MT-SLVR: Multi-Task Self-Supervised Learning for Transformation In(Variant) Representations

Open paper
Benchmark Model Rank Results
few-shot-audio-classification-onMT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#6Top-1 Accuracy(5-Way-1-Shot): 39.11±0.41
few-shot-audio-classification-onSimCLR (FSD50K, RN18)#8Top-1 Accuracy(5-Way-1-Shot): 37.64±0.40
few-shot-audio-classification-onMulti-Label Augmentation Prediction (FSD50K, RN18)#10Top-1 Accuracy(5-Way-1-Shot): 21.72±0.34
few-shot-audio-classification-on-birdclefSimCLR (FSD50K, RN18)#8Top-1 Accuracy(5-Way-1-Shot): 30.93±0.38
few-shot-audio-classification-on-birdclefMT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#9Top-1 Accuracy(5-Way-1-Shot): 29.49±0.38
few-shot-audio-classification-on-birdclefMulti-Label Augmentation Prediction (FSD50K, RN18)#10Top-1 Accuracy(5-Way-1-Shot): 21.04±0.35
few-shot-audio-classification-on-common-voiceMT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#1Top-1 Accuracy(5-Way-1-Shot): 35.22±0.40
few-shot-audio-classification-on-common-voiceSimCLR (FSD50K, RN18)#2Top-1 Accuracy(5-Way-1-Shot): 33.33±0.38
few-shot-audio-classification-on-common-voiceMulti-Label Augmentation Prediction (FSD50K, RN18)#3Top-1 Accuracy(5-Way-1-Shot): 23.00±0.42
few-shot-audio-classification-on-crema-dMT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#1Top-1 Accuracy(5-Way-1-Shot): 29.61±0.38
few-shot-audio-classification-on-crema-dSimCLR (FSD50K, RN18)#2Top-1 Accuracy(5-Way-1-Shot): 29.10±0.36
few-shot-audio-classification-on-crema-dMulti-Label Augmentation Prediction (FSD50K, RN18)#3Top-1 Accuracy(5-Way-1-Shot): 21.68±0.33
few-shot-audio-classification-on-esc-50MT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#4Top-1 Accuracy(5-Way-1-Shot): 69.53±0.39
few-shot-audio-classification-on-esc-50SimCLR (FSD50K, RN18)#8Top-1 Accuracy(5-Way-1-Shot): 63.40±0.39
few-shot-audio-classification-on-esc-50Multi-Label Augmentation Prediction (FSD50K, RN18)#10Top-1 Accuracy(5-Way-1-Shot): 37.76±0.34
few-shot-audio-classification-on-nsynthMT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#6Top-1 Accuracy(5-Way-1-Shot): 71.81±0.39
few-shot-audio-classification-on-nsynthSimCLR (FSD50K, RN18)#8Top-1 Accuracy(5-Way-1-Shot): 66.44±0.40
few-shot-audio-classification-on-nsynthMulti-Label Augmentation Prediction (FSD50K, RN18)#10Top-1 Accuracy(5-Way-1-Shot): 62.52±0.36
few-shot-audio-classification-on-speechSimCLR (FSD50K, RN18)#1Top-1 Accuracy(5-Way-1-Shot): 25.68±0.35
few-shot-audio-classification-on-speechMT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#2Top-1 Accuracy(5-Way-1-Shot): 23.65±0.34
few-shot-audio-classification-on-speechMulti-Label Augmentation Prediction (FSD50K, RN18)#3Top-1 Accuracy(5-Way-1-Shot): 20.08±0.37
few-shot-audio-classification-on-speech-1MT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#1Top-1 Accuracy(5-Way-1-Shot): 28.92±0.37
few-shot-audio-classification-on-speech-1SimCLR (FSD50K, RN18)#2Top-1 Accuracy(5-Way-1-Shot): 26.16±0.34
few-shot-audio-classification-on-speech-1Multi-Label Augmentation Prediction (FSD50K, RN18)#3Top-1 Accuracy(5-Way-1-Shot): 23.08±0.34
few-shot-audio-classification-on-voxceleb1MT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#6Top-1 Accuracy(5-Way-1-Shot): 33.58±0.39
few-shot-audio-classification-on-voxceleb1SimCLR (FSD50K, RN18)#7Top-1 Accuracy(5-Way-1-Shot): 31.18±0.37
few-shot-audio-classification-on-voxceleb1Multi-Label Augmentation Prediction (FSD50K, RN18)#10Top-1 Accuracy(5-Way-1-Shot): 21.68±0.40
few-shot-audio-classification-on-watkinsMT-SLVR (SimCLR + MLAP) w/ Parallel Adapters (FSD50K, RN18)#1Top-1 Accuracy(5-Way-1-Shot): 59.49±0.42
few-shot-audio-classification-on-watkinsSimCLR (FSD50K, RN18)#3Top-1 Accuracy(5-Way-1-Shot): 52.91±0.41
few-shot-audio-classification-on-watkinsMulti-Label Augmentation Prediction (FSD50K, RN18)#5Top-1 Accuracy(5-Way-1-Shot): 28.88±0.39

Rank counts only results with a code link.