Ruth 2025
Last edited: August 8, 2025One-Liner
Governments’ security guidance lack overall consensus in both formatting as well as actual practice; we find this out using treeeeees.
Novelty
- new tree similarity mechanism similar to edit distance
- first hand-labeled ontological dataset for security guidance
Notable Methods
- download security guidance
- label guidance into multi-level controls
- tree distance metric for similarity
SA
Last edited: August 8, 2025an operations research linear solution system
Sadeghian 2021
Last edited: August 8, 2025DOI: 10.3389/fcomp.2021.624594
(Sadeghian, Schaffer, and Zahorian 2021)
One-Liner
Using a genetic algorithm, picked features to optimize fore; achieved \(94\%\) with just MMSE data alone (ok like duh me too). Developed ASR tool to aid.
Novelty
- Developed an ASR methodology for speech, complete with punctuations
- Used a genetic algorithm to do feature selection; NNs performed worse because “space is smaller???”
Notable Methods
Used a GRU to insert punctuations

The paper leveraged the nuke that is a bidirectional GRU, ATTENTION,
Safety for Autonomous Trucking
Last edited: August 8, 2025challenges of autonomous driving
the world is messy—robo-taxies would be confused
It seems like an increasing trend to do data argumentation with generative approaches.
safety standards
ISO 26262
how to do it , instead of what to do
- functional safety standard
- methods about testing and processes
ISO 21448
intended functionality
- scenario based analysis of hazardous situations—“triggering conditions” => “mitigations” shown below
- identification and mitigation of functional insufficiencies
- (requires?) the discovering of unknown unknowns with mitigations
Safety and AI
the actual safety challenges
SAIC: Speech Anonomyzation
Last edited: August 8, 2025Challenge of speech anonymization: cannot develop a model which both preserves speech features well but also effectively anonymizes the speech.
Methodology
- Separate content and speech encoders
- Results in highly concentrated + effective speech content, but with very widespread voiceprint
