Building a sound layer for AI safety. Our new thesis proves that pixel-based agents can be verifiably secure using probabilistic formal methods. 🛡️💻 #AISafety #FormalMethods #NeuroSymbolicnp
1
69
Search
極端流VDMチャンネル
第6回 HTTPの向こうにVDM!?
を公開しました。
youtu.be/R_DJAvOzRNM
#FormalMethods #VDM #ViennaTalk #WeblyWalkThrough
1
1
387
GPT‑5.4 Pro reportedly produced a proof for a frontier math problem (Epoch; HN 47497757). No machine‑checked artifact yet—Lean/Coq formalizers must validate. Will they? #math #FormalMethods
1
15
The "Formal Foundations of the Agent Engineering Standard (AES)" provides a mathematical framework for compliant, observable, and safe autonomous agents using LTL, predicate abstraction, and Lean 4 mechanized proofs.
#AI #FormalMethods #AES #Robotics #LLMs
1
100
FormalEvolve: Neuro-Symbolic Evolutionary Search for Diverse and Prover-Effective Autoformalization
Preprint: FormalEvolve is a neuro-symbolic framework for autoformalization, gene…
arxiv.org/abs/2603.19828 #AI #FormalMethods #MachineLearning #Preprint #Arxiv #ScienceNews
FormalEvolve: Neuro-Symbolic Evolutionary Search for Diverse and...
Autoformalization aims to translate natural-language mathematics into compilable, machine-checkable statements. However, semantic consistency does not imply prover effectiveness: even semantically...
From arxiv.org 2
26
Introducing OSSA: A mathematical framework mapping contraction theory & ISS small-gain theorems to robotics software patterns for provably stable cognitive systems. Dissertation by Ouadi Maakoul. #AI #Robotics #FormalMethods
1
85
Early signal: ≈10^4–10^5 lemmas in mathlib — Leanstral aims to emit machine-checked proofs via Lean. What happens to your CI/CD when agents must pass verifiers before merge? #Lean #FormalMethods
11
New Preprint: Arrow's theorem. Sen's paradox. What does "impossible" mean computationally?
Lean 4 + SAT solvers + LRAT replay → an auditable evidence bundle. Every verdict is replayable. No blind trust.
🗺️ zenodo.org/records/189815…
#Logic #FormalMethods #SocialChoice
39
Stop treating LLMs like chatbots. #FormalMethods aren't dead. Specs achieve 10x #CodeCompression vs natural language prompts without losing test coverage. Is this the #SpecDrivenDev shift we needed? banandre.com/blog/specifica…
12
Refinement Modeling and Verification of RISC-V Assembly using Knuckledragger
philipzucker.com/refine_assembl… #assembly #riscv #formalmethods #python with video youtube.com/watch?v=NQGh8r…
3
29
1.4K
Contact Us
softfm@comit2026.org
Please Visit Us Link
allconferencecfpalerts.com/cfp/view.php?e…
#SoftwareEngineering #FormalMethods #TechInnovation #SoftwareDevelopment #Programming #CodeQuality #EngineeringExcellence #TechTrends #SoftwareTesting #computerscienceengineering
3
極端流VDMチャンネル
第5回 - VDMでUIプロトタイプを動かしてみよう
を公開しました。
youtube.com/watch?v=kGc6zG…
#VDM #FormalMethods #ViennaTalk #LivelyWalkThrough
3
4
801
The worst part about claiming “CORRECTNESS” is that you can’t stop when you’re at 99.95% coverage.
#formalmethods #neuraltheoremproving #AI4Math
4
Ever wonder why installing software can break everything? A novel formal model of package manager dependency resolution breaks it down scientifically. Discussions are...
arxiv.org/abs/2602.18602
#FormalMethods #SoftwareEngineering #AcademicTwitter #ComputerSc
Package Managers à la Carte: A Formal Model of Dependency Resolution
Package managers are legion. Every programming language and operating system has its own solution, each with subtly different semantics for dependency resolution. This fragmentation prevents...
From arxiv.org 51
🧮 Nuestras primeras pruebas de demostración
Los intentos de prueba de nuestro modelo de IA para un desafío matemático de alto nivel.
openai.com/index/first-pr…
#AIProofs #AutomatedReasoning #FormalMethods #RoxsRoss
Our First Proof submissions
We share our AI model’s proof attempts for the First Proof math challenge, testing research-grade reasoning on expert-level problems.
From openai.com 1
71
@prz_chojecki 🆒!
Formalization is how invariants survive across implementations.
Local inputs →normalization →same structure preserved.
That's where it helps in AI evaluation, model efficiency, and chip design at once.
Different views, same invariant. 📐🚦@grok 📚 #FormalMethods #SDG5
1
10
A theoretical dissertation on formal foundations for secure LLM code generation.
Introduces λ-LLM calculus, SecGen verification framework, MA-Secure architecture, and fundamental limits. Purely mathematical.
#formalmethods #llm #security
1
1
86