Datasets:
Tauhait
/
JargonDetection

Dataset card Data Studio Files
xet
 Community
tokens
list
label_ids
list
[ "Similar", "to", "the", "hypothesis", "in", "~", ",", "we", "consider", "that", "malware", "might", "have", "special", "codes", "for", "attacks", ",", "and", "CNNs", "could", "explore", "the", "distance", "between", "malicious", "and", "benign", "samples", "...
[ 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, 1, 0, 0, 1, 0 ]
[ "The", "DSC", "factorizes", "common", "convolution", "into", "depth", "wise", "convolution", "and", "pointwise", "convolution", ",", "which", "is", "proposed", "by", "Howard", "et", "al", "~", "and", "used", "in", "MobileNets", "." ]
[ 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "\\noindent", "Binary", "-", "image", "Converting", "." ]
[ 0, 0, 0, 0, 1, 0 ]
[ "The", "second", "method", "converts", "all", "binary", "codes", "into", "an", "image", "and", "leverages", "image", "classification", "solutions", "to", "perform", "malware", "detection", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "All", "images", "could", "be", "resampled", "to", "the", "same", "size", "using", "the", "bilinear", "interpolation", "algorithm", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "This", "problem", "easily", "makes", "the", "instructions", "far", "from", "each", "other", "in", "the", "original", "sequence", "forced", "to", "be", "integrated", "in", "the", "corresponding", "image", ",", "which", "might", "confuse", "the", "neural", "net...
[ 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 ]
[ "To", "ensure", "the", "consistency", "of", "the", "convolutional", "layer", "inputs", "and", "outputs", ",", "we", "usually", "pad", "some", "zeros", "at", "the", "edge", "of", "the", "input", ",", "and", "the", "neural", "network", "might", "get", "spati...
[ 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, 0, 0, 1, 0, 0 ]
[ "\\item", "Padding", "Problem", ":", "The", "padding", "operation", "might", "make", "the", "model", "hard", "to", "recognize", "the", "beginning", "and", "the", "end", "of", "the", "original", "sequences", ",", "as", "shown", "in", "Figure", "~", "." ]
[ 0, 0, 1, 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 ]
[ "The", "semantic", "loss", "caused", "by", "file", "cutting", "and", "the", "contextual", "confusion", "caused", "by", "binary", "-", "image", "converting", "hinder", "the", "performance", "of", "malware", "detection", "models", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "These", "issues", "might", "confuse", "neural", "networks", "and", "even", "mislead", "them", "to", "make", "diametrically", "opposed", "decisions", "." ]
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Different", "from", "image", "processing", ",", "the", "recognized", "spatial", "information", "possibly", "misleads", "the", "model", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 ]
[ "If", "a", "binary", "instruction", "is", "located", "at", "the", "edge", "of", "the", "image", ",", "line", "breaks", "may", "cut", "off", "the", "instruction", "into", "two", "parts", ",", "as", "shown", "in", "Figure", "~", "." ]
[ 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 ]
[ "Besides", ",", "contextual", "information", "is", "probably", "disrupted", "at", "the", "edge", "due", "to", "the", "break", "of", "strongly", "correlated", "instructions", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "\\item", "Resampling", "Noise", ":", "If", "we", "reshape", "the", "image", "size", ",", "unrelated", "instructions", "in", "different", "lines", "introduce", "contextual", "information", "confusion", ",", "as", "shown", "in", "Figure", "~", "." ]
[ 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 ]
[ "To", "make", "our", "model", "more", "automatic", "and", "avoid", "information", "loss", ",", "we", "focus", "on", "the", "binary", "-", "based", "models", ",", "and", "SeqNet", "applies", "raw", "binaries", "as", "input", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0 ]
[ "We", "apply", "a", "novel", "but", "simple", "binary", "code", "representation", "method", "to", "SeqNet", "and", "maintain", "its", "performance", "when", "reducing", "the", "parameters", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Binary", "Code", "Representation" ]
[ 0, 0, 1 ]
[ "In", "this", "part", ",", "we", "will", "introduce", "several", "major", "binary", "code", "representation", "methods", "applied", "to", "binary", "-", "based", "models", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Therefore", ",", "a", "proper", "binary", "code", "representation", "method", "is", "an", "important", "part", "of", "a", "binary", "-", "based", "malware", "detection", "neural", "network", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Translating", "samples", "into", "the", "input", "of", "neural", "networks", "could", "significantly", "affect", "the", "performance", "of", "the", "models", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Due", "to", "the", "limitation", "of", "memory", "constraints", ",", "many", "works", "set", "an", "artificial", "limit", "on", "the", "maximum", "file", "size", ",", "and", "this", "method", "is", "to", "take", "a", "fixed", "-", "length", "code", "sni...
