NEURAL PROCESSING UNIT FOR ATTENTION-BASED INFERENCE
There is provided a neural processing unit for calculating an attention matrix during machine learning inference. The neural processing unit is configured to calculate: a first score matrix based on differences between a query matrix and a key matrix; a second score matrix based on differences betwe...
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Patent |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext bestellen |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | |
|---|---|
| container_issue | |
| container_start_page | |
| container_title | |
| container_volume | |
| creator | BEU, Jesse Garrett O'CONNOR, Mark John GOPE, Dibakar DATTA, Shounak |
| description | There is provided a neural processing unit for calculating an attention matrix during machine learning inference. The neural processing unit is configured to calculate: a first score matrix based on differences between a query matrix and a key matrix; a second score matrix based on differences between the key matrix and a learned key matrix; a similarity matrix based on a combination of the first score matrix and second score matrix; and an attention matrix comprising applying a normalisation function to the similarity matrix. Also provided is an apparatus comprising at least one said neural processing unit and at least one memory, the memory configured to pass, on demand, a learned key matrix to the neural processing unit. Also provided is a computer program product having computer readable program code stored thereon which, when executed by said neural processing unit, causes the unit to perform said calculations. |
| format | Patent |
| fullrecord | <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_US2024028877A1</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>US2024028877A1</sourcerecordid><originalsourceid>FETCH-epo_espacenet_US2024028877A13</originalsourceid><addsrcrecordid>eNrjZDDxcw0NcvRRCAjyd3YNDvb0c1cI9fMMUXDzD1JwDAlx9Qvx9PfTdXIMdnVR8PRzcw1y9XN25WFgTUvMKU7lhdLcDMpuriHOHrqpBfnxqcUFicmpeakl8aHBRgZGJgZGFhbm5o6GxsSpAgB1zCh7</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>NEURAL PROCESSING UNIT FOR ATTENTION-BASED INFERENCE</title><source>esp@cenet</source><creator>BEU, Jesse Garrett ; O'CONNOR, Mark John ; GOPE, Dibakar ; DATTA, Shounak</creator><creatorcontrib>BEU, Jesse Garrett ; O'CONNOR, Mark John ; GOPE, Dibakar ; DATTA, Shounak</creatorcontrib><description>There is provided a neural processing unit for calculating an attention matrix during machine learning inference. The neural processing unit is configured to calculate: a first score matrix based on differences between a query matrix and a key matrix; a second score matrix based on differences between the key matrix and a learned key matrix; a similarity matrix based on a combination of the first score matrix and second score matrix; and an attention matrix comprising applying a normalisation function to the similarity matrix. Also provided is an apparatus comprising at least one said neural processing unit and at least one memory, the memory configured to pass, on demand, a learned key matrix to the neural processing unit. Also provided is a computer program product having computer readable program code stored thereon which, when executed by said neural processing unit, causes the unit to perform said calculations.</description><language>eng</language><subject>CALCULATING ; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS ; COMPUTING ; COUNTING ; PHYSICS</subject><creationdate>2024</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20240125&DB=EPODOC&CC=US&NR=2024028877A1$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,309,781,886,25566,76549</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20240125&DB=EPODOC&CC=US&NR=2024028877A1$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>BEU, Jesse Garrett</creatorcontrib><creatorcontrib>O'CONNOR, Mark John</creatorcontrib><creatorcontrib>GOPE, Dibakar</creatorcontrib><creatorcontrib>DATTA, Shounak</creatorcontrib><title>NEURAL PROCESSING UNIT FOR ATTENTION-BASED INFERENCE</title><description>There is provided a neural processing unit for calculating an attention matrix during machine learning inference. The neural processing unit is configured to calculate: a first score matrix based on differences between a query matrix and a key matrix; a second score matrix based on differences between the key matrix and a learned key matrix; a similarity matrix based on a combination of the first score matrix and second score matrix; and an attention matrix comprising applying a normalisation function to the similarity matrix. Also provided is an apparatus comprising at least one said neural processing unit and at least one memory, the memory configured to pass, on demand, a learned key matrix to the neural processing unit. Also provided is a computer program product having computer readable program code stored thereon which, when executed by said neural processing unit, causes the unit to perform said calculations.</description><subject>CALCULATING</subject><subject>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2024</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNrjZDDxcw0NcvRRCAjyd3YNDvb0c1cI9fMMUXDzD1JwDAlx9Qvx9PfTdXIMdnVR8PRzcw1y9XN25WFgTUvMKU7lhdLcDMpuriHOHrqpBfnxqcUFicmpeakl8aHBRgZGJgZGFhbm5o6GxsSpAgB1zCh7</recordid><startdate>20240125</startdate><enddate>20240125</enddate><creator>BEU, Jesse Garrett</creator><creator>O'CONNOR, Mark John</creator><creator>GOPE, Dibakar</creator><creator>DATTA, Shounak</creator><scope>EVB</scope></search><sort><creationdate>20240125</creationdate><title>NEURAL PROCESSING UNIT FOR ATTENTION-BASED INFERENCE</title><author>BEU, Jesse Garrett ; O'CONNOR, Mark John ; GOPE, Dibakar ; DATTA, Shounak</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_US2024028877A13</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>eng</language><creationdate>2024</creationdate><topic>CALCULATING</topic><topic>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>BEU, Jesse Garrett</creatorcontrib><creatorcontrib>O'CONNOR, Mark John</creatorcontrib><creatorcontrib>GOPE, Dibakar</creatorcontrib><creatorcontrib>DATTA, Shounak</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>BEU, Jesse Garrett</au><au>O'CONNOR, Mark John</au><au>GOPE, Dibakar</au><au>DATTA, Shounak</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>NEURAL PROCESSING UNIT FOR ATTENTION-BASED INFERENCE</title><date>2024-01-25</date><risdate>2024</risdate><abstract>There is provided a neural processing unit for calculating an attention matrix during machine learning inference. The neural processing unit is configured to calculate: a first score matrix based on differences between a query matrix and a key matrix; a second score matrix based on differences between the key matrix and a learned key matrix; a similarity matrix based on a combination of the first score matrix and second score matrix; and an attention matrix comprising applying a normalisation function to the similarity matrix. Also provided is an apparatus comprising at least one said neural processing unit and at least one memory, the memory configured to pass, on demand, a learned key matrix to the neural processing unit. Also provided is a computer program product having computer readable program code stored thereon which, when executed by said neural processing unit, causes the unit to perform said calculations.</abstract><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | |
| ispartof | |
| issn | |
| language | eng |
| recordid | cdi_epo_espacenet_US2024028877A1 |
| source | esp@cenet |
| subjects | CALCULATING COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS COMPUTING COUNTING PHYSICS |
| title | NEURAL PROCESSING UNIT FOR ATTENTION-BASED INFERENCE |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-18T08%3A27%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=BEU,%20Jesse%20Garrett&rft.date=2024-01-25&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3EUS2024028877A1%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |