2026. Neural Architecture Optimization for Low-Latency Large Language Model Inference. Journal of Advanced Artificial Intelligence Research. 5, 1 (Aug. 2026).