4 papers
The Lossy Horizon: Error-Bounded Predictive Coding for Lossy Text Compression (Episode I)
Nnamdi Aghanya, Jun Li, Kewei Wang
Large Language Models (LLMs) can achieve near-optimal lossless compression by acting as powerful probability models. We investigate their use in the lossy domain, where reconstruct…
Heaven & Hell II: Scale Laws and Robustness in One-Step Heaven-Hell Consensus
Nnamdi Daniel Aghanya, Romain Leemans
We study Heaven-Hell dynamics, a model for network consensus. A known result establishes an exact one-step convergence threshold for systems with a single uniform hub: the per-node…
Heaven & Hell: One-Step Hub Consensus
Nnamdi Daniel Aghanya
Many networked systems require a central authority to enforce a global configuration against local peer influence. We study influence dynamics on finite weighted directed graphs wi…
Queue up for takeoff: a transferable deep learning framework for flight delay prediction
Nnamdi Daniel Aghanya, Ta Duong Vu, Amaëlle Diop +5
Flight delays are a significant challenge in the aviation industry, causing major financial and operational disruptions. To improve passenger experience and reduce revenue loss, fl…