Tight Sequential Confidence Intervals for Online Kernel Regression
Source: arxiv.org/abs/2611.19772
We characterize, in the worst case over kernels, the sharp minimax width of online confidence intervals for elements of a reproducing kernel Hilbert space (RKHS). This resolves the open problem posed by Vakili et al. (2021) in its original discrete-time form. A learner observes $y_i=f(x_i)+ε_i$ at adaptively chosen points, with $\|f\|_{\mathcal H_k}\le B$ and i.i.d. $R$-sub-Gaussian noise, and reports the interval $|f(x)-μ_n(x)|\leρ_n(δ)σ_n(x)$, where $μ_n$ and $σ_n$ are the posterior mean and standard deviation of a surrogate Gaussian process model with regularization parameter $λ$. Writing…
Comments and reviews
No comments yet.
API
curl /api/v1/records/arxiv/2611.19772
curl -X POST /api/v1/records/arxiv/2611.19772/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2611.19772/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"body": "…", "review": true}'