3 problems

0 votes0 replies0 views
The machine-learning-aided robust detection conjecture
A robust hypothesis test is a detector designed to maintain performance under uncertainty about the underlying distributions or model, and machine learning methods can be used to l…
Conjectureopenmachine-learning minimax-detection robust-statisticsarXiv posted 2 days ago
0 votes0 replies0 views
The correct-divergence conjecture for minimax robust detection
In minimax detection, hypotheses and action/uncertainty sets determine the class of admissible detectors, while an $f$ -divergence is a statistical similarity measure used as a surr…
Conjectureopenadversarial-machine-learning f-divergences minimax-detectionarXiv posted 2 days ago
0 votes0 replies0 views
Conjecture on the impossibility of asymptotic equivalence for the spectral cut-off separation radius
Let $\bar{r}_\varepsilon^\star>0$ satisfy $u(\bar{r}_\varepsilon^\star)=t_{1-\alpha}-t_\beta$ for prescribed $\alpha,\beta\in]0,1[$ . Let $\tilde{r}_{\varepsilon,\tilde D}^\star$ be…
Conjectureopenasymptotic-statistics minimax-detectionarXiv posted 2 days ago

The machine-learning-aided robust detection conjecture