Appstract Inventions

• Welcome
• Contact

Appstract Inventions is a human knowledge application company, based in the US northwest.
We make innovative tools to explore and sculpt domain interaction.
We believe in using the power of human knowledge to make a better future.
Bring us your dream!

Leadership

Appstract R&D is led by Stu B. (LinkedIn, Reddit) known by his online handle: @stub22 => "Stu Btwotwo"
Stu's tech blog is Outside The Rectangle

Typesafe Composition => Authorable AI

1. Strong Types = Express invariant concepts as formal structures
2. Pure Functions = Scale reliably, using lazy streams of immutable data
3. Explicit Schemas = Define domain data clearly, aligned with public specifications
4. Symbolic Maths = Leverage equations (algebra, calculus, diff-eqs) across analysis ecosystems
5. Interactive Adjustments = Empower humans to filter, tweak, and experiment
6. Verified Proofs = Validate formal system constraints during builds, or even at runtime

OutRect blog highlights (2020-2023)

• Functional Domain Design
  • (2023-Feb) Noun-space, Type-paint, Morphism-arrows - Introduction to AxLam abstract kinds: a handwaving simplified description of categorical domains
  • (2023-Feb) Questions about Triangles - Concept modeling appetizer
  • (2020-Aug) Asset Allocation Challenge - Example domain model (Chapter 1 in series of 4 articles)
  • (2023-Jun) AxLam implementation types - Axiomagic reference impl of YaflSpec
• Streaming Analytics, Cloud Deployment
  • (2023-Apr) Streaming Analytics in Scala using ZIO 2.x - a minimal, self-contained runnable example of functional streaming
  • (2021-Aug) AxioMagic + Spark in cloud environments
  • (2023-Jun) ZIO interop with AWS cloud - Open source doc contribution to ZIO.dev guides - work in progress
• Knowledge Graphs, Models and Apps
  • (2020-Jul) Knowledge Graphs and Types - Grab bag of useful concepts, with links to background material and examples
  • (2020-Sep) Brief summary of Application Authoring in RDF as impl in glue-ai Glue-ML
• Symbolic Solvers, Formal Proofs, Knowledge Representation
  • (2020-Sep) Composed stochastic proof : Joint distribution monad as dependent type. (Chapter 3 of the Asset Allocation Challenge). Builds upon a reddit post with more detail and notation.
  • (2020-Aug) Solve, Search, or Prove? : Perspective on declarative math programs
  • (2021-Aug) Some links to recent academic work in the field of math knowledge representation

Open Source Software Prototypes

Active open source research code (Scala, RDF, JS) is available on github.
These two projects contain feature prototoypes overlapping with the concepts discussed in the OutRect blog posts above.

• GravAx - Example analytical apps in Scala using our #AxLam method. Some recent work includes
  • Pure math-enabled calculation using Typelevel and ZIO 2.x lazy functional streams
  • Deployment in AWS serverless applications using AWS SAM, Lambda and DynamoDb
  • Streamed approximations to joint probability distributions on high-D vector spaces
• AxioMagic - Dynamic tools for knowledge exploration, composed thru functions over curated model spaces of nouns.