Research map / Learning theory and algorithms / Prediction and decision losses

Learning No-Regret Sparse Generalized Linear Models with Varying Observation(s)

ICLR 2024 · Learning theory and algorithms › Prediction and decision losses · approach: Data attribution and recourse

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes an efficient incremental learning method for sparse generalized linear models that updates the optimal model as observations are added or removed, with no-regret guarantees.

Problem

Training optimal sparse GLMs when observations are frequently added to or removed from the training set, without retraining from scratch.

Approach

Incremental learning for sparse GLMs under varying observations, designed so that the update is efficient and incurs no regret relative to retraining on the modified dataset. Full method details are not available beyond the abstract.

Closest papers by meaning

• Scalable First-order Method for Certifying Optimal k-Sparse GLMs (ICML 2025 · similarity 0.88)
• Generalized Linear Model Regression under Distance-to-set Penalties (NeurIPS 2017 · similarity 0.88)
• Scaled Least Squares Estimator for GLMs in Large-Scale Problems (NeurIPS 2016 · similarity 0.87)
• A Unified Confidence Sequence for Generalized Linear Models, with Applications to Bandits (NeurIPS 2024 · similarity 0.86)
• SLoPe: Double-Pruned Sparse Plus Lazy Low-Rank Adapter Pretraining of LLMs (ICLR 2025 · similarity 0.84)
• New Bounds for Sparse Variational Gaussian Processes (ICML 2025 · similarity 0.84)