3 papers
cs.LG2026
Chronicals: A High-Performance Framework for LLM Fine-Tuning with 3.51x Speedup over Unsloth
Arjun S. Nair
Large language model fine-tuning is bottlenecked by memory: a 7B parameter model requires 84GB--14GB for weights, 14GB for gradients, and 56GB for FP32 optimizer states--exceeding…
cs.LG2025
Calibration Meets Reality: Making Machine Learning Predictions Trustworthy
Kristina P. Sinaga, Arjun S. Nair
Post-hoc calibration methods are widely used to improve the reliability of probabilistic predictions from machine learning models. Despite their prevalence, a comprehensive theoret…
cs.LG2025
Disproving the Feasibility of Learned Confidence Calibration Under Binary Supervision: An Information-Theoretic Impossibility
Arjun S. Nair, Kristina P. Sinaga
We prove a fundamental impossibility theorem: neural networks cannot simultaneously learn well-calibrated confidence estimates with meaningful diversity when trained using binary c…