Skip to search formSkip to main contentSkip to account menu

Backplay: "Man muss immer umkehren"

The approach, Backplay, uses a single demonstration to construct a curriculum for a given task, and analytically characterize the types of environments where Backplay can improve training speed and compare favorably to other competitive methods known to improve sample efficiency.
View PDF on arXiv (opens in a new tab)

Variational Auto-Regressive Gaussian Processes for Continual Learning

Experiments on standard continual learning benchmarks demonstrate the ability of VAR-GPs to perform well at new tasks without compromising performance on old ones, yielding competitive results to state-of-the-art methods.
View PDF on arXiv (opens in a new tab)

Multi-Agent Reinforcement Learning: A Report on Challenges and Approaches

The report concludes with advances in the paradigm of training Multi-Agent Systems called Decentralized Actor, Centralized Critic, based on an extension of MDPs called \textit{Decentralization Partially Observable MDP}s, which has seen a renewed interest lately.
View PDF on arXiv (opens in a new tab)

First-Order Preconditioning via Hypergradient Descent

First-order preconditioning (FOP) is introduced, a fast, scalable approach that generalizes previous work on hypergradient descent and is able to improve the performance of standard deep learning optimizers on visual classification and reinforcement learning tasks with minimal computational overhead.
View PDF on arXiv (opens in a new tab)

On Uncertainty, Tempering, and Data Augmentation in Bayesian Classification

It is shown that a cold posterior, tempered by a power greater than one, often more honestly reflects the authors' beliefs about aleatoric uncertainty than no tempering, providing an explicit link between data augmentation and cold posteriors.
View PDF on arXiv (opens in a new tab)

When are Iterative Gaussian Processes Reliably Accurate?

This work investigates CG tolerance, preconditioner rank, and Lanczos decompositions, and shows that L-BFGS-B is a compelling optimizer for Iterative GPs, achieving convergence with fewer gradient updates.
View PDF on arXiv (opens in a new tab)

SKIing on Simplices: Kernel Interpolation on the Permutohedral Lattice for Scalable Gaussian Processes

This work develops a connection between SKI and the permutohedral lattice used for highdimensional fast bilateral filtering, and provides a CUDA implementation of Simplex-GP, which enables significant GPU acceleration of MVM based inference.
View PDF on arXiv (opens in a new tab)

Pre-Train Your Loss: Easy Bayesian Transfer Learning with Informative Priors

This work shows that highly informative posteriors can be learned from the source task, through supervised or self-supervised approaches, which then serve as the basis for priors that modify the whole loss surface on the downstream task.
View PDF on arXiv (opens in a new tab)

PAC-Bayes Compression Bounds So Tight That They Can Explain Generalization

A compression approach based on quantizing neural network parameters in a linear subspace is developed, profoundly improving on previous results to provide state-of-the-art generalization bounds on a variety of tasks, including transfer learning.
View PDF on arXiv (opens in a new tab)

A Simple and Fast Baseline for Tuning Large XGBoost Models

It is shown that uniform subsampling makes for a simple yet fast baseline to speed up the tuning of large XGBoost models using multi-fidelity hyperparameter optimization with data subsets as the fidelity dimension.
View PDF on arXiv (opens in a new tab)
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)