tokens list | label_ids list |
|---|---|
[
"All",
"models",
"are",
"trained",
"for",
"70",
"epochs",
",",
"and",
"we",
"select",
"the",
"validation",
"results",
"of",
"the",
"last",
"30",
"epochs",
"to",
"get",
"the",
"average",
"accuracy",
"and",
"other",
"metrics",
"."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"We",
"set",
"the",
"batch",
"size",
"to",
"32",
"and",
"choose",
"Adam",
"~",
"as",
"the",
"optimizer",
"of",
"all",
"models",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0
] |
[
"In",
"order",
"to",
"ensure",
"fairness",
"in",
"training",
",",
"we",
"apply",
"the",
"cross-entropy",
"loss",
"to",
"all",
"models",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0
] |
[
"Also",
",",
"we",
"find",
"that",
"the",
"PE",
"headers",
"often",
"affect",
"SeqNet",
"greatly",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
1,
0,
0
] |
[
"This",
"might",
"imply",
"that",
"the",
"PE",
"header",
"contains",
"malicious",
"information",
"in",
"malware",
"."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
1,
0
] |
[
"Although",
"SeqNet",
"performs",
"well",
",",
"there",
"are",
"still",
"several",
"limitations",
"in",
"our",
"works",
"."
] | [
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"Remained",
"Semantic",
"Loss",
"."
] | [
0,
0,
1,
0
] |
[
"Although",
"we",
"effectively",
"reduce",
"the",
"semantic",
"loss",
",",
"the",
"input",
"of",
"SeqNet",
"still",
"cannot",
"contain",
"all",
"semantics",
"."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0
] |
[
"If",
"the",
"sequence",
"is",
"too",
"long",
",",
"the",
"sequence",
"will",
"be",
"compressed",
"during",
"the",
"interpolation",
"process",
",",
"and",
"the",
"compressed",
"sequence",
"is",
"not",
"able",
"to",
"represent",
"all",
"original",
"information"... | [
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
] |
[
"Also",
",",
"if",
"the",
"sequence",
"is",
"too",
"short",
",",
"the",
"sequence",
"will",
"be",
"expanded",
",",
"which",
"might",
"confuse",
"SeqNet",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"The",
"architecture",
"of",
"SeqNet",
"determines",
"that",
"the",
"input",
"must",
"have",
"the",
"same",
"size",
",",
"which",
"is",
"a",
"limitation",
"of",
"SeqNet",
"."
] | [
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"Lack",
"of",
"Benign",
"Samples",
"."
] | [
0,
0,
0,
1,
0
] |
[
"The",
"main",
"difficulty",
"we",
"face",
"is",
"the",
"lack",
"of",
"benign",
"samples",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"We",
"can",
"get",
"plenty",
"of",
"malware",
"collection",
"websites",
",",
"but",
"authoritative",
"benign",
"sample",
"providers",
"are",
"hard",
"to",
"find",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0
] |
[
"In",
"order",
"to",
"sample",
"evenly",
",",
"it",
"is",
"improper",
"to",
"expand",
"the",
"training",
"dataset",
"only",
"by",
"adding",
"malicious",
"samples",
",",
"which",
"might",
"reduce",
"the",
"performance",
"of",
"SeqNet",
"and",
"make",
"the",
... | [
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,
0,
0,
0,
0,
0,
0
] |
[
"As",
"a",
"result",
",",
"it",
"is",
"difficult",
"to",
"train",
"neural",
"networks",
"on",
"a",
"much",
"larger",
"dataset",
"with",
"enough",
"benign",
"samples",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"Besides",
"the",
"lack",
"of",
"benign",
"samples",
",",
"the",
"quality",
"of",
"labels",
"might",
"be",
"a",
"potential",
"problem",
"."
] | [
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"Due",
"to",
"the",
"few",
"authoritative",
"providers",
",",
"we",
"cannot",
"guarantee",
"that",
"all",
"the",
"benign",
"samples",
"in",
"our",
"training",
"and",
"validation",
"datasets",
"are",
"labeled",
"correctly",
"."
] | [
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
] |
[
"All",
"malicious",
"samples",
"in",
"our",
"datasets",
"are",
"collected",
"from",
"VirusShare",
"without",
"manual",
"reconfirmation",
"."
