9e38f78379
Two changes per the user's feedback that the slide had unused horizontal space and needed per-PDF context. Layout - The reference scene is now a 2-column grid inside the metric-stack: PDF iframe at ~1.7fr on the left, description panel at ~0.55fr on the right (min 280px). On narrow viewports (<1100px) it falls back to a vertical stack with the description capped to 240px. - Added #zoom=page-width to the iframe URL so the PDF's page fits its column width instead of leaving margins beside an 8.5x11 page rendered in a wider iframe. - Hide the prose card on the references scene — the description panel inside the stack covers what the prose was saying, and freeing the right edge gives the description proper room. Description content - Backend reads <stem>.md sidecar files alongside each PDF and returns the contents in the /api/references payload. - Frontend renders them with a tiny built-in markdown subset (headings, bold/italic, lists, inline code, paragraphs) — no third-party renderer dependency. - Initial draft sidecar .md files committed for the four PDFs currently in references/. Each describes how the paper informs a specific scene of the deck (which model row, which eval protocol, which channel selection). Edit them in place and the panel updates on the next reload.
26 lines
1.1 KiB
Markdown
26 lines
1.1 KiB
Markdown
# LSTM on event-log sequences
|
|
|
|
DANTE applies a **plain LSTM directly to system-log event sequences**
|
|
to flag insider-threat behavior. Earlier in the literature than the
|
|
transformer wave, and useful here as a methodological baseline.
|
|
|
|
## What we borrowed
|
|
|
|
- **Evidence that simple recurrent models are enough.** The paper
|
|
shows an LSTM on sequence-of-events alone — no per-task feature
|
|
engineering — captures enough temporal structure to beat
|
|
bag-of-events classifiers. That's the empirical ground for the
|
|
*RNN/GRU/LSTM* entries in our model comparison being plain, not
|
|
bespoke.
|
|
- **Negative-evidence framing.** DANTE is also explicit about cases
|
|
where the LSTM under-performs (low-volume users, novel event
|
|
types). Informs the *split-by-sample, not split-by-time* eval
|
|
protocol on the perf scene — generalising to unseen actors is
|
|
the bar.
|
|
|
|
## Where it differs
|
|
|
|
- Operates on log-event token sequences (categorical), not numeric
|
|
resource metrics (continuous). Our channels are floats from
|
|
`/proc`, so we use the temporal structure DANTE validates without
|
|
inheriting the embedding setup.
|