Bridging knowledge gaps in digital forensics using unsupervised explainable AI
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Artificial Intelligence (AI) has found multi-faceted applications in critical sectors including Digital Forensics (DF) which also require eXplainability (XAI) as a non-negotiable for its applicability, such as admissibility of expert evidence in the court of law. The state-of-the-art XAI workflows focus more on utilizing XAI tools for supervised learning. This is in contrast to the fact that unsupervised learning may be practically more relevant in DF and other sectors that largely produce complex and unlabeled data continuously, in considerable volumes. This research study explores the challenges and utility of unsupervised learning-based XAI for DF's complex datasets. A memory forensics-based case scenario is implemented to detect anomalies and cluster obfuscated malware using the Isolation Forest, Autoencoder, K-means, DBSCAN, and Gaussian Mixture Model (GMM) unsupervised algorithms on three categorical levels. The CIC MalMemAnalysis-2022 dataset's binary, and multivariate (4, 16) categories are used as a reference to perform clustering. The anomaly detection and clustering results are evaluated using accuracy, confusion matrices and Adjusted Rand Index (ARI) and explained through Shapley Additive Explanations (SHAP), using force, waterfall, scatter, summary, and bar plots' local and global explanations. We also explore how some SHAP explanations may be used for dimensionality reduction.
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.
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.004 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.003 | 0.015 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| 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 it