Privacy aware data science in Scala with monads...
summary
Proposes a Scala library that enforces information-flow policies at compile time by translating a Haskell seclib design into a monadic security wrapper with flatMap-based declassification and a type-level join (last-write-wins) computed via abstract type members. Implements a neutral security level so nested wrappers collapse predictably, wires into Cats and monad transformers (OptionT) so queries, analysis and reporting compose, and uses ScalaTest’s compile-time checks to reject unsafe reveals; works for non-adversarial, audited code but remains a work-in-progress.
AI-written summary · about
chapters (15)
- 0:00Introduction and context
- 2:12Proposed technical approach
- 3:28Challenges of data workflows
- 5:57Risks of sensitive data
- 8:31Data leakage and modeling
- 11:38Scope and limitations
- 14:46Designing the data box
- 16:56Lattice-based security models
- 21:10Graded monads implementation
- 22:29Scala implementation details
- 24:32Type-level computation
- 26:41Building the security API
- 30:42Benefits for data scientists
- 33:44Pipeline construction
- 35:33Future work and conclusions
related talks
01h 02mPrinciples and tools to cleverly model your data in Scala
38mThinking Less with Scala
35mIntroducing Typelevel Scala into an OO environment
29mLearning Scalaz
01h 01mTypesafe Scala, Typelevel Scala – What’s Going On, Anyway?
32mSafer Code with LLMs: Scala Types Silver Bullet?
58mSeparating Behavior and Data with Type Classes in Scala
24mLet the Scala compiler work for you