] | [
0,
0,
1,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0
] |
[
"Several",
"papers",
"have",
"inspected",
"the",
"quality",
"of",
"malware",
"labels",
"and",
"found",
"it",
"probably",
"cannot",
"reach",
"what",
"we",
"expect",
"~",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"Although",
"these",
"limitations",
"might",
"have",
"a",
"few",
"impacts",
"on",
"the",
"training",
"process",
"of",
"SeqNet",
",",
"we",
"assume",
"that",
"the",
"several",
"incorrect",
"labeled",
"samples",
"could",
"not",
"significantly",
"affect",
"the",
... | [
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
] |
[
"Another",
"reason",
"to",
"use",
"sequence",
"characterization",
"instead",
"of",
"binary",
"-",
"image",
"converting",
"is",
"that",
"executable",
"files",
"have",
"a",
"more",
"pronounced",
"before",
"-",
"and",
"-",
"after",
"correlation",
"than",
"the",
"... | [
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,
0,
0,
1,
0
] |
[
"Based",
"on",
"sequence",
"input",
"and",
"depthwise",
"separable",
"convolution",
"~",
",",
"we",
"propose",
"a",
"method",
"called",
"Sequence",
"Depthwise",
"Separable",
"Convolution",
"~",
"(",
"SDSC",
")",
"which",
"needs",
"fewer",
"parameters",
"and",
... | [
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"In",
"the",
"following",
",",
"we",
"analyze",
"the",
"computation",
"reduction",
"compared",
"to",
"DSC",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"In",
"SDSC",
",",
"We",
"use",
"3",
"\\times",
"1",
"kernels",
"to",
"replace",
"the",
"2D",
"depthwise",
"convolution",
"kernels",
"in",
"DSC",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
1,
0,
1,
0
] |
[
"The",
"convolution",
"kernels",
"of",
"SeqNet",
"only",
"need",
"to",
"extract",
"features",
"on",
"a",
"sequence",
"that",
"has",
"just",
"one",
"dimension",
"."
] | [
0,
0,
1,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"Sequence",
"Depthwise",
"Separable",
"Convolution"
] | [
0,
0,
0,
1
] |
[
"Compared",
"to",
"processing",
"one",
"dimension",
"input",
",",
"extracting",
"features",
"on",
"two",
"dimensions",
"requires",
"larger",
"convolution",
"kernels",
"and",
"more",
"calculations",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0
] |
[
"For",
"example",
",",
"as",
"shown",
"in",
"Figure",
"~",
",",
"a",
"3",
"\\times",
"3",
"kernel",
"used",
"in",
"images",
"needs",
"at",
"least",
"ten",
"parameters",
"~",
"(",
"including",
"bias",
")",
"while",
"a",
"3",
"\\times",
"1",
"kernel",
... | [
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,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"By",
"using",
"the",
"SDSC",
"layers",
",",
"SeqNet",
"has",
"a",
"much",
"smaller",
"size",
"than",
"existing",
"models",
"."
] | [
0,
0,
0,
0,
1,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"The",
"sequence",
"input",
"format",
"not",
"only",
"addresses",
"the",
"problem",
"of",
"semantic",
"loss",
"and",
"contextual",
"confusion",
"but",
"also",
"compresses",
"the",
"scale",
"of",
"SeqNet",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"The",
"robustness",
"of",
"SeqNet",
"still",
"needs",
"further",
"research",
"."
] | [
0,
0,
0,
1,
0,
0,
0,
0,
0
] |
[
"Related",
"Work",
"%",
"\\noindent",
"Feature",
"-",
"based",
"Models",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"In",
"the",
"early",
"works",
",",
"deep",
"learning",
"models",
"are",
"trained",
"from",
"carefully",
"crafted",
"malware",
"features",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"Yuan",
"et",
"al.",
"used",
"deep",
"learning",
"algorithms",
"to",
"detect",
"Android",
"malware",
"based",
"on",
"202",
"manual",
"features",
"extracted",
"through",
"static",
"and",
"dynamic",
"analysis",
"~."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"Saxe",
"et",
"al.",
"designed",
"a",
"network",
"with",
"four",
"layers",
"and",
"trained",
"the",
"model",
"with",
"the",
"static",
"features",
"extracted",
"from",
"PE",
"files",
"~."
