study, which reports observations on how well people decompose complex tasks, and disambiguate sub-tasks. Our findings suggest that disambiguation and decomposition are difficult for inexperienced users. We identify seven types of mistakes made, and suggest new opportunities for actionable feedback based on unsuccessful examples, with design implications for future PBE systems.

Topic modeling is a common tool for understanding large bodies of text, but is typically provided as a “take it or leave it” proposition. Incorporating human knowledge in unsupervised learning is a promising approach to create high-quality topic models. Existing interactive systems and modeling algorithms support a wide range of refinement operations to express feedback. However, these systems’ interactions are primarily driven by algorithmic convenience, ignoring users who may lack expertise in topic modeling. To better understand how non-expert users understand, assess, and refine topics, we conducted two user studies—an in-person interview study and an online crowdsourced study. These studies demonstrate a disconnect between what non-expert users want and the complex, low-level operations that current interactive systems support . In particular, our findings include: (1) analysis of how non-expert users perceive topic models; (2) characterization of primary refinement operations expected by non-expert users and ordered by relative preference; (3) further evidence of the benefits of supporting users in directly refining a topic model; (4) design implications for future human-in-the-loop topic modeling interfaces.

Probabilistic topic models are important tools for indexing, summarizing, and analyzing large document collections by their themes. However, promoting end-user understanding of topics remains an open research problem. We compare labels generated by users given four topic visualization techniques—word lists, word lists with bars, word clouds, and network graphs—against each other and against automatically generated labels. Our basis of comparison is participant ratings of how well labels describe documents from the topic. Our study has two phases: a labeling phase where participants label visualized topics and a validation phase where different participants select which labels best describe the topics’ documents. Although all visualizations produce similar quality labels, simple visualizations such as word lists allow participants to quickly understand topics, while complex visualizations take longer but expose multi-word expressions that simpler visualizations obscure. Automatic labels lag behind user-created labels, but our dataset of manually labeled topics highlights linguistic patterns (e.g., hypernyms, phrases) that can be used to improve automatic topic labeling algorithms.