Advances, Systems and Applications
From: Defect knowledge graph construction and application in multi-cloud IoT
Symbol | Representation |
---|---|
Q & A | question and answer |
DB | database |
ES | elasticsearch |
OSS | Object Storage Service |
MOM | Message Oriented Middleware |
ISC | Internet Service Customer |
S6000 | a network equipment |
\(r(X_0,X_n)\) | the relationship between \(X_0\) and \(X_n\) |
G | the knowledge graph |
\(P_d\) | the representation of training data |
q | the question to be queried |
a | the inference result |
\(p_w\) | the probability of an answer conditioned defined by Reasoning Evaluator |
\(p_\theta\) | the prior probability of an answer conditioned defined by Rule Miner |
R | the representation of rule set |
\(\mathbf {E}\) | the maximum likelihood estimation |
\(w_{rule}\) | the weight of the rule |
\(score_w\) | the score of candidate answer in set A |
exp | the representation of exponential function in softmax |
log | the representation of log function |
H | the score of rules’ quality |
t | the representation of tail entity of knowledge triple |
A | the set of candidate answers |
MD | the representation of polynomial distribution |
N | the sampling number parameter in polynomial distribution |
GRU | the representation of Gate Recurrent Unit |
f | the representation of explicit function mapping used in GRU network |
K | the number of high-quality rules to output |
\(\max \mathcal {O}\) | the representation of maximizing the loglikelihood of the ruleset |