Modified Distortion Matrices for Phrase-Based SMT Arianna Bisazza & Marcello Federico – FBK (Italy) PSMT decoding overview E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali 2 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT PSMT decoding overview E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali Freedom of movement must be encouraged LM scores 3 LM scores A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT PSMT decoding overview E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali Freedom of movement must be encouraged while ensuring that career paths … LM scores 4 LM scores LM scores LM scores A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT PSMT decoding overview E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali Freedom of movement must be encouraged while ensuring that career paths … LM scores 5 LM scores LM scores LM scores A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Reordering Models Many solutions have been proposed with different reo. classes, features, train modes etc. Tillman 04, Zens & Ney 06 AlOnaizan & Papineni 06 Galley & Manning 08 Green & al.10, Feng & al.10 … E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali 6 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Reordering Models Many solutions have been proposed with different reo. classes, features, train modes etc. Tillman04, Zens&Ney06 Tillman 04, Zens & Ney 06 AlOnaizan & Papineni06 AlOnaizan & Papineni 06 Galley & Manning08 Galley & Manning 08 Green &al.10,Feng Feng& &al.10 Green & al.10, al.10 …… E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali No matter what reordering model is used, permutation search space must be limited! The power of all reordering models is bound to the reordering constraints in use 7 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali 8 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali Reordering Constraints #perm.=11!≈40,000,000 9 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali Reordering Constraints #perm.=11!≈40,000,000 D(x,y)=|y-x-1| w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 3 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 w9 10 9 w10 11 10 w2 2 1 0 2 3 4 5 6 7 8 9 w3 3 2 1 0 2 3 4 5 6 7 8 w4 4 3 2 1 0 2 3 4 5 6 7 w5 5 4 3 2 1 0 2 3 4 5 6 w6 6 5 4 3 2 1 0 2 3 4 5 w7 7 6 5 4 3 2 1 0 2 3 4 w8 8 7 6 5 4 3 2 1 0 2 3 w9 9 8 7 6 5 4 3 2 1 0 2 Source-to-Source distortion 10 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT w10 10 9 8 7 6 5 4 3 2 1 0 E' necessario incoraggiare tale mobilità garantendo la sicurezza dei percorsi professionali DL: distortion limit Reordering Constraints #perm.=11!≈40,000,000 D(x,y)=|y-x-1| DL=3 #perm.