[ 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, 0, 0, 0, 0, 1, 0 ]
[ "To", "mitigate", "this", "problem", ",", "MalConvGCT", "~", "improved", "MalConv~", "performance", "by", "expanding", "the", "snippet", "size", "limit", "." ]
[ 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "However", ",", "malicious", "codes", "are", "often", "located", "at", "the", "end", "of", "the", "binary", "files", "." ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 ]
[ "File", "cutting", "suffers", "from", "semantical", "information", "loss", "because", "the", "end", "of", "the", "binary", "will", "be", "ignored", "if", "it", "is", "much", "longer", "than", "the", "fixed", "length", "." ]
[ 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 ]
[ "If", "the", "length", "of", "a", "binary", "program", "is", "less", "than", "the", "length", "of", "the", "required", "code", "snippet", ",", "the", "snippet", "would", "be", "padded", "with", "zeros", "at", "the", "end", "." ]
[ 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, 0, 0, 0, 0 ]
[ "Nowadays", ",", "two", "main", "methods", "are", "proposed", "to", "fully", "represent", "the", "raw", "binary", "codes", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Additionally", ",", "lower", "computational", "overhead", "could", "make", "models", "better", "adapt", "to", "the", "evolution", "of", "malware", "and", "expand", "the", "application", "scenarios", ",", "for", "example", ",", "in", "the", "IoT", "environment",...
[ 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, 1, 0 ]
[ "For", "example", ",", "embedded", "viruses", "usually", "embed", "themselves", "at", "the", "tail", "of", "the", "infected", "files", ",", "which", "might", "help", "them", "evade", "the", "detection", "of", "the", "models", "based", "on", "this", "approach...
[ 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 ]
[ "File", "Cutting", "." ]
[ 0, 1, 0 ]
[ "To", "detect", "malware", "more", "efficiently", ",", "many", "researchers", "proposed", "advanced", "tools", "for", "malware", "analysis", "and", "detection", "~." ]
[ 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0 ]
[ "However", ",", "these", "solutions", "could", "not", "fundamentally", "reduce", "their", "workload", "when", "processing", "such", "a", "large", "number", "of", "malware", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "To", "tackle", "this", "problem", ",", "many", "experts", "and", "scholars", "apply", "machine", "learning", "algorithms", ",", "especially", "deep", "learning", ",", "to", "malware", "detection", "and", "classification", "~." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0 ]
[ "For", "example", ",", "the", "model", "trained", "five", "years", "ago", "might", "be", "very", "weak", "in", "malware", "detection", "today", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 ]
[ "Different", "from", "computer", "vision", "and", "natural", "language", "processing", ",", "malware", "is", "evolving", "continuously", "and", "rapidly", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
[ "Nevertheless", ",", "these", "two", "methods", "probably", "suffer", "from", "contextual", "confusion", "and", "semantic", "loss", ",", "which", "will", "be", "discussed", "later", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "To", "realize", "more", "user", "-", "friendly", "and", "automatic", "detection", ",", "binary", "-", "based", "methods", "have", "been", "proposed", "~." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0 ]
[ "The", "two", "epidemic", "approaches", "for", "raw", "binary", "processing", "are", "file", "cutting", "and", "binary", "-", "image", "converting", "." ]
[ 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Additionally", ",", "model", "aging", "is", "a", "crucial", "problem", "for", "neural", "networks", "~." ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "As", "malware", "continues", "to", "evolve", ",", "deep", "learning", "models", "might", "be", "out", "-", "of", "-", "date", "." ]
[ 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Malware", "detection", "is", "a", "battle", "between", "attackers", "and", "detectors", "." ]
[ 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Neural", "networks", "are", "hard", "to", "recognize", "unseen", "malicious", "behaviors", ",", "which", "could", "cause", "lower", "detection", "accuracy", "and", "easier", "evasion", "." ]
[ 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "It", "is", "impossible", "for", "models", "to", "predict", "the", "features", "of", "the", "future", "malware", "but", "feasible", "to", "quickly", "learn", "the", "knowledge", "of", "detecting", "new", "malware", "." ]
