GANSlider: How Users Control Generative Models for Images using Multiple Sliders with and without Feedforward Information

Hai, Dang,; Mecke, Lukas; Buschek, Daniel

doi:10.1145/3491102.3502141

Cited by 32 publications

(29 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The present research supports the above by providing evidence of how the system responsiveness enables the user comfort. The development of CoAuthor as a writing "collaborator" (M. proved to be bene cial in drawing a conclusion about the signi cance of the current research which highlights how interactions, feedback, and responsiveness play a vital role in fostering user comfort Dang et al 2023. provided insights into the interaction between humans and generative models, contributing to our understanding of how humans and AI can effectively collaborate, moreover, research discovered that an escalation in collaborative activity positively impacted multiple aspects of user experience efforts(Feng et al 2023), which are in consonance with the ndings indicating the effect of user responsiveness and feedback on the comfort of the user.The present study's ndings that indicated the users' comfort in using the voice-based interface is in consonance with the results identifying the use of speech based interactive systems to be coveted andtechnology facilitating natural, authentic and seamless conversations (Savcheva & Foster 2018, Strayer et al 2017, Oesterreich et al 2023, Wang et al 2023).…”

mentioning

confidence: 81%

Understanding User comfort and Expectations in AI-based Systems

Dave¹,

Saxena

Jha³

2023

Preprint

View full text Add to dashboard Cite

This study investigated the impact of prior experience and education levels on user expectations in Artificial Intelligence (AI) based systems. The research aimed to determine whether these factors, individually or interactively, significantly influenced user expectations. Moreover, the effects of system interface, system feedback and system responsiveness on user comfort in AI-based systems were determined as well. The findings highlighted the importance of prior experience in shaping user expectations. It also suggests that educational level may have limited influence on user expectations. The choice of system interface and the responsiveness of the AI-based system significantly impact user comfort. The findings suggest for the creation of more user-friendly and comfortable interfaces. Understanding the various factors that influence user comfort and expectation, can aid the design and development of AI systems tailored to user backgrounds that better meet user needs and enhance their overall experience.

show abstract

mentioning

confidence: 81%

Understanding User comfort and Expectations in AI-based Systems

Dave¹,

Saxena

Jha³

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…However, existing CoT works provide only a simple instruction like "let's do something step by step" and cannot handle intricate tasks. In contrast, our method is AI chain-based [37], [36], [58], which interacts with the model in explicit steps to generate CFGs. While the idea of AI chain has been explored for writing assistants [37], our AI chain involves much more complex task analysis and data flow for a domain-specific CFG generation task.…”

Section: Related Workmentioning

confidence: 99%

“…which cannot handle complex tasks. In contrast, our approach is based on an AI chain [36], [50], [51] that interacts with the model to reuse partial code. While the idea of an AI chain has been explored for writing assistants [36], our AI chain involves complex task analysis and data flow for domainspecific partial code reuse.…”

Section: Related Workmentioning

confidence: 99%

Prompt-tuned Code Language Model as a Neural Knowledge Base for Type Inference in Statically-Typed Partial Code

Huang

Yuan

Xing

et al. 2022

Proceedings of the 37th IEEE/ACM International Conference on Automated Software Engineering

View full text Add to dashboard Cite

Control Flow Graphs (CFGs) are essential for visualizing, understanding and analyzing program behavior. For statically-typed programming language like Java, developers obtain CFGs by using bytecode-based methods for compilable code and Abstract Syntax Tree (AST)-based methods for partially uncompilable code. However, explicit syntax errors during AST construction and implicit semantic errors caused by bad coding practices can lead to behavioral loss and deviation of CFGs. To address the issue, we propose a novel approach that leverages the error-tolerant and understanding ability of pre-trained Large Language Models (LLMs) to generate CFGs. Our approach involves a Chain of Thought (CoT) with four steps: structure hierarchy extraction, nested code block extraction, CFG generation of nested code blocks, and fusion of all nested code blocks' CFGs. To address the limitations of the original CoT's single-prompt approach (i.e., completing all steps in a single generative pass), which can result in an "epic" prompt with hard-to-control behavior and error accumulation, we break down the CoT into an AI chain with explicit sub-steps. Each sub-step corresponds to a separate AI-unit, with an effective prompt assigned to each unit for interacting with LLMs to accomplish a specific purpose. Our experiments confirmed that our method outperforms existing CFG tools in terms of node and edge coverage, especially for incomplete or erroneous code. We also conducted an ablation experiment and confirmed the effectiveness of AI chain design principles: Hierarchical Task Breakdown, Unit Composition, and Mix of AI Units and Non-AI Units. Our work opens up new possibilities for building foundational software engineering tools based on LLMs, as opposed to traditional program analysis methods.

show abstract

“…While, e.g., explanations on the basis of "counterfactuals" [81] may be well suited for testing hypotheses, more research needs to examine how larger numbers of, e.g., setting possibilities affect the interaction. In the exemplary case of generative visual models, the cognitive load of the user increases with the number of adjustable settingswithout a significant effect on performance [31]. Furthermore, it must be considered whether and which additional information is displayed e.g., in a training context or in a daily use context since these may differ considerably with respect to the available time and cognitive resources.…”

Section: Fit Of Performance Measures and Subjective Measures In Xaimentioning

confidence: 99%

How do Users Experience Traceability of AI Systems? Examining Subjective Information Processing Awareness in Automated Insulin Delivery (AID) Systems

Schrills¹,

Franke²

2022

Preprint

View full text Add to dashboard Cite

When interacting with artificial intelligence (AI) in the medical domain, users frequently face automated information processing, which can remain opaque to them. For example, users with diabetes may interact daily with automated insulin delivery (AID). However, effective AID therapy requires traceability of automated decisions for diverse users. Grounded in research on human-automation interaction, we study Subjective Information Processing Awareness (SIPA) as a key construct to research users' experience of explainable AI. The objective of the present research was to examine how users experience differing levels of traceability of an AI algorithm. We developed a basic AID simulation to create realistic scenarios for an experiment with N = 80, where we examined the effect of three levels of information disclosure on SIPA and performance. Attributes serving as the basis for insulin needs calculation were shown to users, who predicted the AID system's calculation after over 60 observations. Results showed a difference in SIPA after repeated observations, associated with a general decline of SIPA ratings over time. Supporting scale validity, SIPA was strongly correlated with trust and satisfaction with explanations. The present research indicates that the effect of different levels of information disclosure may need several repetitions before it manifests. Additionally, high levels of information disclosure may lead to a miscalibration between SIPA and performance in predicting the system's results. The results indicate that for a responsible design of XAI, system designers could utilize prediction tasks in order to calibrate experienced traceability.

show abstract

GANSlider: How Users Control Generative Models for Images using Multiple Sliders with and without Feedforward Information

Cited by 32 publications

References 26 publications

Understanding User comfort and Expectations in AI-based Systems

Understanding User comfort and Expectations in AI-based Systems

Prompt-tuned Code Language Model as a Neural Knowledge Base for Type Inference in Statically-Typed Partial Code

How do Users Experience Traceability of AI Systems? Examining Subjective Information Processing Awareness in Automated Insulin Delivery (AID) Systems

Contact Info

Product

Resources

About