≈7,000 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 3 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 w9 10 9 w10 11 10 w2 2 1 0 2 3 4 5 6 7 8 9 w3 3 2 1 0 2 3 4 5 6 7 8 w4 4 3 2 1 0 2 3 4 5 6 7 w5 5 4 3 2 1 0 2 3 4 5 6 w6 6 5 4 3 2 1 0 2 3 4 5 w7 7 6 5 4 3 2 1 0 2 3 4 w8 8 7 6 5 4 3 2 1 0 2 3 w9 9 8 7 6 5 4 3 2 1 0 2 Source-to-Source distortion 11 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT w10 10 9 8 7 6 5 4 3 2 1 0 w0 w1 w2 w 3 w4 w5 w6 w7 w8 w9 w <s> 0 1 2 3 4 5 6 7 8 9 10 w0 0 1 2 3 4 5 6 7 8 9 w1 2 0 1 2 3 4 5 6 7 8 w2 3 2 0 1 2 3 4 5 6 7 w3 4 3 2 0 1 2 3 4 5 6 w4 5 4 3 2 0 1 2 3 4 5 w5 6 5 4 3 2 0 1 2 3 4 w6 7 6 5 4 3 2 0 1 2 3 w7 8 7 6 5 4 3 2 0 1 2 w8 9 8 7 6 5 4 3 2 0 1 w9 10 9 8 7 6 5 4 3 2 0 w10 11 10 9 8 7 6 5 4 3 2 10 The problem with DL… Arabic-English EN AR EN AR 12 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT w0 w1 w2 w 3 w4 w5 w6 w7 w8 w9 w <s> 0 1 2 3 4 5 6 7 8 9 10 w0 0 1 2 3 4 5 6 7 8 9 w1 2 0 1 2 3 4 5 6 7 8 w2 3 2 0 1 2 3 4 5 6 7 w3 4 3 2 0 1 2 3 4 5 6 w4 5 4 3 2 0 1 2 3 4 5 w5 6 5 4 3 2 0 1 2 3 4 w6 7 6 5 4 3 2 0 1 2 3 w7 8 7 6 5 4 3 2 0 1 2 w8 9 8 7 6 5 4 3 2 0 1 w9 10 9 8 7 6 5 4 3 2 0 w10 11 10 9 8 7 6 5 4 3 2 10 The problem with DL… German-English EN EN DE DE 13 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT #perm.=11! ≈40,000,000 D(x,y)=|y-x-1| DL=3 #perm.≈7,000 Current solution: increase the DLimit w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 3 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 w9 10 9 w10 11 10 w2 2 1 0 2 3 4 5 6 7 8 9 w3 3 2 1 0 2 3 4 5 6 7 8 w4 4 3 2 1 0 2 3 4 5 6 7 w5 5 4 3 2 1 0 2 3 4 5 6 w6 6 5 4 3 2 1 0 2 3 4 5 w7 7 6 5 4 3 2 1 0 2 3 4 w8 8 7 6 5 4 3 2 1 0 2 3 w9 9 8 7 6 5 4 3 2 1 0 2 Source-to-Source distortion 14 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT w10 10 9 8 7 6 5 4 3 2 1 0 #perm.=11! ≈40,000,000 D(x,y)=|y-x-1| DL=3 #perm.≈7,000 DL=7 #perm.≈7,000,000 Generally leads to worse translations! Current solution: increase the DLimit w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 3 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 w9 10 9 w10 11 10 w2 2 1 0 2 3 4 5 6 7 8 9 w3 3 2 1 0 2 3 4 5 6 7 8 w4 4 3 2 1 0 2 3 4 5 6 7 w5 5 4 3 2 1 0 2 3 4 5 6 w6 6 5 4 3 2 1 0 2 3 4 5 w7 7 6 5 4 3 2 1 0 2 3 4 w8 8 7 6 5 4 3 2 1 0 2 3 w9 9 8 7 6 5 4 3 2 1 0 2 Source-to-Source distortion 15 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT w10 10 9 8 7 6 5 4 3 2 1 0 Our solution: #perm.=11! ≈40,000,000 D(x,y)=|y-x-1| DL=3 #perm.≈7,000 DL=7 #perm.≈7,000,000 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 3 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 w9 10 9 w10 11 10 w2 2 1 0 2 3 4 5 6 7 8 9 w3 3 2 1 0 2 3 4 5 6 7 8 w4 4 3 2 1 0 2 3 4 5 6 7 w5 5 4 3 2 1 0 2 3 4 5 6 w6 6 5 4 3 2 1 0 2 3 4 5 w7 7 6 5 4 3 2 1 0 2 3 4 w8 8 7 6 5 4 3 2 1 0 2 3 w9 9 8 7 6 5 4 3 2 1 0 2 Source-to-Source distortion 16 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT w10 10 9 8 7 6 5 4 3 2 1 0 #perm.=11! ≈40,000,000 D(x,y)=|y-x-1| DL=3 #perm.≈7,000 DL=7 #perm.≈7,000,000 DL=3 & modif(D) #perm.