Compression as a structure detector.
Our research thesis — the FLOOR map — is that a single compression methodology, run across many unrelated real domains, reveals which generating primitive each one reduces to, and where each hits its entropy floor: the boundary between determinism and chance. Validated, lossless, and honest about where it finds nothing.
Language family tree from raw bytes
Run the compressor pairwise over raw text in many languages and the language family tree falls out — relatedness rebuilt with no linguistics, just bytes. (It sees surface form: the same language in two alphabets looks far apart until you romanize it.)
The edge of chaos, located
Sweep a family of systems from orderly to random and compression pinpoints the thin band where complexity lives — universal computation (Rule 110) sits at a measured ~0.47 bits per cell, between the trivial and the noise.
Rule-governed games win, statistical style floors
Chess — governed by rules — compresses far past a general compressor, because the rule is the short description. Bach, governed by statistical style rather than a closed rule, floors near the general-codec limit. The method tells the two apart.
High-entropy data floors honestly
Protein, DNA and earthquake records floor near 0% gain — they're close to incompressible, and the detector is not fooled into a false signal. Sunspots, by contrast, reveal a linear recurrence (1.13×). The honesty is the result.
Columns that are math, stored as math (published, not patented)
When a column is exactly a function of others — total = price + tax, or an additive counter of two flow fields — AT-1 discovers the relationship from the data itself and stores the formula instead of the column, rebuilding it bit-for-bit (measured 8.27% on real NetFlow). We publish this openly rather than patent it: discovering exact arithmetic dependencies for lossless columnar compression is a known idea, and our edge is shipping it verified and reversible inside a queryable, erasable container — not the idea alone.
Why this matters for a compression company
The same discipline that finds the shortest description of your data is, in effect, a detector for hidden structure. It's why AT-1 wins where data is rule-governed (logs, ticks, genomics, games) and honestly delegates where it isn't (natural media at its entropy floor). The research isn't a side project — it's the same lens, pointed at the world instead of a single file.
Read the papers
A minimum-description-length instrument for measuring the determinate-to-stochastic ratio of structure across scientific domains. Dylan Wolpe, 2026. The full FLOOR-map methodology and its results across 11+ unrelated real domains — including the honest negatives.
Open-access · DOI: 10.5281/zenodo.20811282
Read itDerived-Column Elimination by Exact Functional-Dependency Discovery: a Verified, Reversible Pre-Pass for Lossless Columnar Compression. Dylan Wolpe, 2026. The derived-column method — published openly as prior art, not patented, with the real-data result.
Open-access · DOI: 10.5281/zenodo.20848308
Read itThe instrument, pointed at eleven domains
Each of these is the same measurement — how much structure a compressor can find, against a null that preserves the boring properties of the data — aimed somewhere new. Where it fails, we say so.
Determinism
At1 determinism returns a single calibrated [0,1] score per file: how much of it is rule-governed structure versus irreducible noise, comparable across data types.
Discover
Point it at a sequence and it recovers the generator that produced it — state recurrences, index-functions, cellular automata, compositions…
Docs & commands →Explain
Point AT-1 at an integer sequence and it searches a program space in shortest-description-first order for the exact generator…
Docs & commands →Fraud Fingerprints
Give every cashier, pump, terminal or account its own compression model.
Intel
Instead of compress(file), ask what a file IS.
MDL-Select
MDL-Select scores whether a fit is a genuine law or a memorization by description length in bits…
Ncd
At1 ncd uses normalized compression distance to measure how similar two files are without parsing them: a code-clone / near-duplicate scan and a clause/document…
Observe
AT-1 Observe scores every log line, metric, or netflow row by how many bits it costs against a model of normal traffic.
Risk Tape
A Compression-VIX: AT-1 reads the structure of an order-flow tape and turns how compressible it is into a model-free stress index that tracks realized volatility…