Datasets:
Tauhait
/
JargonDetection

Dataset card Data Studio Files
xet
 Community
tokens
list
label_ids
list
[ "As", "such", ",", "they", "do", "not", "investigate", "different", "aspects", "of", "generalisability", "such", "as", "the", "effect", "of", "swapping", "the", "source", "and", "target", "datasets", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "More", "importantly", ",", "since", "generalisability", "is", "not", "the", "main", "focus", "of", "these", "works", ",", "they", "do", "not", "investigate", "the", "reason", "behind", "lack", "or", "existence", "of", "the", "generalisability", "between", "tw...
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0 ]
[ "In", "addition", ",", "the", "fact", "that", "~", "uses", "the", "Benign", "class", "of", "the", "same", "dataset", "for", "the", "training", "and", "evaluation", ",", "and", "the", "importance", "of", "the", "Benign", "traffic", "in", "ML", "-", "based...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Table", "~", "shows", "the", "summary", "information", "of", "the", "four", "datasets", "used", "in", "this", "study", ",", "all", "in", "NetFlow", "(", "NF", ")", "format", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "These", "datasets", ",", "which", "include", "NFv2", "-UNSW-", "NB15", ",", "NFv2-CIC-2018", ",", "NFv2-ToN-IoT", "and", "NFv2-BoT-IoT", "~", ",", "are", "converted", "from", "their", "original", "formats", "published", "as", "UNSW", "-", "NB15", "~", ",", ...
[ 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 0 ]
[ "The", "first", "version", "of", "the", "NetFlow", "format", "datasets", "(", "NFv1", ")", "with", "20", "features", "has", "been", "published", "in", "~", ",", "and", "the", "second", "version", ",", "that", "is", "used", "in", "this", "work", "and", ...
[ 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "The", "procedure", "for", "converting", "the", "original", "format", "into", "NetFlow", "and", "the", "labelling", "of", "the", "converted", "flows", "are", "also", "explained", "in", "~", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
[ "The", "three", "original", "datasets", "UNSW", "-", "NB15", ",", "ToN-", "IoT", "and", "BoT", "-", "IoT", "are", "published", "by", "same", "research", "group", "." ]
[ 0, 0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "Since", "the", "network", "setup", "used", "for", "the", "traffic", "generation", "for", "each", "dataset", "is", "very", "different", ",", "they", "represent", "different", "network", "environments", "and", "represent", "a", "valid", "basis", "for", "the", "...
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0 ]
[ "This", "is", "also", "the", "case", "for", "CIC", "-", "2018", ",", "the", "fourth", "considered", "dataset", ",", "which", "has", "been", "generated", "by", "a", "different", "research", "group", "in", "a", "completely", "different", "network", "setup", ...
[ 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Although", "these", "models", "are", "typically", "considered", "unsupervised", "learning", "algorithms", ",", "since", "they", "are", "very", "sensitive", "to", "anomalies", "/", "outliers", "in", "the", "training", "set", "~", ",", "we", "have", "used", "th...
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "This", "means", "the", "models", "are", "exposed", "to", "one", "class", "of", "data", "during", "the", "training", "phase", ",", "but", "are", "tested", "against", "both", "classes", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "To", "achieve", "this", ",", "each", "benchmark", "dataset", "is", "divided", "into", "the", "training", "and", "test", "subsets", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Then", ",", "the", "Attack", "samples", "are", "removed", "from", "the", "training", "subset", "." ]
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0 ]
[ "Hence", ",", "the", "models", "are", "trained", "only", "on", "normal", "/", "Benign", "data", "samples", "and", "evaluated", "against", "the", "test", "subset", ",", "which", "includes", "samples", "of", "both", "the", "Benign", "and", "Attack", "classes",...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Similar", "to", "the", "case", "of", "the", "supervised", "learning", "models", ",", "we", "did", "not", "fine", "tune", "the", "hyperparameters", "of", "the", "models", ",", "for", "achieving", "the", "best", "possible", "performances", ",", "and", "mostly...
