← Back to all news
arXiv AI

Compressed Active Subspaces for Scalable Bayesian Inference

The paper introduces Compressed Active Subspaces (CAS), a scalable method for Bayesian inference in high‑dimensional models. CAS first compresses model parameters via a structured isometric embedding, then constructs the active subspace in this reduced space, dramatically lowering memory requirements. Experiments on neural networks of growing size show that CAS preserves predictive performance and provides robust uncertainty estimates while enabling inference where traditional active subspace methods fail.

Related stories

arXiv Machine Learning
Sep 4

Active learning for data-driven reduced models of parametric differential systems with Bayesian operator inference

The paper presents an active learning framework that enhances data-driven reduced-order models (ROMs) for parametric dynamical systems by intelligently selecting training parameters. Using a Bayesian linear regression version of operator inference, the method quantifies prediction uncertainty to guide sequential adaptive sampling, aiming to improve ROM stability and accuracy across the parameter domain. Numerical experiments on nonlinear PDE systems show that this adaptive strategy outperforms random sampling under the same computational budget.

By Shane A. McQuarrie, Mengwu Guo, Anirban Chaudhuri
More like this →
arXiv AI
Sep 4

Subspace Inference Enables Efficient Active Reward Learning from Preferences

The paper introduces PreferenceEKF, a sample‑efficient method for active reward learning from human preferences. By framing preference learning as a sequential Bayesian filtering problem, it tracks reward model uncertainty using an extended Kalman filter in a low‑dimensional subspace, avoiding costly posterior inference over the full neural network. Experiments on D4RL and V‑D4RL benchmarks show improved sample efficiency, runtime, scalability, and calibration, with reward models that support competitive offline reinforcement learning policies.

By Yutai Zhou, Erdem B{\i}y{\i}k
reinforcement-learningbenchmarks
More like this →
arXiv Machine Learning
4d ago

High-Dimensional Simulation-Based Inference in Latent Spaces

arXiv:2609.37381v1 Announce Type: new Abstract: Neural simulation-based inference (SBI) has been widely successful in inferring a relatively small number of interpretable parameters from potentially...

By Lars K\"uhmichel, Stefan T. Radev, Bhanu Prasanna Koppolu, Masoumeh Davoudi, Jerry M. Huang, Paul-Christian B\"urkner
diffusion
More like this →
arXiv Machine Learning
Sep 4

Linearized subspace refinement framework to expose hidden accuracy in trained neural networks

The paper introduces Linearized Subspace Refinement (LSR), a post‑training framework that uses the local linearized model of a trained neural network to compute a low‑dimensional correction via a reduced least‑squares problem. LSR is architecture‑agnostic and improves accuracy across tasks such as function approximation, operator learning, physics‑informed fine‑tuning, and noisy inverse problems, often achieving order‑of‑magnitude error reductions. The method reveals that standard training can leave significant accuracy plateaus due to numerical ill‑conditioning, and it offers a subspace rank that balances correction strength, stability, and noise sensitivity.

By Wenbo Cao, Weiwei Zhang
fine-tuning
More like this →