[ 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 ]
[ "Therefore", ",", "retraining", "models", "becomes", "one", "of", "the", "few", "methods", "to", "mitigate", "the", "aging", "problem", "." ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "We", "can", "make", "neural", "networks", "quickly", "retrained", "and", "learn", "new", "features", "of", "new", "malware", "so", "that", "they", "could", "recognize", "novel", "attack", "approaches", "." ]
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Their", "efforts", "contribute", "a", "lot", "to", "the", "research", "of", "malware", "analysis", "neural", "networks", "and", "practical", "automatic", "malware", "detection", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0 ]
[ "However", ",", "these", "models", "usually", "require", "various", "feature", "engineering", "to", "help", "neural", "networks", "make", "judgments", ",", "which", "might", "be", "laborious", "and", "easy", "to", "lose", "some", "critical", "information", "." ]
[ 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 ]
[ "In", "this", "paper", ",", "we", "introduce", "SeqNet", ",", "an", "efficient", "automatic", "malware", "detection", "model", "based", "on", "deep", "learning", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0 ]
[ "Compared", "to", "existing", "models", ",", "SeqNet", "has", "a", "much", "smaller", "size", "with", "enough", "detection", "accuracy", ",", "which", "can", "be", "more", "quickly", "trained", "on", "raw", "binaries", "and", "probably", "has", "more", "appl...
[ 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 ]
[ "Additionally", ",", "the", "training", "and", "detection", "process", "of", "SeqNet", "needs", "little", "human", "intervention", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 ]
[ "We", "hope", "this", "model", "will", "inspire", "more", "researchers", "to", "develop", "better", "architectures", "for", "more", "applications", ",", "not", "only", "malware", "detection", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Also", ",", "we", "expect", "that", "deep", "learning", "algorithms", "will", "be", "widely", "used", "in", "practice", "for", "malware", "detection", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "We", "make", "the", "codes", "and", "data", "of", "SeqNet", "available", "to", "the", "research", "community", "to", "promote", "the", "adoption", "of", "SeqNet", "in", "security", "research", "and", "deployment", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
[ "The", "SeqNet", "project", "website", "is", "at", "https://github.com/Darren-8/SeqNet.git.", "\\section*{Acknowledgements}", "%" ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 0 ]
[ "We", "also", "thank", "the", "authors", "of", "VirusShare", "~", ",", "for", "their", "public", "dataset", "used", "in", "our", "evaluation", ",", "and", "VirusTotal", "~", "for", "providing", "us", "the", "malware", "scan", "results", "." ]
[ 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, 1, 0 ]
[ "To", "further", "explore", "the", "performance", "of", "SeqNet", ",", "we", "record", "the", "training", "processes", "and", "the", "prediction", "result", "distribution", "on", "the", "validation", "dataset", "in", "the", "last", "epoch", "." ]
[ 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 ]
[ "We", "see", "SeqNet", "has", "strong", "confidence", "on", "most", "samples", ",", "and", "all", "models", "converge", "quickly", "." ]
[ 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "The", "standard", "SDSC", "blocks", "are", "used", "to", "extract", "the", "features", "when", "downsampling", "the", "sequences", "." ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "For", "the", "first", "convolutional", "layer", ",", "we", "use", "a", "single", "common", "convolutional", "layer", "to", "embed", "the", "original", "input", ",", "and", "the", "size", "of", "the", "kernel", "is", "3", "\\times", "1", "." ]
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 ]
[ "We", "set", "the", "size", "because", "the", "length", "of", "frequently", "used", "CPU", "instructions", "is", "usually", "three", "bytes", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 ]
[ "For", "the", "high", "-", "level", "features", ",", "we", "use", "five", "residual", "SDSC", "blocks", "to", "analyze", ",", "which", "could", "also", "reserve", "the", "contextual", "and", "spatial", "information", "better", "compared", "to", "the", "fully...