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Table", "~", "shows", "the", "results", "of", "applying", "these", "three", "NIDS", "models", "on", "the", "four", "considered", "benchmark", "datasets", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "First", ",", "all", "models", "have", "lower", "performance", "on", "NFv2", "-", "ToN", "-", "IoT", ",", "which", "was", "also", "the", "case", "for", "the", "supervised", "learning", "models", "(", "except", "the", "Feed", "Forward", "model", ")", "." ...
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0 ]
[ "Second", ",", "the", "Isolation", "Forest", "model", "has", "higher", "performances", "across", "all", "the", "four", "datasets", "." ]
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Generalisability", "Evaluation", "%" ]
[ 0, 1, 0 ]
[ "In", "the", "previous", "section", "we", "evaluated", "the", "performance", "of", "both", "supervised", "and", "unsupervised", "learning", "NIDS", "algorithms", "via", "a", "single", "domain", "approach", ",", "which", "is", "the", "current", "de-facto", "stand...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "As", "it", "was", "shown", ",", "we", "were", "able", "to", "get", "close", "to", "the", "high", "performance", "values", "reported", "in", "the", "literature", ",", "using", "our", "simple", "non-fine", "-", "tuned", "models", ",", "for", "most", "of",...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Since", "in", "the", "next", "steps", "of", "this", "study", "we", "are", "going", "to", "evaluate", "the", "performance", "of", "these", "models", "via", "the", "multi-domain", "approach", ",", "hyperparameter", "fine", "tuning", "will", "bias", "the", "pe...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Accordingly", ",", "in", "order", "to", "avoid", "over-fitting", "to", "the", "training", "datasets", ",", "we", "used", "the", "default", "hyperparameters", "for", "the", "different", "models", ",", "as", "provided", "by", "the", "corresponding", "software", ...
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0 ]
[ "Table", "~", "shows", "the", "classification", "performance", "(", "F1-", "Score", ")", "for", "the", "four", "supervised", "learning", "models", "trained", "and", "evaluated", "on", "the", "same", "NetFlow", "dataset", "(", "single", "-", "domain", "evaluati...
[ 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0 ]
[ "As", "can", "be", "seen", ",", "the", "performance", "on", "each", "dataset", "are", "mostly", "consistent", "across", "all", "four", "ML", "-", "based", "NIDSs", "(", "1\\%", "to", "5\\%", "variations", ")", ",", "except", "the", "Feed", "Forward", "NN...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "It", "is", "noticeable", "that", "even", "these", "simple", "models", ",", "without", "fine", "tuning", "of", "their", "hyperparameters", ",", "are", "able", "to", "achieve", "a", "very", "high", "classification", "performance", "in", "most", "cases", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "For", "instance", ",", "all", "these", "simple", "models", "have", "been", "able", "to", "achieve", "a", "F1", "-", "Score", "above", "99\\%", "on", "the", "NFv2", "-", "BoT", "-", "IoT", "dataset", ",", "and", "for", "the", "rest", "of", "the", "da...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0 ]
[ "Unsupervised", "Learning", "%" ]
[ 0, 1, 0 ]
[ "While", "the", "generalisability", "of", "supervised", "learning", "models", "has", "partly", "been", "investigated", "for", "the", "NIDS", "application", "in", "a", "previous", "study", "~", ",", "there", "are", "no", "such", "results", "for", "unsupervised", ...
[ 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "As", "such", ",", "we", "have", "included", "three", "unsupervised", "learning", "models", "in", "our", "investigation", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0 ]
[ "These", "models", ",", "which", "include", "the", "Isolation", "Forest", "(", "IsolationForest", ")", "~", ",", "One", "-", "Class", "Support", "Vector", "Machines", "(", "oSVM", ")", "~", ",", "and", "Stochastic", "Gradient", "Descent", "one", "-", "Clas...