] | [
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"Huang",
"et",
"al.",
"used",
"4.5",
"million",
"samples",
"as",
"the",
"training",
"dataset",
"and",
"used",
"the",
"random",
"projection",
"to",
"reduce",
"feature",
"space",
"from",
"50000",
"dimensions",
"to",
"4000",
"dimensions",
"~."
] | [
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
] |
[
"Zhang",
"et",
"al.",
"combined",
"dynamic",
"analysis",
"techniques",
"with",
"deep",
"learning",
"algorithms",
"and",
"used",
"the",
"API",
"calls",
"with",
"arguments",
"to",
"train",
"their",
"network",
"~."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
1,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
1,
0
] |
[
"Xu",
"et",
"al.",
"devised",
"a",
"two",
"-",
"stage",
"inference",
"framework",
"to",
"detect",
"Android",
"malware",
"on",
"the",
"extracted",
"information",
"~."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
1,
0,
0,
0,
0,
0
] |
[
"Li",
"et",
"al.",
"leveraged",
"Autoencoder",
"~",
"(",
"AE",
")",
"to",
"find",
"the",
"mutation",
"of",
"malicious",
"features",
"and",
"got",
"great",
"performance",
"~."
] | [
0,
0,
0,
0,
1,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0
] |
[
"Han",
"et",
"al.",
"extracted",
"installation",
"graphs",
"from",
"viruses",
"and",
"applied",
"Graph",
"Neural",
"Network",
"~",
"(",
"GNN",
")",
"to",
"malware",
"detection",
"~.",
"\\",
"noindent",
"Binary",
"-",
"based",
"Models",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
1,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"Nowadays",
",",
"automatic",
"feature",
"extraction",
"is",
"one",
"of",
"the",
"trends",
"of",
"neural",
"networks",
"with",
"less",
"human",
"intervention",
"and",
"better",
"performance",
"than",
"traditional",
"feature",
"engineering",
"."
] | [
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"Raff",
"et",
"al.",
"devised",
"an",
"architecture",
"called",
"MalConv",
"which",
"could",
"learn",
"directly",
"from",
"the",
"raw",
"PE",
"binary",
"samples",
"without",
"manual",
"feature",
"selection",
"~."
] | [
0,
0,
0,
0,
0,
1,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0
] |
[
"Krc",
"{",
"\\'{a}",
"}",
"l",
"et",
"al.",
"designed",
"a",
"simple",
"CNN",
"which",
"learns",
"from",
"PE",
"raw",
"byte",
"sequence",
"without",
"domain",
"-",
"specific",
"feature",
"selection",
",",
"and",
"this",
"work",
"achieved",
"a",
"high",
... | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"We",
"have",
"created",
"our",
"own",
"OpenAI",
"-",
"Gym",
"-",
"compatible",
"environment",
",",
"where",
"agents",
"with",
"both",
"episodic",
"and",
"semantic",
"memory",
"systems",
"can",
"be",
"tested",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
1,
0,
0,
1,
0,
0,
0,
1,
0,
0,
0,
0
] |
[
"We",
"showed",
"that",
"when",
"a",
"machine",
"is",
"explicitly",
"given",
"both",
"semantic",
"and",
"episodic",
"memory",
"systems",
",",
"it",
"outperforms",
"the",
"ones",
"that",
"only",
"have",
"one",
"of",
"the",
"two",
"memory",
"systems",
"."
] | [
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
] |
[
"We",
"also",
"showed",
"that",
"when",
"an",
"agent",
"is",
"pretrained",
"with",
"commonsense",
"knowledge",
",",
"it",
"outperforms",
"the",
"one",
"that",
"is",
"not",
"pretrained",
"."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"In",
"the",
"future",
",",
"we",
"want",
"to",
"see",
"if",
"reinforcement",
"learning",
"agents",
"can",
"perform",
"as",
"good",
"as",
"the",
"handcrafted",
"ones",
",",
"to",
"see",
"if",
"such",
"data",
"-",
"driven",
"agents",
"can",
"lead",
"to",
... | [
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
] |
[
"We",
"also",
"hope",
"to",
"make",
"the",
"Room",
"environment",
"even",
"closer",
"to",
"the",
"real",
"human",
"environment",
"(",
"e.g.",
",",
"adding",
"images",
",",
"voices",
",",
"more",
"entities",
"and",
"relations",
",",
"etc.",
")",
"."
