國立虎尾科技大學 |

Human-AI Interface Layers : = Enhancing Communication of Intent for AI-Assisted Creative Pursuits and Social Experiences.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Human-AI Interface Layers :/
Reminder of title:	Enhancing Communication of Intent for AI-Assisted Creative Pursuits and Social Experiences.
Author:	Louie, Ryan.
Description:	1 online resource (220 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-06, Section: B.
Contained By:	Dissertations Abstracts International85-06B.
Subject:	Computer science. -
Online resource:	click for full text (PQDT)
ISBN:	9798381174663

Human-AI Interface Layers : = Enhancing Communication of Intent for AI-Assisted Creative Pursuits and Social Experiences.
Louie, Ryan.

Human-AI Interface Layers :Enhancing Communication of Intent for AI-Assisted Creative Pursuits and Social Experiences. - 1 online resource (220 pages)

Source: Dissertations Abstracts International, Volume: 85-06, Section: B.

Thesis (Ph.D.)--Northwestern University, 2023.

Includes bibliographical references

While early uses of artificial intelligence (AI) aimed to automate repetitive and burdensome tasks, AI shows great potential in assisting users in human domains of personal meaning and importance, from creative pursuits to social experiences. These human domains require that AI have an understanding and sensitivity to a user's personal ideas and knowledge. To assist a novice creator who lacks formal training, a generative AI partner must understand the person's creative ideas. An AI agent can appropriately identify opportune moments to connect with friends/colleagues from afar, only if it has cultural and tacit knowledge of the situations and activities in which people desire to connect. Despite significant advances in AI capabilities, conventional interfaces for AI can make it challenging for people to communicate their ideas and expectations. This dissertation proposes a crucial layer in an AI-powered application's stack called the Human-AI Interface Layer, which reconfigures existing AI capabilities to empower people's effective communication.They provide (1) intuitive constructs for communicating ideas in ways AI systems can act upon and (2) interactive tools for forming, reflecting upon, and clarifying their intentions.I developed three Interface Layers to enhance communication of intent to existing AI capabilities. First, users struggle to convey fine-grained semantics and control pieces of a generated artifact, when the AI's interface only allows indirect influence over how the entire artifact is generated. To resolve this, I propose a Steering Interface Layer that partitions and constrains generated outputs to support incrementally guiding AI outputs toward desired directions. Second, when a person's ideas for their AI-enabled creation are conceptually-rich and sometimes vague, people can struggle to translate these high-level concepts to the lower-level constructs for AI systems to understand and act upon. As such, I contribute an Expression Interface Layer equipped with cognitive bridging tools that help people address this semantic gap by fleshing out their high-level concepts and foraging for relevant and precisely operating AI constructs. Third, when an individual provides explicit instructions to an AI agent, the agent's actions during execution can still disappoint when it encounters constraints that prevent implicit expectations from being satisfied. To resolve this, I developed an Execution Interface Layer for recognizing and stating implicit expectations so an AI agent can appropriately adjust after handoff.More generally, a mismatch might always exist between what people envision for human pursuits and what AI assistants can currently understand about them. This thesis shows that Human-AI Interface Layers for intention formation and articulation can empower creators and designers to communicate their ideas and intentions for AI assistance, despite existing AI's inherently limited interfaces for doing so.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798381174663Subjects--Topical Terms:

573171
Computer science.
Subjects--Index Terms:

Context-aware computingIndex Terms--Genre/Form:

554714
Electronic books.

Human-AI Interface Layers : = Enhancing Communication of Intent for AI-Assisted Creative Pursuits and Social Experiences.
LDR:04416ntm a22003977 4500 001 1142990
005 20240513061040.5
006 m o d
007 cr mn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798381174663
035 $a (MiAaPQ)AAI30639437
035 $a AAI30639437
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Louie, Ryan. $3 1467542
245 1 0 $a Human-AI Interface Layers : $b Enhancing Communication of Intent for AI-Assisted Creative Pursuits and Social Experiences.
264 0 $c 2023
300 $a 1 online resource (220 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-06, Section: B.
500 $a Advisor: Zhang, Haoqi.
502 $a Thesis (Ph.D.)--Northwestern University, 2023.
504 $a Includes bibliographical references
520 $a While early uses of artificial intelligence (AI) aimed to automate repetitive and burdensome tasks, AI shows great potential in assisting users in human domains of personal meaning and importance, from creative pursuits to social experiences. These human domains require that AI have an understanding and sensitivity to a user's personal ideas and knowledge. To assist a novice creator who lacks formal training, a generative AI partner must understand the person's creative ideas. An AI agent can appropriately identify opportune moments to connect with friends/colleagues from afar, only if it has cultural and tacit knowledge of the situations and activities in which people desire to connect. Despite significant advances in AI capabilities, conventional interfaces for AI can make it challenging for people to communicate their ideas and expectations. This dissertation proposes a crucial layer in an AI-powered application's stack called the Human-AI Interface Layer, which reconfigures existing AI capabilities to empower people's effective communication.They provide (1) intuitive constructs for communicating ideas in ways AI systems can act upon and (2) interactive tools for forming, reflecting upon, and clarifying their intentions.I developed three Interface Layers to enhance communication of intent to existing AI capabilities. First, users struggle to convey fine-grained semantics and control pieces of a generated artifact, when the AI's interface only allows indirect influence over how the entire artifact is generated. To resolve this, I propose a Steering Interface Layer that partitions and constrains generated outputs to support incrementally guiding AI outputs toward desired directions. Second, when a person's ideas for their AI-enabled creation are conceptually-rich and sometimes vague, people can struggle to translate these high-level concepts to the lower-level constructs for AI systems to understand and act upon. As such, I contribute an Expression Interface Layer equipped with cognitive bridging tools that help people address this semantic gap by fleshing out their high-level concepts and foraging for relevant and precisely operating AI constructs. Third, when an individual provides explicit instructions to an AI agent, the agent's actions during execution can still disappoint when it encounters constraints that prevent implicit expectations from being satisfied. To resolve this, I developed an Execution Interface Layer for recognizing and stating implicit expectations so an AI agent can appropriately adjust after handoff.More generally, a mismatch might always exist between what people envision for human pursuits and what AI assistants can currently understand about them. This thesis shows that Human-AI Interface Layers for intention formation and articulation can empower creators and designers to communicate their ideas and intentions for AI assistance, despite existing AI's inherently limited interfaces for doing so.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Computer science. $3 573171
650 4 $a Communication. $3 556422
653 $a Context-aware computing
653 $a Generative AI
653 $a Human-AI communication
653 $a Human-AI interaction
653 $a Human-AI interface layers
655 7 $a Electronic books. $2 local $3 554714
690 $a 0984
690 $a 0459
690 $a 0800
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Northwestern University. $b Technology and Social Behavior. $3 1193212
773 0 $t Dissertations Abstracts International $g 85-06B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30639437 $z click for full text (PQDT)