[ 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Supervised", "Learning" ]
[ 0, 1 ]
[ "Initially", ",", "we", "evaluate", "supervised", "learning", "methods", "via", "the", "single", "domain", "evaluation", "approach", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0 ]
[ "There", "are", "many", "previous", "studies", "which", "achieve", "a", "very", "high", "detection", "performance", "using", "this", "approach", ",", "based", "on", "the", "same", "benchmark", "NIDS", "datasets", "that", "we", "consider", "in", "this", "paper"...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 0, 0 ]
[ "Since", "the", "NetFlow", "version", "of", "these", "datasets", "have", "been", "published", "relatively", "recently", ",", "there", "are", "only", "a", "few", "studies", "such", "as", "~", ",", "and", ",", "that", "have", "used", "them", "." ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "However", ",", "we", "cannot", "use", "their", "results", "in", "our", "comparisons", "because", "we", "need", "to", "evaluate", "the", "same", "model", "later", "in", "the", "multi-domain", "setup", ",", "and", "these", "studies", "do", "not", "provide", ...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "As", "such", ",", "we", "use", "the", "result", "of", "our", "experiment", ",", "even", "for", "the", "single", "-", "domain", "evaluation", ",", "which", "might", "be", "found", "in", "the", "literature", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "We", "chose", "four", "simple", "supervised", "learning", "models", "including", "two", "deep", "and", "two", "shallow", "learning", "methods", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "For", "the", "deep", "learning", "we", "chose", "the", "same", "number", "of", "layers", "and", "nodes", "from", "two", "different", "architectures", "." ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0 ]
[ "The", "first", "model", "is", "a", "simple", "Feed", "Forward", "neural", "network", "with", "5", "hidden", "layers", ",", "each", "with", "10", "nodes", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0 ]
[ "The", "second", "model", "is", "a", "Long", "Short", "-", "Term", "Memory", "(", "LSTM", ")", "network", "with", "the", "same", "number", "of", "layers", "and", "nodes", "on", "each", "layer", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0 ]
[ "For", "the", "shallow", "learning", "methods", ",", "we", "used", "a", "Random", "Forest", "and", "an", "Extra-", "Tree", "classifier", ",", "which", "allow", "us", "to", "easily", "implement", "our", "NIDSs", "without", "much", "effort", "to", "tune", "h...
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0 ]
[ "It", "is", "possible", "to", "search", "and", "fine", "tune", "the", "hyperparameters", "of", "the", "ML", "-", "models", "to", "achieve", "the", "best", "possible", "performances", "on", "a", "training", "dataset", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "However", ",", "this", "increases", "the", "chance", "of", "over-fitting", "to", "the", "training", "dataset", "and", "reduces", "the", "performance", "on", "other", "datasets", ",", "not", "seen", "during", "the", "training", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "While", "the", "accuracy", ",", "detection", "rate", "and", "other", "performance", "metrics", "reported", "in", "many", "of", "these", "studies", "are", "near", "perfect", ",", "and", "the", "studies", "are", "designed", "and", "performed", "rigorously", ","...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "We", "believe", "that", "an", "investigation", "into", "the", "generalisability", "properties", "of", "ML", "-", "based", "NIDSs", "is", "an", "important", ",", "but", "so", "far", "under", "-investigated", "step", "towards", "bridging", "the", "gap", "betwee...