] | [
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
] |
[
"As",
"for",
"the",
"collaboration",
",",
"the",
"current",
"setup",
"is",
"limited",
"to",
"only",
"agents",
"working",
"together",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0
] |
[
"It",
"would",
"be",
"especially",
"interesting",
"to",
"encourage",
"agents",
"to",
"ask",
"humans",
"questions",
",",
"when",
"it",
"is",
"not",
"sure",
"how",
"to",
"answer",
"them",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"Collaboration",
"will",
"not",
"always",
"be",
"straightforward",
",",
"especially",
"if",
"there",
"are",
"conflicts",
"and",
"different",
"degrees",
"of",
"trust",
"among",
"humans",
"and",
"agents",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"The",
"reason",
"why",
"the",
"observations",
"and",
"questions",
"are",
"given",
"as",
"RDF",
"-",
"triple",
"-",
"like",
"format",
"is",
"two",
"folds",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0
] |
[
"Second",
"is",
"that",
"we",
"can",
"use",
"existing",
"knowledge",
"graphs",
",",
"such",
"as",
"ConceptNet",
"~",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
1,
0,
0
] |
[
"The",
"commonsense",
"knowledge",
"we",
"use",
"is",
"from",
"ConceptNet",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"All",
"of",
"the",
"four",
"probabilities",
"account",
"for",
"the",
"Bernoulli",
"distributions",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"Episodic",
"and",
"Semantic",
"Memory",
"Systems"
] | [
0,
0,
0,
0,
1
] |
[
"This",
"can",
"be",
"done",
"by",
"having",
"memory",
"systems",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"In",
"our",
"work",
",",
"we",
"model",
"our",
"agent",
"to",
"have",
"a",
"human",
"-",
"like",
"explicit",
"memory",
"system",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"This",
"means",
"that",
"it",
"has",
"both",
"episodic",
"and",
"semantic",
"memory",
"systems",
"."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
1,
0
] |
[
"If",
"the",
"agent",
"has",
"an",
"episodic",
"memory",
"<",
"Karen",
"'s",
"cat",
",",
"At",
"Location",
",",
"Karen",
"'s",
"office",
",",
"21",
">",
",",
"which",
"it",
"has",
"seen",
"two",
"time",
"steps",
"ago",
",",
"this",
"episodic",
"memor... | [
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,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"The",
"Room",
"environment",
"does",
"not",
"restrict",
"the",
"number",
"of",
"agents",
"."
] | [
0,
0,
1,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"If",
"agent_{1}",
"has",
"an",
"episodic",
"memory",
"<",
"Karen",
"'s",
"cat",
",",
"AtLocation",
",",
"Karen",
"'s",
"office",
",",
"21",
">",
"and",
"agent_{2}",
"has",
"an",
"episodic",
"memory",
"<",
"Karen",
"'s",
"cat",
",",
"AtLocation",
",",
... | [
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,
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... |
[
"Or",
",",
"it",
"is",
"also",
"possible",
"for",
"them",
"to",
"use",
"commonsense",
"knowledge",
"that",
"â",
"Cats",
"are",
"often",
"found",
"on",
"the",
"lap",
"of",
"their",
"owners",
".",
"â",
",",
"if",
"the",
"agents",
"have",
"encoded",
"... | [
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
] |
[
"Data",
"Collection",
"and",
"the",
"Environment",
"Hyperparameters"
] | [
0,
0,
0,
0,
0,
1
] |
[
"In",
"order",
"to",
"simplify",
"the",
"experiment",
"setup",
",",
"we",
"decided",
"to",
"use",
"a",
"subset",
"of",
"the",
"ConceptNet",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"This",
"means",
"that",
"the",
"environment",
"terminates",
"after",
"an",
"agent",
"has",
"taken",
"1,000",
"steps",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0
] |
[
"Not",
"every",
"observation",
"has",
"to",
"be",
"saved",
"in",
"the",
"episodic",
"memory",
"system",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"If",
"the",
"agent",
"has",
"an",
"episodic",
"memory",
"of",
"this",
"event",
",",
"then",
"it",
"can",
"use",
"it",
"."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"This",
"commonsense",
"knowledge",
"forms",
"the",
"semantic",
"memory",
"of",
"the",
"agent",
",",
"since",
"semantic",
"memory",
"has",
"to",
"do",
"with",
"the",
"general",
"knowledge",
"of",
"the",
"world",
"."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
1,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"Similar",
"to",
"episodic",
"memories",
",",
"semantic",
"memories",
"are",
"also",
"quadruples",
"."
