List your company

Byzantine Attack Detection on FashionMNIST (test)

100Detection AccuracyClassical MultiKrum
58.469.28090.8May 14, 2026

Evaluation results

MethodLinks
Classical MultiKrumAttack=Gaussian Noise, Rounds=302026.05100100
Classical MultiKrumAttack=Scale, Rounds=302026.05100100
Classical MultiKrumAttack=Targeted, Rounds=302026.05100100
Classical MultiKrumAttack=Clustered, Rounds=302026.05100100
Classical MultiKrumAttack=Same Value, Rounds=302026.05100100
Classical MultiKrumAttack=Blatant Lie, Rounds=302026.05100100
Classical MultiKrumAttack=Label Flip, Rounds=302026.05100100
Classical MultiKrumAttack=Stealthy, Rounds=302026.05100100
QUBO-based client selectionAttack=ALIE, Rounds=302026.0595.1187.78
QUBO-based client selectionAttack=Blatant Lie, Rounds=302026.058870
QUBO-based client selectionAttack=Lie, Rounds=302026.0586.6766.67
QUBO-based client selectionAttack=Stealthy, Rounds=302026.0586.2265.56
Classical MultiKrumAttack=ALIE, Rounds=302026.0581.3353.33
QUBO-based client selectionAttack=Sparse Lie, Rounds=302026.0578.2245.56
Classical MultiKrumAttack=Lie, Rounds=302026.0574.6736.67
QUBO-based client selectionAttack=Targeted, Rounds=302026.0574.2235.56
QUBO-based client selectionAttack=Clustered, Rounds=302026.0573.3333.33
QUBO-based client selectionAttack=Same Value, Rounds=302026.0573.3333.33
Classical MultiKrumAttack=Shuffle, Rounds=302026.0570.2225.56
QUBO-based client selectionAttack=Sign Flip, Rounds=302026.0569.7824.44
Classical MultiKrumAttack=Sparse Lie, Rounds=302026.0565.7814.44
QUBO-based client selectionAttack=Gaussian Noise, Rounds=302026.05600
Classical MultiKrumAttack=Sign Flip, Rounds=302026.05600
QUBO-based client selectionAttack=Scale, Rounds=302026.05600
QUBO-based client selectionAttack=Label Flip, Rounds=302026.05600
QUBO-based client selectionAttack=Shuffle, Rounds=302026.05600