Master’s Thesis Presentation • Artificial Intelligence — End-to-end Neural Information RetrievalExport this event to calendar

Friday, April 5, 2019 10:00 AM EDT

Wei Yang, Master’s candidate
David R. Cheriton School of Computer Science

In recent years, we have witnessed many successes of neural networks in the information retrieval community with lots of labeled data. Yet it remains unknown whether the same techniques can be easily adapted to search social media posts where the text is much shorter. In addition, we find that most neural information retrieval models are compared against weak baselines. 

In this thesis, we build an end-to-end neural information retrieval system using two toolkits: Anserini and MatchZoo. In addition, we also propose a novel neural model to capture the relevance of short and varied tweet text, named MP-HCNN. 

With the information retrieval toolkit Anserini, we build a reranking architecture based on various traditional information retrieval models (QL, QL+RM3, BM25, BM25+RM3), including a strong IR baseline: RM3. With the neural network toolkit MatchZoo, we offer an empirical study of a number of popular neural network ranking models (DSSM, CDSSM, KNRM, DUET, DRMM). Experiments on datasets from the TREC Microblog Tracks and TREC Robust Retrieval Track show that \textit{most} existing neural network models cannot beat a simple language model baseline. However, DRMM provides a significant improvement over the baseline (BM25+RM3) on the Robust04 dataset and DUET, DRMM and MP-HCNN can provide significant improvements over the baseline (QL+RM3) on the microblog datasets. 

Further detailed analyses suggest that searching social media and searching news articles exhibit several different characteristics that require customized model design, shedding light on future work.

Location 
DC - William G. Davis Computer Research Centre
3317
200 University Avenue West

Waterloo, ON N2L 3G1
Canada

S M T W T F S
30
31
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
1
2
3
  1. 2024 (100)
    1. April (23)
    2. March (27)
    3. February (25)
    4. January (25)
  2. 2023 (296)
    1. December (20)
    2. November (28)
    3. October (15)
    4. September (25)
    5. August (30)
    6. July (30)
    7. June (22)
    8. May (23)
    9. April (32)
    10. March (31)
    11. February (18)
    12. January (22)
  3. 2022 (245)
  4. 2021 (210)
  5. 2020 (217)
  6. 2019 (255)
  7. 2018 (217)
  8. 2017 (36)
  9. 2016 (21)
  10. 2015 (36)
  11. 2014 (33)
  12. 2013 (23)
  13. 2012 (4)
  14. 2011 (1)
  15. 2010 (1)
  16. 2009 (1)
  17. 2008 (1)