] | [
0,
0,
0,
1,
0,
0,
1,
0,
0,
0,
0
] |
[
"However",
",",
"the",
"last",
"element",
"of",
"a",
"semantic",
"memory",
"is",
"not",
"a",
"timestamp",
",",
"but",
"the",
"strength",
"of",
"the",
"semantic",
"memory",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"For",
"example",
",",
"<",
"laptop",
",",
"AtLocation",
",",
"desk",
",",
"10",
">",
"is",
"a",
"stronger",
"semantic",
"memory",
"than",
"<",
"laptop",
",",
"AtLocation",
",",
"garage",
",",
"5",
">",
",",
"since",
"the",
"agent",
"has",
"seen",
"l... | [
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,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"Notice",
"that",
"semantic",
"memories",
"do",
"not",
"include",
"the",
"names",
"of",
"people",
",",
"since",
"this",
"type",
"of",
"memory",
"is",
"not",
"person",
"-",
"specific",
"."
] | [
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"Both",
"episodic",
"and",
"semantic",
"memory",
"systems",
",",
"\\bm{M}_{E}",
"and",
"\\bm{M}_{S}",
",",
"respectively",
",",
"are",
"bounded",
"in",
"size",
"."
] | [
0,
1,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"It",
"should",
"also",
"learn",
"that",
"some",
"of",
"its",
"episodic",
"memories",
"can",
"be",
"summarized",
"into",
"one",
"semantic",
"memory",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"For",
"example",
",",
"if",
"it",
"always",
"sees",
"humans",
"placing",
"their",
"laptops",
"on",
"their",
"desks",
",",
"then",
"perhaps",
"one",
"semantic",
"memory",
"<",
"laptop",
",",
"AtLocation",
",",
"desk",
">",
"is",
"enough",
"to",
"answer",
... | [
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
] |
[
"For",
"a",
"fair",
"comparison",
",",
"memory",
"capacity",
"was",
"kept",
"the",
"same",
"(",
"i.e.",
",",
"as",
"for",
"the",
"two",
"agent",
"setup",
",",
"each",
"agent",
"can",
"store",
"16",
"memories",
",",
"while",
"the",
"agent",
"in",
"the",... | [
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,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"It",
"shows",
"that",
"two",
"agents",
"working",
"together",
"were",
"able",
"to",
"answer",
"more",
"questions",
"than",
"one",
"agent",
"working",
"alone",
"."
] | [
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0
] |
[
"This",
"is",
"due",
"to",
"the",
"fact",
"that",
"the",
"two",
"agents",
"were",
"exploring",
"the",
"room",
"in",
"different",
"directions",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"This",
"led",
"them",
"to",
"cover",
"more",
"area",
"than",
"one",
"agent",
"acting",
"alone",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0
] |
[
"After",
"studying",
"related",
"literature",
",",
"we",
"observed",
"that",
"papers",
"that",
"are",
"theoretically",
"similar",
"to",
"ours",
"are",
"mostly",
"cognitive",
"science",
"papers",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0
] |
[
"ACT",
"-",
"R",
"~",
"and",
"Soar",
"~",
"put",
"a",
"big",
"emphasis",
"on",
"theories",
",",
"but",
"they",
"lack",
"of",
"computational",
"experiments",
",",
"which",
"makes",
"it",
"hard",
"to",
"compare",
"."
] | [
1,
1,
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
] |
[
"There",
"was",
"a",
"work",
"~",
"that",
"studied",
"how",
"episodic",
"and",
"semantic",
"memory",
"systems",
"play",
"a",
"role",
"in",
"recall",
"of",
"objects",
",",
"but",
"their",
"experiments",
"were",
"human",
"-",
"based",
"empirical",
"results",
... | [
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,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"Second",
"is",
"that",
"although",
"some",
"computer",
"science",
"based",
"papers",
"do",
"computational",
"experiments",
"that",
"are",
"a",
"bit",
"similar",
"to",
"ours",
",",
"they",
"often",
"do",
"not",
"study",
"episodic",
"and",
"semantic",
"memory",... | [
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,
0,
0,
1,
0,
0
] |
[
"For",
"example",
",",
"Episodic",
"Memory",
"Reader",
"~",
"also",
"learns",
"what",
"to",
"forget",
"in",
"their",
"memory",
"system",
",",
"and",
"they",
"also",
"use",
"question",
"-",
"answering",
"to",
"evaluate",
"their",
"method",
"."