≈20,000 Simplifies the task of reordering models! Our solution: modify distortion for each test sentence w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 3 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 w9 10 9 w10 11 10 w2 2 1 0 2 3 4 5 6 7 8 9 w3 3 2 1 0 2 3 4 5 6 2 8 w4 4 3 2 1 0 2 3 4 5 2 7 w5 5 4 3 2 1 0 2 3 4 5 6 w6 6 5 4 3 2 1 0 2 3 4 5 w7 7 6 0 0 3 2 1 0 2 3 4 w8 8 7 0 0 4 3 2 1 0 2 3 w9 9 8 7 6 5 4 3 2 1 0 2 Source-to-Source distortion 17 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT w10 10 9 8 7 6 5 0 3 2 1 0 Rest of the talk: How to modify the distortion matrix? What effect on translation quality? What effect on baseline runtimes? 18 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Chunk-based fuzzy reordering rules Shallow syntax chunking: • cheaper and easier than deep parsing • constrains reorderings in a softer way Fuzzy (non-determinisic) reordering rules: • generate N permutations for each matching sequence • final reordering decision is taken during translation, guided by all SMT models (reoM, LM...) Few rules for language pair, to only capture long reordering 19 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Chunk-based fuzzy reordering rules w- $Ark and took part CC1 20 VC2 Arabic-English “Move verb chunk (and following chunk) to the right by 1 to N chunks” fy AltZAhrp E$rAt AlmslHyn in the march PC3 mn AlktA}b . dozens of militants from the Brigades NC4 PC5 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Pct6 Chunk-based fuzzy reordering rules w- $Ark and took part CC1 CC1 CC1 CC1 21 VC2 PC3 PC3 PC3 Arabic-English “Move verb chunk (and following chunk) to the right by 1 to N chunks” fy AltZAhrp E$rAt AlmslHyn in the march PC3 VC2 NC4 NC4 mn AlktA}b . dozens of militants from the Brigades NC4 PC5 Pct6 NC4 VC2 PC5 PC5 PC5 VC2 Pct6 Pct6 Pct6 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Chunk-based fuzzy reordering rules w- $Ark and took part CC1 CC1 CC1 CC1 CC1 CC1 22 VC2 PC3 PC3 PC3 NC4 NC4 Arabic-English “Move verb chunk (and following chunk) to the right by 1 to N chunks” fy AltZAhrp E$rAt AlmslHyn in the march mn AlktA}b . dozens of militants from the Brigades PC3 VC2 NC4 NC4 VC2 PC5 NC4 PC5 Pct6 NC4 VC2 PC5 PC5 PC5 VC2 Pct6 Pct6 Pct6 PC5 PC3 Pct6 Pct6 PC3 VC2 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Chunk-based fuzzy reordering rules Reordered source LM Reordering selection w- $Ark and took part CC1 CC1 CC1 CC1 CC1 CC1 23 VC2 PC3 PC3 PC3 NC4 NC4 fy AltZAhrp E$rAt AlmslHyn in the march mn AlktA}b . dozens of militants from the Brigades PC3 VC2 NC4 NC4 VC2 PC5 NC4 PC5 Pct6 NC4 VC2 PC5 PC5 PC5 VC2 Pct6 Pct6 Pct6 PC5 PC3 Pct6 Pct6 PC3 VC2 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT 0.7 0.1 0.1 0.4 0.9 Chunk-based fuzzy reordering rules Reordered source LM Reordering selection w- $Ark and took part CC1 CC1 VC2 PC3 fy AltZAhrp E$rAt AlmslHyn in the march mn AlktA}b . dozens of militants from the Brigades PC3 VC2 NC4 PC5 Pct6 NC4 PC5 Pct6 0.7 Pct6 0.1 0.1 0.4 0.9 Reorderings to encode in the distortion matrix CC1 24 NC4 PC5 VC2 PC3 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT CC1 VC2 Modifying the distortion matrix CC1 VC2 PC3 NC4 PC5 Pct6 CC1 PC3 VC2 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 3 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 PC3 w2 2 1 0 2 3 4 5 6 7 NC4 w3 3 2 1 0 2 3 4 5 6 