To design a successful Multiplayer Online Battle Arena (MOBA) game, the ratio of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">snowballing</i> and comeback occurrences to all matches played must be maintained at certain level ensure its fairness engagement. Although it is easy identify these two types occurrences, game developers often find difficult determine their causes triggers with so many choices parameters involved. In addition, huge...

The message a designer wants to convey plays pivotal role in directing the design of an infographic, yet most authoring workflows start with creating visualizations or graphics first without gauging whether they fit message. To address this gap, we propose Epigraphics, web-based system that treats "epigraph" as first-class object, and uses it guide infographic asset creation, editing, syncing. text-based recommend visualizations, graphics, data filters, color palettes, animations. It further...

Visual data stories can effectively convey insights from data, yet their creation often necessitates intricate exploration, insight discovery, narrative organization, and customization to meet the communication objectives of storyteller. Existing automated storytelling techniques, however, tend overlook importance user during story authoring process, limiting system's ability create tailored narratives that reflect user's intentions. We present a novel generation workflow leverages adaptive...

Bipartite graphs model the key relations in many large scale real-world data: customers purchasing items, legislators voting for bills, people's affiliation with different social groups, faults occurring vehicles, etc. However, it is challenging to visualize bipartite tens of thousands or even more nodes edges. In this paper, we propose a novel visual summarization technique based on minimum description length (MDL) principle. The method simultaneously groups two set and constructs...

The brachial plexus is a complex network of peripheral nerves that enables sensing from and control the movements arms hand. Nowadays, coordination between muscles to generate simple still not well understood, hindering knowledge how best treat patients with this type nerve injury. To acquire enough information for medical data analysis, physicians conduct motion analysis assessments produce rich dataset electromyographic signals multiple recorded joint during real-world tasks. However,...

With the increasing sophistication of machine learning models, there are growing trends developing model explanation techniques that focus on only one instance (local explanation) to ensure faithfulness original model. While these provide accurate interpretability various data primitive (e.g., tabular, image, or text), a holistic Explainable Artificial Intelligence (XAI) experience also requires global and dataset enable sensemaking in different granularity. Thus, is vast potential...

Understanding the interpretation of machine learning (ML) models has been paramount importance when making decisions with societal impacts, such as transport control, financial activities, and medical diagnosis. While local explanation techniques are popular methods to interpret ML on a single instance, they do not scale understanding model's behavior whole dataset. In this article, we outline challenges needs visually analyzing explanations propose SUBPLEX, visual analytics approach help...

Understanding the interpretation of machine learning (ML) models has been paramount importance when making decisions with societal impacts such as transport control, financial activities, and medical diagnosis. While current model methodologies focus on using locally linear functions to approximate or creating self-explanatory that give explanations each input instance, they do not at subpopulation level, which is understanding interpretations across different subset aggregations in a...

Chieh-Yang Huang, Ting-Yao Hsu, Ryan Rossi, Ani Nenkova, Sungchul Kim, Gromit Yeuk-Yin Chan, Eunyee Koh, C Lee Giles, Ting-Hao Huang. Proceedings of the 16th International Natural Language Generation Conference. 2023.

When reading financial news, although there are critics explaining the fluctuation of Economic indexes in articles everyday, news often assert bias on authors' favorite opinions. On other hand, amount published these days is staggering with diversified opinions same issues. Unless acumen for whole environment, audience will find hard to stay objective and identify useful information from mass media. If computer granted ability analyze all generate a narrative that addresses concerns related...

The Data Polygamy framework allows users to uncover interesting patterns and interactions in the data exhaust from different components of an urban environment. But analyzing plethora relationships derived by is challenging. In this demo, we show how visualization can help discovery that are potentially allowing query explore relationship set intuitive way. We will demonstrate effectiveness visual interface through case studies, demo visitors also interact with polygamous relationships.

Online visitor behaviors are often modeled as a large sparse matrix, where rows represent visitors and columns behavior. To discover customer segments with different hierarchies, marketers need to cluster the data in splits. Such analyses require clustering algorithm provide real-time responses on user parameter changes, which current techniques cannot support. In this paper, we propose algorithm, density peaks, for large-scale data. It pre-processes input points compute annotations...

With the increasing use of black-box Machine Learning (ML) techniques in critical applications, there is a growing demand for methods that can provide transparency and accountability model predictions. As result, large number local explainability models have been developed popularized. However, machine learning explanations are still hard to evaluate compare due high dimensionality, heterogeneous representations, varying scales, stochastic nature some these methods. Topological Data Analysis...

With the increasing use of black-box Machine Learning (ML) techniques in critical applications, there is a growing demand for methods that can provide transparency and accountability model predictions. As result, large number local explainability models have been developed popularized. However, machine learning explanations are still hard to evaluate compare due high dimensionality, heterogeneous representations, varying scales, stochastic nature some these methods. Topological Data Analysis...

Augmented Reality assistance are increasingly popular for supporting users with tasks like assembly and cooking. However, current practice typically provide reactive responses initialized from user requests, lacking consideration of rich contextual user-specific information. To address this limitation, we propose a novel AR system, Satori, that models both states environmental contexts to deliver proactive guidance. Our system combines the Belief-Desire-Intention (BDI) model state-of-the-art...

Ensuring large language models' (LLMs) responses align with prompt instructions is crucial for application development. Based on our formative study industry professionals, the alignment requires heavy human involvement and tedious trial-and-error especially when there are many in prompt. To address these challenges, we introduce CoPrompter, a framework that identifies misalignment based assessing multiple LLM criteria. It proposes method to generate evaluation criteria questions derived...

Online marketing platforms often store millions of website visitors' behavior as a large sparse matrix with rows visitors and columns behavior. These allow marketers to conduct Audience Expansion, technique identify new audiences similar the original target audiences. In this paper, we propose method achieve interactive Expansion from visitor data efficiently. Unlike other methods that undergo significant computations upon inputs, our approach provides responses when marketer inputs...

Good figure captions help paper readers understand complex scientific figures. Unfortunately, even published papers often have poorly written captions. Automatic caption generation could aid writers by providing good starting that can be refined for better quality. Prior work treated as a vision-to-language task. In this paper, we show it more effectively tackled text summarization task in documents. We fine-tuned PEGASUS, pre-trained abstractive model, to specifically summarize...

Detecting structures and components in business emails is vital for editor software to convert third-party so that designers can edit them without needing know how HTML works. In a production environment, the challenge make model easy be understood maintained by different stakeholders. We propose detecting email with collection of constraints written Answer Set Programming (ASP). Hard detect well-defined like layouts, soft incorporate ML custom buttons titles emails. Using constraints,...

Given a large corpus of HTML-based emails (or websites, posters, documents) collected from the web, how can we train model capable learning such rich heterogeneous data for style recommendation tasks as recommending useful design styles or suggesting alternative HTML designs? To address this new task, first decompose each document in into sequence smaller fragments where fragment may consist set entities buttons, images, textual content (titles, paragraphs) and stylistic background-style,...

The ability to extract insights from large amounts of data in a timely manner is crucial problem. Exploratory Data Analysis (EDA) commonly used by analysts uncover using sequence SQL commands and associated visualizations. However, many cases, this process carried out non-programmers who must work within tight time constraints, such as marketing campaign where marketer quickly analyse reach target revenue. This paper presents ApproxEDA - system that combines natural language processing (NLP)...