] | [
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,
0
] |
[
"However",
",",
"this",
"work",
"only",
"focuses",
"on",
"episodic",
"memory",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
[
"Also",
",",
"their",
"memory",
"system",
"is",
"not",
"composed",
"of",
"RDF",
"-",
"like",
"data",
",",
"but",
"rather",
"numeric",
"embeddings",
",",
"which",
"are",
"hard",
"to",
"interpret",
"what",
"they",
"have",
"captured",
"."
] | [
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
] |
[
"For",
"a",
"fair",
"comparison",
",",
"every",
"agent",
"has",
"the",
"same",
"total",
"memory",
"capacity",
"."
] | [
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
1,
0
] |
[
"As",
"for",
"the",
"Handcrafted",
"3",
"agent",
",",
"the",
"episodic",
"and",
"semantic",
"memory",
"systems",
"have",
"the",
"same",
"capacity",
",",
"since",
"this",
"agent",
"does",
"not",
"know",
"which",
"one",
"is",
"more",
"important",
"a",
"prior... | [
0,
0,
0,
0,
0,
1,
0,
0,
1,
0,
0,
0,
1,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"As",
"for",
"the",
"Handcrafted",
"4",
"agent",
",",
"if",
"there",
"is",
"space",
"left",
"in",
"the",
"semantic",
"memory",
"system",
"after",
"filling",
"it",
"up",
",",
"it",
"will",
"give",
"the",
"rest",
"of",
"the",
"space",
"to",
"the",
"episo... | [
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,
0,
0,
0,
0,
1,
0
] |
[
"In",
"order",
"to",
"show",
"the",
"validity",
"of",
"our",
"handcrafted",
"agents",
",",
"we",
"compare",
"them",
"with",
"the",
"agents",
"that",
"forget",
"and",
"answer",
"uniform",
"-",
"randomly",
"."
] | [
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
] |
[
"The",
"multiple",
"agent",
"policies",
"work",
"in",
"the",
"same",
"manner",
"as",
"the",
"single",
"agent",
"policies",
",",
"except",
"that",
"they",
"can",
"use",
"their",
"combined",
"memory",
"systems",
"to",
"answer",
"questions",
"."
] | [
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0
] |
[
"It",
"shows",
"that",
"when",
"the",
"memory",
"capacity",
"is",
"low",
",",
"having",
"only",
"episodic",
"memory",
"system",
"is",
"better",
"than",
"the",
"others",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0
] |
[
"As",
"the",
"memory",
"capacity",
"increases",
",",
"however",
",",
"it",
"shows",
"that",
"having",
"a",
"semantic",
"memory",
"system",
"helps",
",",
"as",
"it",
"learns",
"to",
"generalize",
"the",
"world",
"."
] | [
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"It",
"is",
"especially",
"interesting",
"to",
"see",
"the",
"Handcrafted",
"4",
"agent",
",",
"which",
"has",
"an",
"episodic",
"system",
"and",
"a",
"pretrained",
"semantic",
"system",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
1,
0
] |
[
"Since",
"it",
"already",
"knows",
"the",
"general",
"world",
"knowledge",
",",
"it",
"could",
"focus",
"more",
"on",
"the",
"episodic",
"part",
",",
"leading",
"to",
"better",
"performance",
"."
] | [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0
] |
[
"Handcrafted",
"1",
":",
"Only",
"episodic",
",",
"forget",
"the",
"oldest",
"and",
"answer",
"the",
"latest",
"."
] | [
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0
] |
[
"This",
"agent",
"only",
"has",
"an",
"episodic",
"memory",
"system",
"."
] | [
0,
1,
0,
0,
0,
0,
0,
1,
0
] |
[
"When",
"the",
"episodic",
"memory",
"system",
"is",
"full",
",",
"it",
"will",
"forget",
"the",
"oldest",
"episodic",
"memory",
"."
] | [
0,
0,
0,
0,
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
0
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.