<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="it">
	<id>https://wiki.mindmaker.it/index.php?action=history&amp;feed=atom&amp;title=SPLADE</id>
	<title>SPLADE - Cronologia</title>
	<link rel="self" type="application/atom+xml" href="https://wiki.mindmaker.it/index.php?action=history&amp;feed=atom&amp;title=SPLADE"/>
	<link rel="alternate" type="text/html" href="https://wiki.mindmaker.it/index.php?title=SPLADE&amp;action=history"/>
	<updated>2026-05-01T11:43:28Z</updated>
	<subtitle>Cronologia della pagina su questo sito</subtitle>
	<generator>MediaWiki 1.41.0</generator>
	<entry>
		<id>https://wiki.mindmaker.it/index.php?title=SPLADE&amp;diff=2880&amp;oldid=prev</id>
		<title>Mindmakerbot il 16:26, 17 ago 2024</title>
		<link rel="alternate" type="text/html" href="https://wiki.mindmaker.it/index.php?title=SPLADE&amp;diff=2880&amp;oldid=prev"/>
		<updated>2024-08-17T16:26:53Z</updated>

		<summary type="html">&lt;p&gt;&lt;/p&gt;
&lt;table style=&quot;background-color: #fff; color: #202122;&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;it&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Versione meno recente&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Versione delle 16:26, 17 ago 2024&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l31&quot;&gt;Riga 31:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Riga 31:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |title=Your page title&lt;/div&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |title=Your page title&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |title_mode=append&lt;/div&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |title_mode=append&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;−&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |keywords=&lt;del style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;&quot;&lt;/del&gt;SPLADE, Sparse Lexical and Expansion Model, modello linguistico, embedding, rappresentazione semantica, BERT, Word-Pieces, espansione dei termini, apprendimento automatico, Pinecone, Masked-Language-Modeling, matrice di embedding, HuggingFace, Naver Labs&lt;del style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;&quot;&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |keywords=SPLADE, Sparse Lexical and Expansion Model, modello linguistico, embedding, rappresentazione semantica, BERT, Word-Pieces, espansione dei termini, apprendimento automatico, Pinecone, Masked-Language-Modeling, matrice di embedding, HuggingFace, Naver Labs&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;−&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |description=&lt;del style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;&quot;&lt;/del&gt;Questo articolo descrive SPLADE (Sparse Lexical and Expansion Model), un modello che combina la sparsità degli approcci tradizionali come TF-IDF con la ricchezza semantica dei modelli densi come BERT. SPLADE apprende a espandere i termini e a generarne di nuovi, migliorando la rappresentazione del testo e le attività di recupero delle informazioni.&lt;del style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;&quot;&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |description=Questo articolo descrive SPLADE (Sparse Lexical and Expansion Model), un modello che combina la sparsità degli approcci tradizionali come TF-IDF con la ricchezza semantica dei modelli densi come BERT. SPLADE apprende a espandere i termini e a generarne di nuovi, migliorando la rappresentazione del testo e le attività di recupero delle informazioni.&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |image=Temexpansion.png&lt;/div&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             |image=Temexpansion.png&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             }}&lt;/div&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;             }}&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Mindmakerbot</name></author>
	</entry>
	<entry>
		<id>https://wiki.mindmaker.it/index.php?title=SPLADE&amp;diff=2752&amp;oldid=prev</id>
		<title>Mindmakerbot il 13:36, 17 ago 2024</title>
		<link rel="alternate" type="text/html" href="https://wiki.mindmaker.it/index.php?title=SPLADE&amp;diff=2752&amp;oldid=prev"/>
		<updated>2024-08-17T13:36:10Z</updated>

		<summary type="html">&lt;p&gt;&lt;/p&gt;
