ukubona × who india · θᵗ → L₀+Σwᵢ·Lᵢ → f(σ²,λ,ε) → γ|ε_FGT|² → L(θᵗ⁺¹)

AI Tools for Literature Review

Five levels. Five sessions each. A ROYGBIV loss gradient — from what we already believe to what we now know — built for WHO India health economists.

θᵗ  →  L₀+Σwᵢ·Lᵢ  →  f(σ²,λ,ε)  →  γ|ε_FGT|²  →  L(θᵗ⁺¹)
Level I · Red · θᵗ A Priori What we already believe
I · 1The Existing Process I · 2The Tools Available I · 3The Framework I · 4The Protocol I · 5The Evidence Brief
Level II · Orange · L₀+Σwᵢ·Lᵢ Bias & Weights How we attend
II · 1Bias in the Machine II · 2Prompt Engineering II · 3Critical Appraisal II · 4Meta-Analysis Basics II · 5Surveillance Use Cases
Level III · Yellow · f(σ²,λ,ε) Composite Loss The civilizational stack
III · 1Ethics of AI Evidence III · 2Institutional Integration III · 3Rapid Evidence Reviews III · 4Global Evidence Ecosystems III · 5Future of Evidence
Level IV · Green · γ|ε_FGT|² Error What the model cannot account for
IV · 1Residual Bias IV · 2Irreducible Stochasticity IV · 3The Anchor Constraint IV · 4Transferability Limits IV · 5Operating Under Uncertainty
Level V · Blue · L(θᵗ⁺¹) Posteriori The updated prior
V · 1Brief as Parameter Update V · 2Durable Institutional Memory V · 3Continuous Surveillance V · 4The Federated Evidence Twin V · 5The Integrator's Horizon
loss gradient · TMVES · tensor → matrix → vector → eigenmode → scalar
Red · A Priori · θᵗ Orange · Bias & Weights · L₀+Σwᵢ·Lᵢ Yellow · Vector · f(σ²,λ,ε) Green · Eigenmode · γ|ε_FGT|² Blue · Posteriori · L(θᵗ⁺¹)
five AI layers · evidence synthesis reference
I · Tensor · θᵗ World AI General science · ambient intelligence The silent substrate — general scientific understanding across all domains already in the room II · Matrix · L₀+Σwᵢ·Lᵢ Perception AI Ingestion · journals · grey sources Ingests specific literature across journals, languages, and grey sources via PDF upload and API III · Vector · f(σ²,λ,ε) Agentic AI Screening · extraction · repetitive tasks Automates data extraction on identified literature — the highest-ROI layer for adoption right now IV · Eigenmode · γ|ε_FGT|² Generative AI Synthesis · reports · policy briefs Synthesizes findings into decision-ready documents — think Claude, not just ChatGPT V · Scalar · L(θᵗ⁺¹) Embodied AI Human in loop · audit · conscience Bias detection, equity checks, PRISMA compliance — the layer that makes the others trustworthy

No tools match — try a different search or filter.