• Aucun résultat trouvé

Knowledge-based Transfer Learning Explanation

N/A
N/A
Protected

Academic year: 2021

Partager "Knowledge-based Transfer Learning Explanation"

Copied!
11
0
0

Texte intégral

Loading

Figure

Figure 1 displays some axiom examples of an LSO annotated by property-value pairs S := {dat : 01/01/2018, car : DL, ori : LAX, des : J F K}
Figure 2: Transfer Learning with Convolutional Neural Networks.
Table 1: Average Number of Root Entailments, Root Individuals and External Axioms per Learning Domain.
Figure 5 [Right] reports that (19.9%, 11.6%, 4.8%) of all the (2, 3, 4)-dimension entailment subsets have  signif-icant correlation analysis with FTI (i.e., ρ(X ) < 0.05), while (13.6%, 1.8%, 0.2%) are valid core contexts (i.e., ρ(X ) < 0.05 and kγ(X

Références

Documents relatifs

In our experiments, this version worked better in the Unsupervised and Transfer Learning Challenge, characterized by a non-discriminant linear classifier and very few labeled

(b) a simple yet efficient methodology to project the source em- beddings into a unique one, which is accurate both on the source and the target knowledge.. is introduced to

We remark that the same method used to prove (1.5) and (1.6) can also be used for computing the average root number for all the other families given in (1.3) (conditionally on

We systematically evaluate different teacher and student models, metric learning and knowledge transfer loss functions on the new asymmetric testing as well as the standard

We thus had one group of participants which transferred stimulus thus had one group of participants which transferred stimulus- - related knowledge, one group who transferred

This presents a representation learning model called SetE by modeling a predicate into a subspace in a semantic space where entities are vectors.. Within SetE, a type as unary

The aim is the transfer of proce- dural knowledge from a source into a target domain and we study the feasibility of using a process-oriented ontology as a means for the transfer..

The use of Kolmogorov complexity in machine learning is accepted by the community, but mainly in a stationary point of view (when the key concept does not vary); we proposed to