NC4 w4 4 3 2 1 0 2 3 4 5 w5 5 4 3 2 1 0 2 3 4 PC5 w6 6 5 4 3 2 1 0 2 3 w7 7 6 5 4 3 2 1 0 Pct6 w8 8 7 6 5 4 3 2 1 0 2 PC5 Pct6 Reorderings to encode in the distortion matrix CC1 25 NC4 PC5 VC2 PC3 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Pct6 CC1 VC2 Modifying the distortion matrix CC1 VC2 PC3 NC4 PC5 Pct6 CC1 PC3 VC2 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 3 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 PC3 w2 2 0 0 2 3 4 5 6 7 NC4 w3 3 0 1 0 2 3 4 5 6 NC4 w4 4 3 2 1 0 2 3 4 5 w5 5 4 3 2 1 0 2 3 4 PC5 w6 6 5 4 3 2 1 0 2 3 w7 7 6 5 4 3 2 1 0 Pct6 w8 8 7 6 5 4 3 2 1 0 2 PC5 Pct6 Reorderings to encode in the distortion matrix CC1 26 NC4 PC5 VC2 PC3 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Pct6 CC1 VC2 Modifying the distortion matrix CC1 VC2 PC3 NC4 PC5 Pct6 CC1 PC3 VC2 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 2 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 PC3 w2 2 0 0 2 3 4 5 6 7 NC4 w3 3 0 1 0 2 3 4 5 6 NC4 w4 4 3 2 1 0 2 3 4 5 w5 5 4 3 2 1 0 2 3 4 PC5 w6 6 5 4 3 2 1 0 2 3 w7 7 6 5 4 3 2 1 0 Pct6 w8 8 7 6 5 4 3 2 1 0 2 PC5 Pct6 Reorderings to encode in the distortion matrix CC1 27 NC4 PC5 VC2 PC3 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Pct6 CC1 VC2 Modifying the distortion matrix CC1 VC2 PC3 NC4 PC5 Pct6 CC1 PC3 VC2 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 2 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 PC3 w2 2 0 0 2 3 4 5 6 7 NC4 w3 3 0 1 0 2 3 4 5 6 NC4 w4 4 3 0 1 0 2 3 4 5 w5 5 4 0 2 1 0 2 3 4 PC5 w6 6 5 4 3 2 1 0 2 3 w7 7 6 5 4 3 2 1 0 Pct6 w8 8 7 6 5 4 3 2 1 0 2 PC5 Pct6 Reorderings to encode in the distortion matrix CC1 28 NC4 PC5 VC2 PC3 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Pct6 CC1 VC2 Modifying the distortion matrix CC1 VC2 PC3 NC4 PC5 Pct6 CC1 PC3 VC2 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 2 w4 5 4 w5 6 5 w6 7 6 w7 8 7 w8 9 8 PC3 w2 2 0 0 2 3 4 5 6 7 NC4 w3 3 0 1 0 2 3 4 5 6 NC4 w4 4 0 0 1 0 2 3 4 5 w5 5 0 0 2 1 0 2 3 4 PC5 w6 6 5 4 3 2 1 0 2 3 w7 7 6 5 4 3 2 1 0 Pct6 w8 8 7 6 5 4 3 2 1 0 2 PC5 Pct6 Reorderings to encode in the distortion matrix CC1 29 NC4 PC5 VC2 PC3 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Pct6 CC1 VC2 Modifying the distortion matrix CC1 VC2 PC3 NC4 PC5 Pct6 CC1 PC3 VC2 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 2 w4 5 4 w5 6 5 w6 7 2 w7 8 2 w8 9 8 PC3 w2 2 0 0 2 3 4 5 6 7 NC4 w3 3 0 1 0 2 3 4 5 6 NC4 w4 4 0 0 1 0 2 3 4 5 w5 5 0 0 2 1 0 2 3 4 PC5 w6 6 5 4 3 2 1 0 2 3 w7 7 6 5 4 3 2 1 0 Pct6 w8 8 7 6 5 4 3 2 1 0 2 PC5 Pct6 Reorderings to encode in the distortion matrix CC1 30 NC4 PC5 VC2 PC3 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Pct6 CC1 VC2 Modifying the distortion matrix CC1 VC2 PC3 NC4 PC5 Pct6 CC1 PC3 VC2 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 2 w4 5 4 w5 6 5 w6 7 2 w7 8 2 w8 9 8 PC3 w2 2 0 0 2 3 4 5 6 7 NC4 w3 3 0 1 0 2 3 4 5 6 NC4 w4 4 0 0 1 0 2 3 4 5 w5 5 0 0 2 1 0 2 3 4 PC5 w6 6 5 4 3 2 1 0 2 3 w7 7 6 5 4 3 2 1 0 Pct6 w8 8 7 6 0 0 3 2 1 0 2 PC5 Pct6 Reorderings to