Auto saved by Logseq
This commit is contained in:
@@ -0,0 +1,125 @@
|
||||
file:: [KARS_2023_submission_7_1692635907726_0.pdf](../assets/KARS_2023_submission_7_1692635907726_0.pdf)
|
||||
file-path:: ../assets/KARS_2023_submission_7_1692635907726_0.pdf
|
||||
|
||||
- CF models calculate recommendations leveraging similarities in interaction/preference patterns of similar users.
|
||||
ls-type:: annotation
|
||||
hl-page:: 1
|
||||
hl-color:: green
|
||||
id:: 64ec59ef-c3c6-40f6-a075-7dd1b4649b50
|
||||
- To compensate for this lack of information, an increasing number of approaches combine collaborative information with auxiliary content attributes, such as tags, metadata, and geographical data
|
||||
ls-type:: annotation
|
||||
hl-page:: 1
|
||||
hl-color:: green
|
||||
id:: 64ec5a10-97c8-4be7-a2e0-bc9c0937ff50
|
||||
- Knowledge graphs (KGs) have the advantage of covering a variety of heterogeneous domains, thus making it easier to transpose the advances in one domain to another.
|
||||
ls-type:: annotation
|
||||
hl-page:: 1
|
||||
hl-color:: green
|
||||
id:: 64ec5a2f-55a9-4e2b-8e29-99357dfc8505
|
||||
- existing dataset and the knowledgeaware recommendation works that use them
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec60f6-83fa-416e-aa29-f5b65059d4f8
|
||||
- challenges we have faced and methodologies we have adopted
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec6100-9c7f-4e67-8332-48e2c1e060b2
|
||||
- most used datasets in the recent literature
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec611d-b8e4-4df5-83dc-0c58ba11d611
|
||||
- Unsurprisingly, they relate with streaming and e-commerce platforms and cover the typical domains of movies, music, books, and shops
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec6129-a7f5-444d-8ac5-1027e041950d
|
||||
- ovieLens 1M and DBook datasets to the DBpedia KG and investigate the idea of transfer learning between item recommendations and KG completion tasks leveraging a co-factorization model.
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec613e-5871-478a-b680-6b6aef08170e
|
||||
- This large amount of work using side information and enriched datasets shows the potential of these tools
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec6199-961d-402c-8466-0dad64b1707e
|
||||
- hese works use datasets whose linking strategies are highly different from each other and lack details for their reproducibility, putting reuse within other works at risk.
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec61a7-9cb1-48f6-a49e-6d33b2e80c50
|
||||
- Wang et al. [30] link MovieLens 1M with IMDb information, such as genres, actors, directors, and writers to use it in a knowledgeaware model.
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec61c6-fafe-4357-a1f2-aea78c1f8fbb
|
||||
- mapping of MovieLens 1M to the DBpedia KG
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec61d2-04b7-48ec-bc4c-e7f12e7b891a
|
||||
- hao et al. [32] propose KB4Rec, a public linking of MovieLens 20M, LastFM 1B, and Amazon Book to the popular KB Freebase
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec61de-a82f-4275-b09e-188adc047b73
|
||||
- LibraryThing to a KG, specifically to DBpedia, likely due to some domain limitations detailed in Section 4.
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: green
|
||||
id:: 64ec61ee-2ac8-4575-88b6-fa5087587969
|
||||
- These drawbacks have animated us to release extensive linkings of the large MovieLens 25M and LibraryThing datasets to DBpedia, Wikidata, and Freebase and describe our methodology in detail.
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: yellow
|
||||
id:: 64ec62eb-0601-45cb-97b7-b80f79dce40f
|
||||
- evaluation of our linking strategy,
|
||||
ls-type:: annotation
|
||||
hl-page:: 2
|
||||
hl-color:: blue
|
||||
id:: 64ec633a-5274-4166-94e2-c8d5bca747df
|
||||
hl-stamp:: 1693213500751
|
||||
- nriched versions of MovieLens25M (ML25M) and LibraryThing (LT), thus guaranteeing the reproducibility of any experiment making use of additional knowledge.
|
||||
ls-type:: annotation
|
||||
hl-page:: 3
|
||||
hl-color:: yellow
|
||||
id:: 64ec636a-1ddd-4858-97df-096a2d682014
|
||||
hl-stamp:: 1693213552695
|
||||
- These datasets constantly increase in size regarding ratings, users, and items, with the release that we hereby consider including about 25 million ratings. We have complemented the items in MovieLens 25M with links to resources of Wikidata, DBpedia, and Freebase knowledge graphs
|
||||
ls-type:: annotation
|
||||
hl-page:: 3
|
||||
hl-color:: yellow
|
||||
id:: 64ec63ee-0096-49e3-96db-6ec160582571
|
||||
- we have not filtered the retrieved information, leaving each researcher the choice to select
|
||||
ls-type:: annotation
|
||||
hl-page:: 3
|
||||
hl-color:: green
|
||||
id:: 64ecc65a-e06f-4786-91d5-a22729da04c8
|
||||
- In this section, we present methodologies, challenges, and solutions adopted for performing the item linkings described in Section 3.
|
||||
ls-type:: annotation
|
||||
hl-page:: 3
|
||||
hl-color:: green
|
||||
id:: 64ecc66c-20f1-4e5e-a357-1b218c1c2186
|
||||
- Our goal is to sketch best practices and highlight issues that raise from item linking, with the aim of encouraging the discussion about the need of a common workflow to collect augumented datasets.
|
||||
ls-type:: annotation
|
||||
hl-page:: 3
|
||||
hl-color:: blue
|
||||
id:: 64ecc678-99cb-44b0-98e3-98b69d3ff8b3
|
||||
- This work provides supplementary data of two wellknown and widely used recommendation datasets (i.e., MovieLens 25M and LibraryThing) that can be utilized in tasks involving Knowledge Graphs
|
||||
ls-type:: annotation
|
||||
hl-page:: 5
|
||||
hl-color:: blue
|
||||
id:: 64ecc96c-6c2e-49ac-8f0b-8c5ec69b264a
|
||||
- differently from a traditional linking task where a textual context is provided, only titles and authors are retrieved from LibraryThing
|
||||
ls-type:: annotation
|
||||
hl-page:: 5
|
||||
hl-color:: blue
|
||||
id:: 64ecc98f-586a-4140-bd78-f0ff29ce1ceb
|
||||
- In fact, the MovieLens dataset contains titles, IMDb, and TMDb IDs. Therefore, we can exploit it to evaluate the performance of the linkers by fixing the available side information, i.e., the label of the film and its director taken from IMDb
|
||||
ls-type:: annotation
|
||||
hl-page:: 5
|
||||
hl-color:: blue
|
||||
id:: 64ecc9a6-08a4-43be-a915-b435c575c830
|
||||
Reference in New Issue
Block a user