🏢 Amazon

Agnostic boosting gets a major efficiency upgrade! A new algorithm leverages sample reuse to drastically reduce the data needed for accurate learning, closing the gap with computationally expensive al…

Risk-Averse RLHF fine-tunes LLMs to minimize toxic outputs while maintaining performance.

Deleting data from machine learning models exposes individuals to highly accurate reconstruction attacks, even when models are simple; this research demonstrates the vulnerability.

Researchers achieved high-accuracy differentially private (DP) models by using a novel DP continual pre-training strategy with only 10% public data, mitigating the performance degradation common in DP…

Causal assumptions drastically alter predictor merging, with CMAXENT revealing logistic regression for causal and LDA for anticausal directions.