http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-114972839-A
Outgoing Links
| Predicate | Object |
|---|---|
| assignee | http://rdf.ncbi.nlm.nih.gov/pubchem/patentassignee/MD5_be2c55a9de09e08e3404c0a5a1ae4dc7 |
| classificationCPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-2431 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-214 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06N3-08 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06N3-047 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06N3-045 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-2415 |
| classificationIPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06N3-04 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06V10-82 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06V10-764 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06N3-08 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06V10-774 |
| filingDate | 2022-03-30-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| inventor | http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_3f8589d5eccad5638c9b817b65705ada http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_90bb1089bfb64ebb27ef84bd44be1a9d |
| publicationDate | 2022-08-30-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| publicationNumber | CN-114972839-A |
| titleOfInvention | A Generalized Continuous Classification Method Based on Online Contrastive Distillation Network |
| abstract | The invention discloses a generalized continuous classification method based on an online comparison distillation network. A classification model based on knowledge distillation is established; a buffer zone is established and a reservoir sampling method is used to update the buffer zone; S samples are randomly sampled from the buffer zone and respectively Input into the teacher and student models, and get the classification output and feature embedding corresponding to the two models; calculate and classify the output quality score according to the teacher model, adjust the weight of the knowledge distillation loss function of different samples and calculate the distillation loss Compare the feature embeddings between the two models, and calculate the distillation loss of the comparison relationship between the two models. Computing Student Model Self-Supervised Loss Comparing Learning Losses with Supervised Contrast Calculate Student Model Cross-Entropy Classification Loss The weighted summation of the above losses determines the maximum optimization objective to optimize the parameters of the student model. The parameters of the teacher model are updated by the parameters of the student model. The present invention has good classification accuracy for both new tasks and old tasks. |
| isCitedBy | http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-116502621-A http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-116502621-B http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-115457042-A |
| priorityDate | 2022-03-30-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| type | http://data.epo.org/linked-data/def/patent/Publication |
Incoming Links
Total number of triples: 29.