Vareon Inc., Irvine, California, USA
April 2026
Abstract
MatterSpace Vital was evaluated under a sealed-target blind rediscovery contract spanning three longevity-relevant single-cell challenge families: senescence clearance, transient epigenetic reprogramming, and muscle stem-cell rejuvenation. The benchmark contract is visible, but the exact target remains sealed until post-hoc evaluation. Under the current frozen benchmark package, MatterSpace Vital recovered the sealed target exactly at rank 1 in all three families: GSM8297138, Aged_SO, and 15-20. The public baseline missed the exact hidden target in all three cases, even when neighborhood similarity remained high. This is computational benchmark evidence rather than wet-lab or clinical proof, but it supports the narrower claim that MatterSpace Vital can recover exact hidden targets under a strict longevity blind-rediscovery contract.
Open scoring can reward answers that look young-like, match part of a desired transcriptional program, or remain close to the correct hidden answer without actually recovering it. Blind rediscovery is stricter. The question is not whether a system can produce something plausible after the fact. The question is whether it can recover the sealed target when the target is withheld during generation.
That distinction is especially important in longevity research, where post-hoc storytelling is easy and exact hidden-target recovery is harder. The MatterSpace Vital publication is therefore framed around the exact-hit contract rather than around vague similarity alone.
| Challenge family | Sealed target | System top-1 | Baseline top-1 | Exact hit |
|---|---|---|---|---|
| Senescence clearance | GSM8297138 | GSM8297138 | GSM8297143 | Yes |
| Epigenetic reprogramming | Aged_SO | Aged_SO | Aged_O | Yes |
| Stem-cell rejuvenation | 15-20 | 15-20 | 55-60 | Yes |
The current frozen summary therefore records 3/3 exact blind rediscoveries for MatterSpace Vital and 0/3 exact hits for the public baseline.
The public baseline was not trivial. In the epigenetic and stem-cell settings it remained numerically close on similarity metrics, but it still failed the exact-hit gate. That is precisely what the benchmark is meant to reveal. A system can approach the right neighborhood without recovering the right answer.
MatterSpace Vital's result is therefore best understood as exact sealed-target recovery under a strict benchmark contract across three independent challenge families, not merely as a soft-score improvement over a weak baseline.
The publication is backed by a dedicated blind_rediscovery/ package in the MatterSpace Vital repository. The package includes a machine-readable frozen manifest, generated paper assets, and rerun scripts for local and private-HF execution paths. This keeps the public summary tied to inspectable benchmark evidence rather than to manually copied results.
blind_rediscovery/results/frozen/blind_rediscovery/paper/MatterSpace Vital now carries a public blind-rediscovery result on the Vareon website alongside MatterSpace Lattice. The important claim is narrow and testable: under the current frozen benchmark package, the system recovered the exact sealed target at rank 1 in all three evaluated longevity challenge families while the public baseline did not.