Collections
Curated reading lists. Each collection is a coherent selection of papers around a theme — a problem, a method, or a research thread. Use these instead of the full paper archive unless you already know exactly what you're looking for.
◆
Start with the Latent
Three papers that introduce the core framework
If you want to understand what the research program is actually about, read these three papers in order. They define the Latent framework, show how proof discovery works as neural architecture search, and demonstrate the…
3 papers →
★
Millennium & Smale Problems
Attacks on six open problems
Formal attempts at problems from the Clay Millennium Prize list and Smale's 1998 list. Each paper has a Lean-verified proof skeleton. These are ongoing research — the papers present proof strategies and verified componen…
6 papers →
✓
Machine-Verified Proofs
Papers with Lean 4 formal verification
Papers whose central claims have been formalized and checked in Lean 4. These are the highest trust-level items in the corpus — the math has been re-verified by a proof assistant, not just reviewed by humans.
20 papers →
▲
AI: Safety, Scaling, Interpretability
Machine learning results with formal guarantees
Results about large language models, neural scaling, adversarial robustness, and AI safety — all framed as mathematical theorems with Lean-verified components. Some of these rederive known empirical scaling laws from fir…
8 papers →
$
Quantitative Finance
Exact pricing, risk, and distributions
Closed-form results and machine-verified theorems for financial mathematics — rough volatility, expected shortfall, spectral importance sampling, and exact distribution theory.
8 papers →
π
Number Theory & the Riemann Hypothesis
Paths toward RH and related prime distribution results
Three distinct analytic paths toward RH: via Euler product smoothness, via Fourier–Euler symmetry, and via GUE moment positivity. Each paper stands on its own and together they form a parallel-strategy research program.
4 papers →
⚙
Method & Meta-Theory
How the research program works
Papers that describe the methodology itself: proof discovery as neural search, cross-domain bridges, the Latent framework, and how results compose. Read these if you want to understand the overall approach before diving …
4 papers →
◷
Latest Work
Recently created or revised papers
Papers that have been added or substantially updated recently. Ordered by creation date. Maturity varies — some are polished, some are early drafts still being developed.
20 papers →
■
Most Mature
Working papers, fully developed
Papers classified as Working Papers in the maturity taxonomy — over 4000 words, Lean-verified or human-reviewed, ready to be cited or shared. The trust-layer entry points.
20 papers →
∴
Proof-Heavy
Papers with many verified theorems
Papers where most of the content is formal mathematics — theorem statements, proof sketches, and verified claims. Use these as a reference corpus, not as a narrative introduction.
15 papers →
Looking for a specific paper? Try the full paper index, the domains view, or the connection graph.