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "In", "particular", ",", "we", "assume", "that", "the", "setup", "/", "environment", "in", "which", "benchmark", "NIDS", "datasets", "are", "created", "/", "collected", "varies", "among", "the", "different", "available", "benchmark", "datasets", ",", "and", "m...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Applying", "machine", "learning", "for", "detecting", "unseen", "datasets", "and", "unseen", "classes", "of", "the", "same", "datasets", "has", "been", "explored", "in", "many", "previous", "works", "." ]
[ 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "This", "includes", "studies", "that", "investigate", "zero", "-", "day", "/", "unseen", "attack", "detection", "such", "as", "~", "and", "or", "studies", "that", "examine", "domain", "adaptation", "and", "transfer", "learning", "in", "the", "field", "of", "...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 ]
[ "Though", ",", "none", "of", "these", "works", "specifically", "investigates", "the", "generalisability", "in", "ML", "-", "based", "NIDSs", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0 ]
[ "Perhaps", "the", "only", "previous", "works", "that", "partially", "/", "indirectly", "investigate", "the", "generalisability", "of", "ML", "-", "based", "NIDSs", "are", "~", "and", "in", "which", "the", "cross-evaluation", "of", "ML", "-", "based", "NIDSs", ...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0 ]
[ "Apart", "from", "these", "studies", "that", "have", "been", "discussed", "in", "the", "related", "works", "of", "this", "paper", ",", "we", "have", "not", "been", "able", "to", "find", "other", "works", "considering", "the", "generalisability", "of", "these...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0 ]
[ "Our", "work", "aims", "to", "address", "this", "gap", "by", "taking", "advantage", "of", "four", "recently", "published", "NIDS", "benchmark", "datasets", "~", ",", "which", "have", "been", "converted", "from", "their", "original", "format", "into", "a", "c...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "We", "evaluate", "the", "generalisability", "in", "both", "directions", "across", "different", "datasets", ",", "i.e.", ",", "each", "dataset", "is", "used", "as", "the", "training", "and", "test", "dataset", ",", "against", "all", "other", "datasets", "." ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
[ "In", "addition", ",", "we", "evaluate", "the", "generalisability", "of", "both", "supervised", "as", "well", "as", "unsupervised", "learning", "models", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "The", "next", "two", "model", ",", "oSVM", "and", "SGD", "-", "oSVM", ",", "shown", "in", "Figure", "~-", "(", "b", ")", "and", "(", "c", ")", "respectively", ",", "have", "much", "lower", "variations", "compared", "to", "Isolation", "Forest", "." ]
[ 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "As", "can", "be", "seen", ",", "the", "variance", "of", "the", "mean", "SHAP", "values", "of", "features", "in", "the", "case", "of", "Isolation", "Forest", ",", "Figure", "~-", "(", "a", ")", ",", "is", "significantly", "larger", "than", "the", "othe...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "This", "can", "be", "easily", "verified", "by", "comparing", "the", "average", "decay", "per", "model", "as", "shown", "in", "Table", "~", ",", "which", "is", "49.73\\%", ",", "17.90\\%", "and", "17.36\\%", "for", "the", "Isolation", "Forest", ",", "oSVM"...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 0, 0, 0 ]
[ "The", "oSVM", "and", "SGD", "-", "oSVM", "models", ",", "in", "addition", "to", "the", "lower", "variance", "of", "mean", "SHAP", "values", ",", "share", "the", "same", "order", "/", "rank", "for", "many", "of", "the", "features", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "This", "similarity", "in", "the", "distribution", "of", "mean", "SHAP", "value", "of", "features", ",", "which", "is", "an", "indicator", "of", "model", "behaviour", "similarity", ",", "explains", "their", "close", "average", "model", "decays", "(", "17.90\\%...
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
[ "To", "answer", "this", "kind", "of", "questions", "about", "the", "behaviour", "of", "the", "models", ",", "we", "need", "a", "more", "detailed", "analysis", "of", "the", "SHAP", "values", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "As", "such", ",", "it", "is", "expected", "that", "their", "behaviour", "is", "more", "similar", "across", "different", "combinations", "of", "source", "-", "targets", ",", "which", "means", "these", "two", "model", "are", "more", "generalisable", "compared",...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Similarly", ",", "the", "considerable", "variations", "of", "the", "importance", "and", "order", "of", "the", "features", "across", "different", "source", "-", "target", "combinations", "for", "the", "Isolation", "Forest", "model", "indicates", "variations", "of"...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "There", "is", "no", "supervised", "learning", "-", "based", "NIDS", "generalising", "over", "all", "combination", "of", "source", "-", "target", "domains", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "\\item", "Deep", "learning", "-", "based", "NIDSs", "generalise", "better", "than", "shallow", "learning", "-", "based", "NIDSs", "in", "average", "." ]
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0 ]
[ "Figure", "~", "visualises", "the", "results", "for", "the", "three", "considered", "unsupervised", "algorithms", ",", "and", "also", "includes", "the", "single", "-", "domain", "results", "on", "the", "diagonal", ",", "corresponding", "to", "Figure", "~", "."...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Table", "~", "shows", "the", "results", "of", "generalisability", "evaluation", "of", "the", "unsupervised", "learning", "algorithms", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Similar", "to", "the", "supervised", "learning", "algorithms", ",", "we", "evaluated", "the", "generalisability", "of", "unsupervised", "learning", "-", "based", "NIDSs", "using", "the", "same", "four", "benchmark", "datasets", "as", "we", "used", "for", "the",...