&lt;table style=&quot;background-color: #fff; color: #202122;&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;it&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Versione meno recente&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Versione delle 13:36, 17 ago 2024&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l27&quot;&gt;Riga 27:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Riga 27:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;br&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;br&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;[[Categoria:Modello]]&lt;/div&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot;&gt;&lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;[[Categoria:Modello]]&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-side-deleted&quot;&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-side-deleted&quot;&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;{{#seo:&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-side-deleted&quot;&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;            |title=Your page title&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-side-deleted&quot;&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;            |title_mode=append&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-side-deleted&quot;&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;            |keywords=&quot;SPLADE, Sparse Lexical and Expansion Model, modello linguistico, embedding, rappresentazione semantica, BERT, Word-Pieces, espansione dei termini, apprendimento automatico, Pinecone, Masked-Language-Modeling, matrice di embedding, HuggingFace, Naver Labs&quot;&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-side-deleted&quot;&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;            |description=&quot;Questo articolo descrive SPLADE (Sparse Lexical and Expansion Model), un modello che combina la sparsità degli approcci tradizionali come TF-IDF con la ricchezza semantica dei modelli densi come BERT. SPLADE apprende a espandere i termini e a generarne di nuovi, migliorando la rappresentazione del testo e le attività di recupero delle informazioni.&quot;&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-side-deleted&quot;&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;            |image=Temexpansion.png&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-side-deleted&quot;&gt;&lt;/td&gt;&lt;td class=&quot;diff-marker&quot; data-marker=&quot;+&quot;&gt;&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;ins style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;            }}&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Mindmakerbot</name></author>
	</entry>
	<entry>
		<id>https://wiki.mindmaker.it/index.php?title=SPLADE&amp;diff=2421&amp;oldid=prev</id>
		<title>Alesaccoia: Creata pagina con &quot;Acronimo di  &#039;&#039;&#039;Sp&#039;&#039;&#039;arse &#039;&#039;&#039;L&#039;&#039;&#039;exical &#039;&#039;&#039;a&#039;&#039;&#039;n&#039;&#039;&#039;d&#039;&#039;&#039; &#039;&#039;&#039;E&#039;&#039;&#039;xpansion model: modello che produce embedding &#039;&#039;&#039;sparsi&#039;&#039;&#039; come gli approcci Bag-Of-Words, TF-IDF o BM25, ma arricchiti da una rappresentazione semantica come nei modelli &#039;&#039;&#039;densi&#039;&#039;&#039;.  L&#039;idea di base è che un modello di linguaggio pre-addestrato come BERT possa identificare dei collegamenti fra parole/sub-words (chiamati Word-Pieces) e utilizzare quella conoscenza per aumentare l&#039;embedding sp...&quot;</title>
		<link rel="alternate" type="text/html" href="https://wiki.mindmaker.it/index.php?title=SPLADE&amp;diff=2421&amp;oldid=prev"/>
		<updated>2024-06-20T15:57:56Z</updated>

		<summary type="html">&lt;p&gt;Creata pagina con &amp;quot;Acronimo di  &amp;#039;&amp;#039;&amp;#039;Sp&amp;#039;&amp;#039;&amp;#039;arse &amp;#039;&amp;#039;&amp;#039;L&amp;#039;&amp;#039;&amp;#039;exical &amp;#039;&amp;#039;&amp;#039;a&amp;#039;&amp;#039;&amp;#039;n&amp;#039;&amp;#039;&amp;#039;d&amp;#039;&amp;#039;&amp;#039; &amp;#039;&amp;#039;&amp;#039;E&amp;#039;&amp;#039;&amp;#039;xpansion model: modello che produce &lt;a href=&quot;/index.php?title=Embedding&amp;amp;action=edit&amp;amp;redlink=1&quot; class=&quot;new&quot; title=&quot;Embedding (la pagina non esiste)&quot;&gt;embedding&lt;/a&gt; &amp;#039;&amp;#039;&amp;#039;sparsi&amp;#039;&amp;#039;&amp;#039; come gli approcci &lt;a href=&quot;/index.php?title=Bag-Of-Words&amp;amp;action=edit&amp;amp;redlink=1&quot; class=&quot;new&quot; title=&quot;Bag-Of-Words (la pagina non esiste)&quot;&gt;Bag-Of-Words&lt;/a&gt;, TF-IDF o BM25, ma arricchiti da una rappresentazione semantica come nei modelli &amp;#039;&amp;#039;&amp;#039;densi&amp;#039;&amp;#039;&amp;#039;.  