[ 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 ]
[ "Besides", ",", "the", "residual", "SDSC", "block", "can", "make", "the", "model", "converge", "rapidly", "by", "preventing", "the", "gradient", "from", "vanishing", "." ]
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0 ]
[ "The", "last", "two", "layers", "are", "a", "fully", "connected", "layer", "and", "a", "softmax", "layer", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0 ]
[ "The", "fully", "connected", "layer", "is", "used", "to", "classify", "the", "analysis", "result", "given", "by", "the", "front", "of", "the", "model", "." ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "The", "softmax", "layer", "could", "translate", "the", "result", "into", "the", "possibility", "format", "using", "the", "following", "formula", "." ]
[ 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "In", "this", "formula", ",", "P", "means", "the", "translated", "result", ",", "and", "x", "means", "the", "input", "of", "the", "softmax", "layer", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "For", "the", "pooling", "layers", ",", "we", "use", "average", "pooling", "." ]
[ 0, 0, 0, 1, 0, 0, 0, 0, 1, 0 ]
[ "In", "our", "experiments", ",", "we", "find", "that", "when", "the", "number", "of", "residual", "SDSC", "blocks", "is", "five", "with", "128", "channels", "input", "and", "the", "number", "of", "fully", "connected", "layers", "is", "just", "one", ",", ...
[ 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 ]
[ "SeqNet", "outputs", "the", "malicious", "possibility", "of", "a", "sample", ",", "and", "if", "the", "possibility", "is", "more", "than", "50", "\\", "%", ",", "the", "model", "will", "regard", "it", "as", "malware", "." ]
[ 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 ]
[ "Depending", "on", "the", "output", ",", "for", "the", "loss", "function", ",", "we", "use", "the", "cross-entropy", "function", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "In", "total", ",", "SeqNet", "has", "only", "about", "136", "K", "parameters", "which", "are", "nearly", "one", "-", "tenth", "of", "MalConv", ",", "which", "we", "will", "discuss", "in", "Section", "~", "." ]
[ 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 ]
[ "\\", "noindent", "Padding", "Problem", "Avoidance", "." ]
[ 0, 0, 1, 0, 0, 0 ]
[ "The", "padding", "problem", "could", "be", "solved", "because", "we", "only", "need", "to", "pad", "at", "both", "ends", "of", "the", "sequence", "before", "convolving", "." ]
[ 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Therefore", ",", "this", "operation", "is", "necessary", "to", "reduce", "the", "semantic", "loss", "caused", "by", "the", "linear", "interpolation", "algorithm", ",", "as", "shown", "in", "Figure", "~", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "Therefore", ",", "the", "model", "could", "recognize", "the", "correct", "location", "of", "the", "instruction", "according", "to", "the", "padding", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "The", "semantic", "loss", "could", "be", "reduced", "because", "we", "input", "the", "whole", "binary", "instead", "of", "just", "taking", "a", "snippet", "." ]
[ 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0 ]
[ "\\noindent", "Semantic", "Loss", "Reduction", "." ]
[ 0, 1, 0, 0, 0 ]
[ "What", "is", "more", ",", "compared", "to", "adding", "improper", "information", "to", "the", "image", ",", "the", "padding", "in", "the", "sequence", "will", "effectively", "mark", "the", "beginning", "and", "end", "of", "the", "corresponding", "program", ...
[ 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 ]
[ "In", "this", "case", ",", "embedded", "viruses", "could", "also", "be", "included", "because", "all", "instructions", "in", "a", "program", "are", "input", "into", "the", "model", "." ]
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "By", "the", "linear", "interpolation", "algorithm", ",", "we", "can", "compress", "the", "semantic", "instead", "of", "ignoring", "it", "." ]
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 ]
[ "This", "representation", "method", "also", "takes", "advantage", "of", "spatial", "locality", "in", "codes", "because", "the", "model", "will", "pay", "more", "attention", "to", "the", "instructions", "nearby", "instead", "of", "those", "far", "from", "each", ...
