Bibliographic record
Abstract
This thesis introduces new approaches for enhancing the security and privacy of client-server commu- nications by addressing two critical issues. First, we investigate the risks posed by user-impersonation attacks, where client-side malware forges or tampers with web requests, initiating security-sensitive actions such as financial transactions or voting without the user’s true intent. Current defenses, including user authentication and network encryption, fall short because they assume the integrity of the client device. We present vWitness, a system that certifies user intent by “witnessing” client interactions with the web page, ensuring that outgoing requests align with the interactions that the user made with the help of specifications set by the server. Using computer vision, vWitness distin- guishes benign rendering variations from potential malicious alterations, achieving 99.97% accuracy with minimal performance impact, adding only 197ms on average to the interaction session. This solution enables web servers to verify that requests truly originate from a user’s intended actions. Additionally, we address privacy challenges arising from web tracking, where existing tools strug- gle to balance effective tracker blocking with site compatibility. Many tools rely on manual updates and user feedback to resolve “breakages” caused by overblocking, where essential site functional- ity is inadvertently blocked. To address this, we propose Duumviri, an automated approach that enhances tracking detection by integrating a machine learning-based breakage detection model. Du- umviri improves the accuracy of tracker detection by using differential features of web requests and incorporates breakage detection to mitigate overblocking. This method is particularly novel in han- dling “mixed” trackers, trackers that blend tracking with functional information into a single request, by enabling analysis at partial-request granularity to distinguish between tracking components and functional ones. Evaluations of Duumviri demonstrate high accuracy in detecting non-mixed trackers (97.44%) and mixed trackers (74.19%). Together, vWitness and Duumviri represent significant advancements in web security and privacy. vWitness introduces an innovative method for certifying user intent to counter user-impersonation attacks, while Duumviri provides an automated framework for robust tracker detection and breakage management to safeguard user privacy without compromising functionality.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".