AI Paper Atlas

Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes

Characterizing Deep Gaussian Processes via Nonlinear Recurrence Systems

Anh Tong, Jaesik Choi

AAAI 2021 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Infinite-width networks as Gaussian processes · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Constrain squared-exponential kernel hyperparameters to avoid pathological regionsZero-mean DGP · MNISTclassification accuracy: ≈ 10% -> 91.21%Improved
Constrain squared-exponential kernel hyperparameters to avoid pathological regionsZero-mean DGP · Boston housing and diabetesRMSEImproved
Use a spectral mixture kernel instead of a kernel with a pathological parameter regionDeep Gaussian Process · Deep-layer output distancesRMSDImproved
Increase the layer width so that m > ℓ²/σ²Zero-mean DGP with squared-exponential kernel · Deep-layer output distancesRMSDImproved
Set cosine-kernel hyperparameters so that π²σ² / p² > 1DGP with cosine kernel · Deep-layer output distancesRMSDImproved
Use nonzero-mean rather than zero-mean layersDeep Gaussian Process · Deep-layer output distancesexpected squared output distanceImproved
Connect each layer to the inputDGP with squared-exponential kernel · Deep-layer output distancesexpected squared output distanceImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy