Compact Aspect Embedding for Diversified Query Expansions
Authors: Xiaohua Liu, Arbi Bouchoucha, Alessandro Sordoni, Jian-Yun Nie
AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We test our method on several TREC diversification data sets, and show that our method significantly outperforms the state-of-the-art search result diversification approaches. |
| Researcher Affiliation | Academia | Xiaohua Liu, Arbi Bouchoucha, Alessandro Sordoni and Jian-Yun Nie Dept. of Computer Science and Operations Research University of Montreal Montreal (Quebec), Canada {liuxiao, bouchoar, sordonia and nie}@iro.umontreal.ca |
| Pseudocode | Yes | Algorithm 1 Query Term Diversification. |
| Open Source Code | No | The paper does not provide any specific link or statement indicating that the source code for their methodology is publicly available. |
| Open Datasets | No | The paper mentions using 'Clue Web09 (category B) dataset', 'test queries from TREC 2009, 2010 and 2011 Web tracks', and 'log data of Microsoft Live Search 2006'. While TREC data is widely known, the paper does not provide concrete access information (e.g., specific links, DOIs, or formal citations with access details) for all datasets used, particularly the Microsoft Live Search logs. The footnote for ClueWeb09 links to a spam ranking page, not the dataset itself. |
| Dataset Splits | Yes | We build a development data set consisting of 10 randomly selected queries (from TREC 2009, 2010 and 2011 Web tracks) to fine tune parameters |
| Hardware Specification | No | The paper does not provide any specific hardware details (e.g., GPU/CPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper mentions submitting queries to 'Indri' and provides a general project link ('http://www.lemurproject.org/indri.php'), but it does not specify a version number for Indri or any other software dependencies. |
| Experiment Setup | Yes | Parameter Setting. We build a development data set consisting of 10 randomly selected queries (from TREC 2009, 2010 and 2011 Web tracks) to fine tune parameters, and obtain the following configuration: β, the trade-off parameter of MMRE defined in Formula 4 is set to 0.5; η, the parameter that controls the trade-off between Frobenius-norm loss and trace norm loss, is set to 1; the number of dimensions of an expansion term vector (N) is fixed to 30; K, the number of expansion term candidates, is set to 100; the parameter that controls the trade-off between relevance and nonredundancy in MMREQ (Bouchoucha, Liu, and Nie 2014) is set to 0.6. |