SYSTEM LIVE / PSYCHIATRYAI.COM / AI RAG ARCHITECTURE v1.0
// RAISR4D Hybrid Knowledge Architecture

PSYCHIATRYAI.COM AI Algorithm

A real-time vector retrieval system powering high-dimensional clinical semantic discovery across mental health and psychiatry knowledge nodes derived from the latest research from PubMed.

AI Architecture Pipeline

Article Published │ ▼ AI Embedding Model │ ▼ Article converted into vector [0.23, -0.41, 0.76 ...] │ ▼ Semantic Vector Space │ ▼ Compare against 23,000+ articles │ ▼ Cosine Similarity Engine │ similarity ≥ 0.70 │ ▼ AI-ranked Related Posts

The Core Algorithm

Each article becomes a mathematical vector:

$$\text{Article}_A \rightarrow \vec{A}$$ $$\text{Article}_B \rightarrow \vec{B}$$

The AI calculates the angle between them:

$$\text{Similarity} = \cos(\theta) = \frac{\vec{A} \cdot \vec{B}}{||\vec{A}||\,||\vec{B}||}$$

The closer the vectors point in the same direction, the more semantically related they are.

$$\mathbf{a} \cdot \mathbf{b} = a_x b_x + a_y b_y = \lVert\mathbf{a}\rVert \lVert\mathbf{b}\rVert \cos\theta$$
$$\mathbf{a} \cdot \mathbf{b} = 4.87(-4.25) + -3.5(-4) = -6.71$$

The angle is obtuse, so the vectors point partly apart and the dot product is negative.

Interactive Vector Space

Drag the endpoints of vector a (purple) or b (blue) below to re-calculate cosine angle and dot product in real-time:

Inputs: x, y, θ
Vector a: (4.87, -3.50)
Vector b: (-4.25, -4.00)
Angle θ: 101.05°
Dot Product (a · b): -6.71
Cosine Similarity: -0.192

The 0.7 Cosine Threshold

The AI algorithm applies:

$$\cos(\theta) \geq 0.7$$

Meaning:

The angular boundary is:

$$\theta = \cos^{-1}(0.7) \approx 45.6^\circ$$

Psychiatryai.com is effectively saying:

"Only show related articles whose semantic direction is within approximately 45 degrees of the original article."

Example in Psychiatry

Input article: "Treatment-resistant schizophrenia: clozapine monitoring and outcomes"

Traditional Keyword Engine

  • • schizophrenia
  • • clozapine
  • • treatment

AI Semantic Engine Discovers

  • First episode psychosis
  • Antipsychotic optimisation
  • Negative symptoms
  • Early intervention services
  • Clozapine guidelines

...because these concepts occupy nearby regions in semantic space.

Hybrid Intelligence Layer

PsychiatryAI effectively combines:

Related Posts Engine │ ┌────────────┴────────────┐ ▼ ▼ Contextual AI Semantic AI Tags Embeddings Categories Vector geometry Keywords Cosine similarity └────────────┬────────────┘ ▼ Higher-quality recommendations

Psychiatryai.com leverages novel RAISR4D (Real-Time AI Scoping Review) technology to deploy real-time AI clinical modeling and semantic discovery across complex mental health and psychiatric domains.

"Psychiatryai.com leverages cosine vector similarity (0.7) to turn static data into a real-time, high-dimensional clinical knowledge AI engine for modern mental health and psychiatric CME/CPD and research.