When Volatility Travels: Mapping Global Spillovers with Rough Multivariate Models
A mechanism-first look at how multivariate rough volatility models can turn cross-market covariance patterns into a more useful map of volatility spillovers.
Cognaptus Insights
Industry thoughts, trends, and commentary.
From Black Box to Glass Box: DeepVIS Makes Data Visualization Explain Itself
DeepVIS shows how AI chart generation becomes more useful when users can inspect and edit the reasoning between a business question and the final visualization.
From Chaos to Choreography: The Future of Agent Workflows
A practical reading of agent workflow systems as orchestration infrastructure, not just smarter prompt chains.
From Stage to Script: How AMADEUS Keeps AI Characters in Character
A mechanism-first reading of AMADEUS, a training-free RAG framework for keeping role-playing agents consistent when users ask questions beyond the script.
Meta-Game Theory: What a Pokémon League Taught Us About LLM Strategy
A small Pokémon tournament shows why LLM evaluation should measure strategy, rationale, and constraint exploitation—not just polished reasoning.
Quantum Bridges: Crossing the Label Gap with ILQSSL and IPQSSL
A grounded reading of ILQSSL and IPQSSL, two quantum graph semi-supervised learning methods that look promising on small benchmarks but still demand discipline around noise, separability, and deployment claims.
FAITH in Numbers: Stress-Testing LLMs Against Financial Hallucinations
How FAITH turns financial tables into an auditable stress test for LLM hallucinations, and what its results imply for finance teams using generative AI.
From Zero to Reasoning Hero: How R-Zero Teaches Itself Without Human Data
R-Zero shows how reasoning models can self-generate useful training curricula, but its real lesson is disciplined reward design rather than magic data-free learning.
Mind the Gap: How Tool Graph Retriever Fixes LLMs’ Missing Links
A dependency-aware retriever shows why agentic AI fails when it retrieves the obvious tool but misses the prerequisite one.
The Diligent but Brittle Student Inside Every LLM
A year-long simulated classroom shows why LLMs can look studious, confident, and improving while still failing the tests that require transferable understanding.