boosting
PulseAugur coverage of boosting — every cluster mentioning boosting across labs, papers, and developer communities, ranked by signal.
-
LLM cascades fail due to similar model errors, violating ensemble theory
LLM cascades, which use a cheaper model for initial responses and escalate to a more powerful one if needed, often fail to deliver cost savings because the models are trained on similar data and thus make similar errors…
-
Mentored Decoding: ML Boosting Theory Enhances LLM Inference Speed and Quality
Researchers have introduced "Mentored Decoding," a novel approach that enhances language model inference speed and quality by drawing parallels to machine learning boosting theory. This method formally defines lossy spe…
-
New Fibonacci Ensembles method enhances machine learning aggregation
Researchers have introduced a novel ensemble learning method called Fibonacci Ensembles, inspired by the mathematical properties of the Fibonacci sequence and the golden ratio. This approach offers an alternative to tra…
-
Classical ML models show power-law scaling on tabular data
A new study published on arXiv benchmarks classical machine learning models on tabular data, revealing that power laws accurately describe learning curves across various datasets and model families. The research found t…
-
New framework compresses ensemble models with sparse bagging
A new research paper introduces Simplex-Constrained Sparse Bagging (SCSB), a novel framework designed to compress and calibrate bootstrap-based bagging ensembles. SCSB addresses the limitations of standard bagging metho…