encode in the distortion matrix CC1 31 NC4 PC5 VC2 PC3 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Pct6 CC1 VC2 Modifying the distortion matrix CC1 VC2 PC3 NC4 PC5 Pct6 CC1 PC3 VC2 w0 w1 <s> 0 1 w0 0 w1 2 w2 3 2 w3 4 2 w4 5 4 w5 6 5 w6 7 2 w7 8 2 w8 9 8 PC3 w2 2 0 0 2 3 4 5 6 7 NC4 w3 3 0 1 0 2 3 4 5 6 NC4 w4 4 0 0 1 0 2 3 4 5 w5 5 0 0 2 1 0 2 3 4 PC5 w6 6 5 4 3 2 1 0 2 3 w7 7 6 5 4 3 2 1 0 Pct6 w8 8 7 6 0 0 3 2 1 0 2 PC5 Pct6 Reorderings to encode in the distortion matrix CC1 32 NC4 PC5 VC2 PC3 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Pct6 Experiments • Tasks: NIST-MT09 for Ar-En, WMT10 for De-En • Systems based on Moses, include state-of-the-art hierarchical lexicalized reordering models [Tillmann 04; Koehn & al 05; Galley & Manning 08] • Baseline Distortion Limits: 5 in Ar-En, 10 in De-En • Evaluation by: - BLEU for lexical match & local order - KRS for global order 33 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Arabic-English: BLEU modif‐DL5 50.9 +0.9 BLEU +0.6 KRS (signif.) 50.7 50.5 50.3 50.1 49.9 Translation Time plain‐DL5 plain‐DL8 49.7 83.0 Translation Quality 83.5 84.0 KRS 84.5 85.0 ms/word 389 400 350 Test set: eval09-NW Distortion modified with 3-best reorderings per rule-matching sequence 300 250 200 plain‐DL5 35 273 263 plain‐DL8 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT modif‐DL5 German-English: BLEU modif‐DL4 +0.4 BLEU +0.7 KRS (signif.) 20.6 20.4 20.2 Translation Quality plain‐DL10 Translation Time 20.0 plain‐DL20 19.8 plain‐DL4 19.6 66.0 67.0 68.0 69.0 KRS 70.0 71.0 350 292 300 Test set: newstest10 Distortion modified with 3-best reorderings per rule-matching sequence 369 ms/word 250 200 158 163 150 100 plain‐DL4 plain‐DL10 plain‐DL20 modif‐DL4 37 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT Conclusions • Modified distortion allows for finer & linguistically motivated definition of search space • We achieve better translation & faster decoding in language pairs where long reordering concentrates on few patterns • Our method is complementary to reordering modeling • For now, few reordering rules are needed to modify distortion • We are currently working on a fully data-driven approach to replace the rules 38 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT w0 w1 <s> 0 1 w0 0 w1 2 w2 3 T w3 4 H w4 5 A w5 6 N w6 7 K w7 8 S w8 9 8 w9 10 9 w10 11 10 39 w2 2 1 0 2 T 4 5 6 7 8 9 w3 3 2 1 0 T 3 4 5 6 7 8 w4 4 3 2 1 0 E 2 3 4 5 6 7 w5 5 4 3 2 1 N 2 3 4 5 6 w6 6 5 4 3 2 T 0 F 2 3 4 5 w7 7 6 5 4 3 I 1 O 2 3 4 w8 8 7 6 5 Y O U R 0 w9 w10 9 10 8 9 7 8 6 7 5 6 N ! 3 4 2 3 1 2 0 1 2 0 3 2 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT w0 w1 <s> 0 1 w0 0 w1 2 w2 3 T w3 4 H w4 5 A w5 6 N w6 7 K w7 8 S w8 9 8 w9 10 9 w10 11 10 40 w2 2 1 0 2 T 4 5 6 7 8 9 w3 3 2 1 0 T 3 4 5 6 7 8 w4 4 3 2 1 0 E 2 3 4 5 6 7 w5 5 4 3 2 1 N 2 3 4 5 6 w6 6 5 4 3 2 T 0 F 2 3 4 5 w7 7 6 5 4 3 I 1 O 2 3 4 w8 8 7 6 5 Y O U R 0 w9 w10 9 10 8 9 7 8 6 7 5 6 N ! 3 4 2 3 1 2 0 1 2 0 3 2 A. Bisazza & M. Federico – Modified Distortion Matrices for PSMT