Meet User Generated Lists Da Cao 1,2 , Liqiang Nie 3 , Xiangnan He 4 - - PowerPoint PPT Presentation

▶
meet user generated lists
SMART_READER_LITE
LIVE PREVIEW

Meet User Generated Lists Da Cao 1,2 , Liqiang Nie 3 , Xiangnan He 4 - - PowerPoint PPT Presentation

Mar 24, 2023 467 likes •753 views

Individual Items Meet User Generated Lists Da Cao 1,2 , Liqiang Nie 3 , Xiangnan He 4 , Xiaochi Wei 5 , Shunzhi Zhu 2 , Shunxiang Wu 1 , Tat-Seng Chua 4 1. Xiamen University; 2. Xiamen University of Technology; 3. Shandong University; 4. National

slide-1
SLIDE 1

Individual Items Meet User Generated Lists

Da Cao1,2, Liqiang Nie3, Xiangnan He4, Xiaochi Wei5, Shunzhi Zhu2, Shunxiang Wu1, Tat-Seng Chua4

  • 1. Xiamen University; 2. Xiamen University of Technology; 3. Shandong University;
  • 4. National University of Singapore; 5. Beijing Institute of Technology

8/19/2017 1

slide-2
SLIDE 2

Outline

  • Background
  • Proposed Method
  • Experiments and Results
  • Conclusion

8/19/2017 2

slide-3
SLIDE 3

User Generated Booklists

8/19/2017 3

slide-4
SLIDE 4

User Generated Playlists

8/19/2017 4

The illustrations of 1) a user’s preference over lists; 2) the user’s preference

  • ver items within lists; and 3) relationships among items and lists.
slide-5
SLIDE 5

To the best of our knowledge

8/19/2017 5

  • Factorization approaches &

embedding-based algorithms

  • User generated list

recommendation task First Second

slide-6
SLIDE 6

Challenges

  • The relationship among items within a list
  • New-item cold-start
  • User-item and user-list recommendation

8/19/2017 6

slide-7
SLIDE 7

Outline

  • Background
  • Proposed Method
  • Experiments and Results
  • Conclusion

8/19/2017 7

slide-8
SLIDE 8

Proposed Method

8

  • Bayesian Personalized Ranking [Rendle et al. 2009]
  • Word Embedding as Matrix Factorization [Levy et al. 2014]
  • Embedding Model for Sentences
  • Utilizing Lists as Side-Information
  • Jointly Recommending Items and Lists

8/19/2017

slide-9
SLIDE 9

Framework

9 8/19/2017

Utilizing Lists as Side- Information (EFM-Side) Bayesian Personalized Ranking Word Embedding as Matrix Factorization Jointly Recommending Items and Lists (EFM-Joint) Bayesian Personalized Ranking Embedding Model for Sentences

slide-10
SLIDE 10

Sentence2vec

10 8/19/2017

… wm,n-c wm,n-1 wm,n+1 projection wm,n wm,n+c … sm … wm,1 wm,2 projection wm,N sm words in the context of a sentence words and a sentence in the context of a word

slide-11
SLIDE 11

Utilizing Lists as Side-Information

11 8/19/2017

slide-12
SLIDE 12

Jointly Recommending Items and Lists

12 8/19/2017

slide-13
SLIDE 13

New-Item Cold-Start

8/19/2017 13

The illustration of the new-item cold-start problem where cold-start items only exist in lists and are never consumed by users.

slide-14
SLIDE 14

Outline

  • Background
  • Proposed Method
  • Experiments and Results
  • Conclusion

8/19/2017 14

slide-15
SLIDE 15

Data Statistics

8/19/2017 15

slide-16
SLIDE 16

Research Questions

(RQ1) Overall performance comparison w.r.t. individual item recommendation. (RQ2) New-item cold-start problem. (RQ3) Performance analysis w.r.t. items. (RQ4) Overall performance comparison w.r.t. item and list recommendation. (RQ5) Importance of items within a list.

8/19/2017 16

slide-17
SLIDE 17

Baseline Methods

  • BPR [Rendle et al. 2009] (benchmark method)
  • BPR-map [Gantner et al. 2010] (two-step model)
  • LIRE [Liu et al. 2014] (list recommendation)
  • CoFactor [Liang et al. 2016] (relationship among items)

8/19/2017 17

slide-18
SLIDE 18

Individual Items Recommendation (RQ1)

8/19/2017 18

Overall performance comparison under the EFM-Side framework

slide-19
SLIDE 19

New-Item Cold-Start Problem (RQ2)

8/19/2017 19

Models comparison in handling the new-item cold-start problem

slide-20
SLIDE 20

Performance Analysis w.r.t. Items (RQ3)

8/19/2017 20

Micro-analysis w.r.t. items with different scale of accumulated ratings.

slide-21
SLIDE 21

8/19/2017 21

Jointly Recommend Items and Lists (RQ4)

Overall performance comparison under the EFM-Joint framework w.r.t. item recommendation. Overall performance comparison under the EFM-Joint framework w.r.t. list recommendation.

slide-22
SLIDE 22

8/19/2017 22

Importance of Items within a List (RQ5)

The similarity between the list and its contained items.

slide-23
SLIDE 23

Outline

  • Background
  • Proposed Method
  • Experiments and Results
  • Conclusion

8/19/2017 23

slide-24
SLIDE 24

Challenges Solved

8/19/2017 24

  • The relationship among items within a list
  • New-item cold-start
  • User-item and user-list recommendation
slide-25
SLIDE 25

Website https://listrec.wixsite.com/efms

8/19/2017 25

slide-26
SLIDE 26

8/19/2017 26

Thanksgiving

slide-27
SLIDE 27

8/19/2017 27

Da Cao Assistant Professor in Hunan University caoda@hnu.edu.cn; caoda0721@gmail.com