L&amp;#039;idea di base è che un modello di linguaggio pre-addestrato come &lt;a href=&quot;/index.php/BERT&quot; title=&quot;BERT&quot;&gt;BERT&lt;/a&gt; possa identificare dei collegamenti fra parole/sub-words (chiamati &lt;a href=&quot;/index.php?title=Word-Pieces&amp;amp;action=edit&amp;amp;redlink=1&quot; class=&quot;new&quot; title=&quot;Word-Pieces (la pagina non esiste)&quot;&gt;Word-Pieces&lt;/a&gt;) e utilizzare quella conoscenza per aumentare l&amp;#039;embedding sp...&amp;quot;&lt;/p&gt;
&lt;p&gt;&lt;b&gt;Nuova pagina&lt;/b&gt;&lt;/p&gt;&lt;div&gt;Acronimo di  &amp;#039;&amp;#039;&amp;#039;Sp&amp;#039;&amp;#039;&amp;#039;arse &amp;#039;&amp;#039;&amp;#039;L&amp;#039;&amp;#039;&amp;#039;exical &amp;#039;&amp;#039;&amp;#039;a&amp;#039;&amp;#039;&amp;#039;n&amp;#039;&amp;#039;&amp;#039;d&amp;#039;&amp;#039;&amp;#039; &amp;#039;&amp;#039;&amp;#039;E&amp;#039;&amp;#039;&amp;#039;xpansion model: modello che produce [[embedding]] &amp;#039;&amp;#039;&amp;#039;sparsi&amp;#039;&amp;#039;&amp;#039; come gli approcci [[Bag-Of-Words]], TF-IDF o BM25, ma arricchiti da una rappresentazione semantica come nei modelli &amp;#039;&amp;#039;&amp;#039;densi&amp;#039;&amp;#039;&amp;#039;.&lt;br /&gt;
&lt;br /&gt;
L&amp;#039;idea di base è che un modello di linguaggio pre-addestrato come [[BERT]] possa identificare dei collegamenti fra parole/sub-words (chiamati [[Word-Pieces]]) e utilizzare quella conoscenza per aumentare l&amp;#039;embedding sparso.&lt;br /&gt;
[[File:Temexpansion.png|miniatura|Term Expansion (da Pinecone)]]&lt;br /&gt;
In tal modo si pesa il contributo di termini diversi (per esempio l&amp;#039;articolo &amp;#039;&amp;#039;&amp;#039;il&amp;#039;&amp;#039;&amp;#039; avrà meno rilevanza di una parola meno comune, e permette l&amp;#039;&amp;#039;&amp;#039;espansione dei termini&amp;#039;&amp;#039;, ovvero l&amp;#039;inclusione di termini relativi ma rilevanti al di là di quelli che si trovano nella sequenza originale.&lt;br /&gt;
&lt;br /&gt;
La cosa più importante di SPLADE non è necessariamente che possa creare termini sostitutivi, ma che possa &amp;#039;&amp;#039;&amp;#039;apprenderli&amp;#039;&amp;#039;&amp;#039;: esso può utilizzare i migliori modelli di linguaggio per imparare queste somiglianze e anche personalizzarli in base al contesto della frase.&lt;br /&gt;
&lt;br /&gt;
=== Apprendimento in SPLADE ===&lt;br /&gt;
[[File:Embeddingmatrix.png|miniatura|I Vettori della matrice di embedding representano un token in uno spazio vettoriale]]&lt;br /&gt;
Normalmente si parte da [[BERT]] utilizzando una &amp;#039;&amp;#039;head&amp;#039;&amp;#039; di [[Masked-Language-Modeling (MLM)]]: per maggiori  dettagli si veda il link sul sito di Pinecone in basso. Quello che più importa è che, alla fine del processo di apprendimento del transformer, in cui le distribuzioni di probabilità per tutti i token di input su tutto il vocabolario (I x V) vengono aggretate per ottenere l&amp;#039;&amp;#039;&amp;#039;&amp;#039;importanza di ogni parola del vocabolario rispetto alla nostra frase&amp;#039;&amp;#039;&amp;#039;: questo è il vettore sparso prodotto da SPLADE.&lt;br /&gt;
&lt;br /&gt;
=== Implementazioini ===&lt;br /&gt;
Esistono diverse implementazioni, tra cui:&lt;br /&gt;
&lt;br /&gt;
* L&amp;#039;implementazione fine-tuned su HuggingFace: naver/splade-cocondenser-ensembledistil&lt;br /&gt;
* L&amp;#039;implementazione di Naver Labs: &amp;lt;nowiki&amp;gt;https://github.com/naver/splade.git&amp;lt;/nowiki&amp;gt;&lt;br /&gt;
&lt;br /&gt;
=== Links ===&lt;br /&gt;
[https://arxiv.org/abs/2109.10086 SPLADE v2: Sparse Lexical and Expansion Model for Information Retrieval]&lt;br /&gt;
&lt;br /&gt;
[https://colab.research.google.com/github/pinecone-io/examples/blob/master/learn/search/semantic-search/sparse/splade/splade-vector-generation.ipynb Colab Ufficiale (Google)]&lt;br /&gt;
&lt;br /&gt;
[https://www.pinecone.io/learn/splade/ SPLADE for Sparse Vector Search Explained] ([[Pinecone]])&lt;br /&gt;
&lt;br /&gt;
https://github.com/pinecone-io/examples/blob/master/learn/search/hybrid-search/ecommerce-search/ecommerce-search.ipynb&lt;br /&gt;
&lt;br /&gt;
[[Categoria:Modello]]&lt;/div&gt;</summary>
		<author><name>Alesaccoia</name></author>
	</entry>
</feed>