[ 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, 1, 0 ]
[ "Because", "of", "the", "continuity", "of", "the", "real", "number", "field", ",", "the", "float", "format", "could", "represent", "more", "information", "than", "the", "integer", "format", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Before", "scaling", "into", "the", "same", "length", ",", "we", "first", "normalize", "the", "whole", "sequences", "to", "make", "the", "value", "of", "elements", "between", "minus", "one", "and", "one", "stored", "in", "the", "float", "format", "." ]
[ 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, 1, 0 ]
[ "In", "this", "paper", ",", "we", "propose", "a", "lightweight", "malware", "detection", "model", "called", "SeqNet", "which", "could", "be", "trained", "at", "high", "speed", "with", "low", "memory", "required", "on", "the", "raw", "binaries", "." ]
[ 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, 1, 0 ]
[ "Malware", "is", "a", "serious", "cyber", "security", "threat", ",", "which", "could", "cause", "severe", "damage", "to", "individual", "and", "corporate", "systems", ",", "for", "example", ",", "the", "dramatic", "slowdown", "or", "breakdown", ",", "critical"...
[ 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 ]
[ "The", "enormous", "volume", "of", "new", "malware", "variants", "renders", "manual", "malware", "analysis", "inefficient", "and", "time", "-", "consuming", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
[ "We", "demonstrate", "the", "effectiveness", "of", "our", "methods", "and", "the", "low", "training", "cost", "requirement", "of", "SeqNet", "in", "our", "experiments", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0 ]
[ "Besides", ",", "we", "make", "our", "datasets", "and", "codes", "public", "to", "stimulate", "further", "academic", "research", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "However", ",", "existing", "deep", "learning", "detection", "models", "need", "manual", "feature", "engineering", "or", "require", "high", "computational", "overhead", "for", "long", "training", "processes", ",", "which", "might", "be", "laborious", "to", "select...
[ 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, 0, 0, 0, 0, 1, 0, 0 ]
[ "Malware", "continues", "to", "evolve", "rapidly", ",", "and", "more", "than", "450,000", "new", "samples", "are", "captured", "every", "day", ",", "which", "makes", "manual", "malware", "analysis", "impractical", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 ]
[ "By", "avoiding", "contextual", "confusion", "and", "reducing", "semantic", "loss", ",", "SeqNet", "maintains", "the", "detection", "accuracy", "when", "reducing", "the", "number", "of", "parameters", "to", "only", "136", "K", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "AVTest", "reports", "that", ",", "on", "average", ",", "over", "450,000", "new", "malicious", "programs", "and", "potentially", "unwanted", "applications", "are", "detected", "every", "day", "." ]
[ 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "We", "set", "the", "ratio", "of", "malicious", "and", "benign", "samples", "to", "about", "1", "to", "ensure", "the", "result", "is", "reliable", "." ]
[ 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "For", "example", ",", "if", "the", "dataset", "only", "has", "malicious", "samples", ",", "the", "model", "might", "detect", "all", "malware", "by", "recognizing", "\"", "4D", "5A", "\"", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "On", "the", "contrary", ",", "if", "we", "add", "enough", "benign", "samples", "into", "the", "dataset", ",", "the", "model", "could", "learn", "the", "real", "difference", "between", "malicious", "and", "benign", "samples", "." ]
[ 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, 1, 0 ]
[ "In", "our", "experiments", ",", "we", "measure", "SeqNet", "from", "two", "aspects", ",", "training", "cost", "and", "accuracy", "." ]
[ 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
[ "Each", "parameter", "occupies", "constant", "memory", "during", "the", "training", "and", "predicting", "process", "." ]
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
[ "In", "order", "to", "measure", "the", "computational", "overhead", "for", "model", "inference", "accurately", ",", "we", "calculate", "the", "floating", "point", "of", "operations", "~", "(", "Flops", ")", "on", "each", "model", "by", "inputting", "a", "ran...
[ 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, 1, 0 ]
[ "We", "also", "measure", "the", "speed", "of", "a", "model", "by", "recording", "the", "time", "needed", "for", "an", "epoch", ",", "including", "the", "training", "and", "validation", "processes", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "Besides", "accuracy", ",", "the", "measurements", "for", "performance", "we", "also", "use", "are", "precision", ",", "recall", ",", "and", "F1", "score", "." ]
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
[ "where", "TP", "means", "the", "malicious", "samples", "which", "the", "model", "predicts", "correctly", ",", "FP", "means", "the", "benign", "samples", "which", "the", "model", "predicts", "incorrectly", ",", "and", "FN", "means", "the", "malicious", "samples...
[ 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 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 ]