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "There", "is", "an", "average", "performance", "decay", "of", "56.28\\%", "when", "a", "supervised", "learning", "model", "is", "evaluated", "on", "a", "dataset", "other", "than", "its", "training", "dataset", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "As", "per", "the", "single", "domain", "evaluation", ",", "training", "is", "performed", "using", "only", "the", "normal", "/", "Benign", "class", "of", "the", "source", "domain", ",", "while", "the", "trained", "model", "is", "exposed", "to", "both", "th...
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "For", "each", "of", "the", "four", "datasets", "used", "as", "the", "source", "domain", ",", "we", "consider", "the", "other", "three", "as", "the", "target", "domain", ",", "resulting", "in", "12", "target", "/", "source", "domain", "combinations", "that...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "As", "can", "be", "seen", ",", "while", "the", "performance", "of", "the", "models", "is", "not", "as", "high", "as", "the", "supervised", "learning", "algorithms", ",", "in", "the", "single", "domain", "experiments", ",", "there", "are", "many", "cases",...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Similar", "to", "supervised", "learning", "algorithms", ",", "each", "model", "is", "trained", "on", "one", "dataset", "and", "evaluated", "on", "the", "other", "three", "datasets", "that", "were", "not", "used", "for", "training", "." ]
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Many", "of", "the", "proposed", "machine", "learning", "(", "ML", ")", "based", "network", "intrusion", "detection", "systems", "(", "NIDSs", ")", "achieve", "near", "perfect", "detection", "performance", "when", "evaluated", "on", "synthetic", "benchmark", "da...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Though", ",", "there", "is", "no", "record", "of", "if", "and", "how", "these", "results", "generalise", "to", "other", "network", "scenarios", ",", "in", "particular", "to", "real", "-", "world", "networks", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "In", "this", "paper", ",", "we", "investigate", "the", "generalisability", "property", "of", "ML", "-", "based", "NIDSs", "by", "extensively", "evaluating", "seven", "supervised", "and", "unsupervised", "learning", "models", "on", "four", "recently", "published",...
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Interestingly", ",", "our", "results", "also", "indicate", "that", "the", "generalisability", "has", "a", "high", "degree", "of", "asymmetry", ",", "i.e.", ",", "swapping", "the", "source", "and", "target", "domains", "can", "significantly", "change", "the", ...
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "Our", "investigation", "also", "indicates", "that", "overall", ",", "unsupervised", "learning", "methods", "generalise", "better", "than", "supervised", "learning", "models", "in", "our", "considered", "scenarios", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 ]
[ "Using", "SHAP", "values", "to", "explain", "these", "results", "indicates", "that", "the", "lack", "of", "generalisability", "is", "mainly", "due", "to", "the", "presence", "of", "strong", "correspondence", "between", "the", "values", "of", "one", "or", "more...
