Home |  DEUTSCH |  Contact |  Imprint |  Data Protection |  Login |  KIT

Thema4483/en

Aus Aifbportal

Wechseln zu: Navigation, Suche


Artificial Intelligence: Unblackboxing Artificial Intelligence - Towards a Definition



Information on the Thesis

Type of Final Thesis: Bachelor, Master
Supervisor: Scott ThiebesKonstantin Pandl
Research Group: Critical Information Infrastructures

Archive Number: 4.483
Status of Thesis: Open


Further Information

Background:

Artificial intelligence (AI) and machine learning are among today’s most disruptive technologies. Despite the significant advances in research and industry in recent years (e.g., deep Learning in general or AlphaGo), the field is still defined very vaguely. This starts with the basic and interdisciplinary question of what intelligence and what AI actually are. Extant approaches to define AI appear to be fuzzy and different fields such as computer science, psychology, philosophy, and biology may have different answers. Diving deeper, even more questions arise: What is machine learning? Is AI possible without machine learning? How is deep learning different from machine learning? Can you perform deep learning without using neural networks? What are limitations of AI, can AI surpass human intelligence? How does AI differ from terms such as computational intelligence, soft computing, heuristics, and others?


Objective(s):

Possible topics include, but are not limited to:

  • Review of the intelligence and AI literature to create a consolidated definition
  • Review of boundaries and interrelations of terms such as AI, machine learning, computational intelligence, deep learning, heuristics with the goal of a clear definition and differentiation
  • Review of the current AI and machine learning literature to evaluate various algorithms towards their fit with different AI and machine learning definition concepts

This is an umbrella topic since topics of interest change rapidly. Students are encouraged to propose a topic that is of interest to them within the topic area. The thesis allows you to gain a broad and deep knowledge in artificial intelligence and to make a significant contribution towards a scientifically sound fundament.


Introductory literature:

Dobrev, Dimiter (2012): A Definition of Artificial Intelligence. In arXiv preprint arXiv:1210.1568. Available online at https://arxiv.org/pdf/1210.1568.pdf.

Goodfellow, Ian; Bengio, Yoshua; Courville, Aaron (2016): Deep learning: MIT press. Available online at https://www.deeplearningbook.org/.

Legg, Shane; Hutter, Marcus; others (2007): A collection of definitions of intelligence. In Frontiers in Artificial Intelligence and applications 157, p. 17. Available online at https://arxiv.org/pdf/0706.3639.pdf.