Explanation and ontological unpacking in AI-Assisted Domain Modeling: Connecting Human, Symbolic, and Subsymbolic Intelligence

作者
Dana Kulbayeva
出处
期刊:CERN European Organization for Nuclear Research - Zenodo [European Organization for Nuclear Research]
标识
DOI:10.5281/zenodo.21130659
摘要

The integration of artificial intelligence into domain modeling introduces a fun- damental tension between the efficiency of automated knowledge extraction and the seman- tic accountability that meaningful conceptual modeling requires. When modelers delegate modeling tasks to AI systems, they risk losing value in creation, which, among other things, represents the deep understanding of domain semantics that emerges through the process of constructing a model itself. At the same time, modelers do not get a sufficient explanation of what the AI has done and why. This thesis proposes that explanation in AI-assisted domain modeling must be understood as a structural property of the process through which human intelligence, symbolic AI, and subsymbolic AI together construct and maintain a shared conceptual model. Drawing on the notion of ontological unpacking, the thesis develops two theoretical contributions: the Three- Intelligence Domain Modeling Explanation Framework, which identifies a qualitatively dis- tinct mode of explanation for each of the four directions of interaction between the three intelligence types and the conceptual model; and the Three-Intelligence Domain Modeling Process Framework, which sequences these intelligences across four lifecycle phases: domain elicitation, symbolic structuring, subsymbolic validation, and collaborative convergence. The framework is partially grounded through the analysis of the three existing systems: DoMoRe. DoMoBOT, and the superimposition method, whose complementary coverage confirms that genuine explainability requires the coordinated operation of all three intelligence types across the full modeling lifecycle.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
大模型的应助被小燕尾条采纳,获得10
1秒前
zhu发布了新的文献求助10
1秒前
情怀的应助被咎如天采纳,获得10
1秒前
molihuakai的应助被HJH采纳,获得10
1秒前
2秒前
2秒前
2秒前
啦啦啦发布了新的文献求助10
3秒前
LIGE发布了新的文献求助10
4秒前
4秒前
虎虎发布了新的文献求助10
5秒前
Dht完成签到,获得积分10
6秒前
的法国队发布了新的文献求助10
7秒前
wsx4321发布了新的文献求助10
8秒前
9秒前
Lee发布了新的文献求助10
9秒前
酷波er的应助被kuba采纳,获得10
9秒前
9秒前
yu完成签到,获得积分20
10秒前
Yogita完成签到,获得积分10
10秒前
Eli发布了新的文献求助10
11秒前
11秒前
独特振家完成签到,获得积分10
11秒前
11秒前
12秒前
妮妮的马甲裙完成签到 ,获得积分10
13秒前
朴实流沙完成签到,获得积分10
13秒前
13秒前
十年晨曦完成签到 ,获得积分10
14秒前
zz发布了新的文献求助10
15秒前
15秒前
16秒前
ttt关注了科研通微信公众号
17秒前
muzi完成签到,获得积分10
17秒前
六瓶瓶发布了新的文献求助10
18秒前
JamesPei的应助被孤独听雨的猫采纳,获得10
18秒前
18秒前
伊布完成签到,获得积分10
18秒前
19秒前
19秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
John McDowells Bildungstheorie 800
八旗制度の研究 600
plantTFDB 508
Resource Recovery Handbook 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8175142
求助须知:如何正确求助?哪些是违规求助? 9566928
关于积分的说明 21789123
捐赠科研通 7769522
什么是DOI,文献DOI怎么找? 3431822
关于科研通互助平台的介绍 2556078
邀请新用户注册赠送积分活动 2477222