[ 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "A", "quick", "search", "of", "the", "academic", "literature", "in", "network", "intrusion", "detection", "systems", "(", "NIDSs", ")", "reveals", "that", "there", "are", "hundreds", "of", "proposals", "based", "on", "machine", "learning", "(", "ML", ")", "a...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "In", "most", "of", "these", "proposals", ",", "the", "performance", "is", "evaluated", "using", "a", "publicly", "available", "benchmark", "NIDS", "dataset", "such", "as", "UNSW", "~", ",", "CIC", "-", "IDS", "~", ",", "and", "KDD99", "~", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 1, 0, 0 ]
[ "For", "this", "purpose", ",", "like", "in", "other", "fields", "that", "use", "machine", "learning", ",", "the", "benchmark", "dataset", "is", "divided", "into", "the", "training", "and", "test", "subsets", ",", "the", "ML", "models", "are", "trained", "o...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "That", "is", ",", "the", "ML", "model", "is", "trained", "and", "tested", "on", "each", "benchmark", "dataset", "separately", "." ]
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Figure", "~-", "(", "a", ")", "and", "(", "b", ")", "are", "examples", "of", "SHAP", "summary", "plots", "for", "such", "a", "case", "in", "which", "the", "Feed", "Forward", "model", "is", "trained", "on", "NFv2", "-", "UNSW", "-", "NB15", "dataset"...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Most", "of", "the", "Blue", "dots", "(", "low", "feature", "values", ")", "have", "positive", "SHAP", "values", "(", "Benign", ")", "and", "most", "of", "the", "Red", "dots", "(", "high", "feature", "values", ")", "have", "negative", "SHAP", "values", ...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0 ]
[ "The", "Extra", "Tree", "model", "in", "(", "a", ")", "is", "trained", "on", "NFv2", "-", "UNSW", "-", "NB15", "dataset", "and", "evaluated", "on", "NFv2", "-", "BoT", "-", "IoT", "dataset", ",", "and", "in", "(", "b", ")", "it", "is", "trained", ...
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Figure", "~", "-", "(", "a", ")", "and", "(", "b", ")", "illustrate", "an", "example", "of", "SHAP", "summary", "plots", "for", "such", "a", "case", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 ]
[ "This", "is", "obviously", "due", "to", "the", "difference", "of", "the", "feature", "distribution", "in", "the", "new", "target", "domain", ",", "NFv2", "-", "ToN", "-", "IoT", "with", "the", "source", "domain", "NFv2", "-", "UNSW", "-", "NB15", "datase...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "The", "result", "shown", "in", "Figure", "~-", "(", "c", ")", "confirms", "our", "conclusion", "from", "the", "SHAP", "summary", "plots", "and", "show", "a", "significant", "performance", "drop", "for", "the", "this", "case", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "The", "other", "aspect", "of", "generalisability", "observed", "in", "our", "results", "is", "its", "asymmetric", "behaviour", ",", "i.e.", ",", "while", "a", "model", "has", "a", "high", "performance", "in", "a", "combination", "of", "the", "source", "-", ...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0 ]
[ "This", "is", "a", "simple", "multi-rule", "classifier", "that", "can", "separate", "the", "Attack", "and", "Benign", "classes", "based", "on", "the", "simple", "rules", "in", "terms", "of", "the", "feature", "values", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "In", "Figure", "~-", "(", "b", ")", ",", "however", ",", "this", "correspondence", "of", "the", "feature", "values", "with", "the", "Attack", "and", "Benign", "classes", "hardly", "can", "be", "seen", "and", "most", "of", "the", "feature", "values", "ei...
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
[ "As", "can", "be", "seen", ",", "in", "Figure", "~-", "(", "a", ")", "the", "feature", "values", "have", "a", "clear", "correspondence", "with", "the", "Benign", "and", "Attack", "classes", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "SHapley", "Additive", "exPlanations", "(", "SHAP", ")", "values", "~", "is", "one", "of", "the", "recent", "trends", "in", "explaining", "and", "interpreting", "the", "output", "of", "the", "AI", "/", "ML", "models", "in", "terms", "of", "the", "features"...
[ 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Explaining", "the", "behavior", "of", "a", "machine", "learning", "model", "usually", "requires", "to", "investigate", "the", "impact", "of", "the", "features", "of", "input", "data", "on", "output", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Hence", ",", "these", "values", "depend", "on", "the", "training", "dataset", ",", "the", "ML", "model", ",", "and", "the", "evaluation", "dataset", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
[ "Table", "~", "quantifies", "these", "outcomes", "in", "terms", "of", "the", "performance", "(", "F1", "-", "Score", ")", "decays", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0 ]
[ "Each", "column", "indicates", "an", "unsupervised", "learning", "model", ",", "each", "row", "indicates", "a", "source", "domain", ",", "and", "each", "value", "indicates", "the", "average", "of", "three", "F1", "-", "Score", "values", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "There", "is", "no", "unsupervised", "learning", "model", "that", "generalises", "over", "all", "combination", "of", "source", "-", "target", "domains", "." ]
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]