Table 1 Pareto MTL versus scalarization

\({\overline{U}}_n(\varvec{\phi }, \varvec{\theta })\) versus \({\overline{V}}_n(\varvec{\phi })\) when \(V_n(\varvec{\phi }) = MINE(\varvec{\phi })\)

TM-MARROW

\(\varvec{0.88} \pm 0.08\)

\(0.57 \pm 0.17\)

\(\varvec{0.58} \pm 0.16\)

\(0.42 \pm 0.15\)

\(0.19 \pm 0.01\)

\(0.19 \pm 0.01\)

\(0.13 \pm 0.03\)

\(0.13 \pm 0.04\)

\(\varvec{8.1} \pm 1.20\)

\(3.0 \pm 0.47\)

\(\varvec{5.8} \pm 1.55\)

\(2.5 \pm 0.71\)

MACAQUE-RETINA

\(\varvec{0.83} \pm 0.10\)

\(0.46 \pm 0.09\)

\(\varvec{0.47} \pm 0.13\)

\(0.35 \pm 0.13\)

\(\varvec{0.51} \pm 0.01\)

\(0.50 \pm 0.01\)

\(0.35 \pm 0.02\)

\(0.35 \pm 0.02\)

\(\varvec{7.5} \pm 1.27\)

\(3.7 \pm 0.95\)

\(\varvec{4.2} \pm 1.03\)

\(2.3 \pm 1.06\)

\({\overline{U}}_n(\varvec{\phi },\varvec{\theta })\) versus \({\overline{V}}_n(\varvec{\phi })\) when \(V_n(\varvec{\phi })=MMD(\varvec{\phi })\)

TM-MARROW

\(\varvec{0.84} \pm 0.06\)

\(0.58 \pm 0.06\)

\(\varvec{0.55} \pm 0.08\)

\(0.49 \pm 0.07\)

\(26.68 \pm 0.41\)

\(\varvec{26.91} \pm 0.37\)

\(25.95 \pm 1.32\)

\(\varvec{26.78} \pm 1.03\)

\(\varvec{8.4} \pm 0.70\)

\(6.6 \pm 0.70\)

\(\varvec{5.8} \pm 1.03\)

\(5.3 \pm 0.82\)

MACAQUE-RETINA

\(\varvec{0.78} \pm 0.10\)

\(0.75 \pm 0.10\)

\(\varvec{0.61} \pm 0.09\)

\(0.55 \pm 0.06\)

\(752.35 \pm 17.90\)

\(\varvec{755.46} \pm 6.19\)

\(623.00 \pm 83.73\)

\(\varvec{700.07} \pm 32.06\)

\(\varvec{8.0} \pm 0.47\)

\(6.9 \pm 1.10\)

\(\varvec{6.0} \pm 0.94\)

\(4.7 \pm 0.82\)