Tiwari, Raghav (2025) Reverse-Engineering Expert Tetris Play: Identifying Core Visual-Spatial Cues. Bachelor's Thesis, Artificial Intelligence.
|
Text
RaghavTiwariBachelorThesisFinal.pdf Download (3MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
Toestemming.pdf Restricted to Registered users only Download (157kB) |
Abstract
This thesis investigates the question of how humans make rapid, high-pressure decisions. It examines this by asking whether above-average Tetris play can be reverse-engineered into a small set of visual-spatial features. In Experiment 1, four new feature sets, each motivated by a distinct cognitive theory were optimised via the Cross-Entropy Method and compared against a well-established baseline set known to emulate human play. Although none matched the baseline's performance, they revealed three key heuristics: hole avoidance, skyline smoothness, and an explicit incentive for high-value line clears. Building on these insights, Experiment 2 evaluated nine new feature sets that integrated these pillars in different ways. Three of these achieved up to 77\% of the baseline's mean score and at least 80\% of its criterion score, while closely matching its line clearing pattern. Together, these results suggest a generalizable framework for reverse-engineering expert strategies in dynamic domains.
| Item Type: | Thesis (Bachelor's Thesis) |
|---|---|
| Supervisor name: | Sibert, C.L. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Bachelor's Thesis |
| Language: | English |
| Date Deposited: | 28 Jul 2025 10:52 |
| Last Modified: | 28 Jul 2025 10:52 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/36556 |
Actions (login required)
 |
View Item |