Semantics-Native Communication via Contextual Reasoning (SNC)

• Implementation and experiment code for Semantics-Native Communication via Contextual Reasoning
• 

Architecture

Structure

• system1.py — system 1 SNC: action–concept relevance, A2C/C2A, concept–symbol mapping, Theorem 1 (SR bit-length bounds)
• system2.py — system 2 SNC: individual context (S, L), objective G, self-SNC, Algorithm 1
• experiments/
  • world.py — world generation for section IV experiments (|A|=|C|=100, Dirichlet)
  • reliability.py — system 1 vs system 2 reliability γ
  • fig4.py–fig8.py — scripts to reproduce paper Figures 4–8
• resulsts/ - architecture, output figures from the experiment

Results

Figure	Description
Fig 4	G convergence and stationary rA2C heatmap
Fig 5	Reliability γ vs α, β (3D surface; r = 10, 20, 100)
Fig 6	Reliability γ vs communication rounds
Fig 7	SR length achieving γ = 1 (System 1 vs System 2)
Fig 8	Robustness to asynchronous context (γ vs perturbation ε)

Preview: