Mohamad Al-Zawahreh
Co-Author / Lead, LLM Context Collapse
Independent researcher and lead author of Solving LLM Context Collapse: A Thermodynamic No-Go Theorem for AI Cognitive Architecture and Biological Memory, a cross-domain paper unifying transformer context degradation with biological memory decay under a single thermodynamic framework. The Semantic Landauer Limit derived in this work establishes an irreducible energy cost for semantic compression and reconstruction, with empirical validation against both GPT-4 RULER benchmark data and clinical curves from anti-amyloid Alzheimer's monotherapy.
Published Work
Papers in the Lumiea Series
Solving LLM Context Collapse: A Thermodynamic No-Go Theorem for AI Cognitive Architecture and Biological Memory
Lead author
Mohamad Al-Zawahreh, Gary Floyd, Mario Martín Cuniglio
View in catalog →
About
Bio
[Bio content pending — please contact for collaboration inquiries]
Bio details pending. Reach out via the Lumiea Systems R&D contact channel for collaboration inquiries.
Find Mohamad Online
Verified academia.edu profile (linked from the LLM Context Collapse paper):