Don't miss our weekly PhD newsletter | Sign up now Don't miss our weekly PhD newsletter | Sign up now

  Query Performance Prediction for Neural Information Retrieval: Application to Conversational Search


   School of Computing, Engineering & the Built Environment

This project is no longer listed on FindAPhD.com and may not be available.

Click here to search FindAPhD.com for PhD studentship opportunities
  Dr Md Zia Ullah  Applications accepted all year round  Self-Funded PhD Students Only

About the Project

In information retrieval (IR), query performance prediction (QPP) aims to predict the search effectiveness for a given query without resorting to relevance judgments. QPP may be advantageous in many ways, such as signaling an IR system whether a search query would be effective or underperforming. Based on that information, the system can either apply a query reformulation [6] or an adaptive retrieval configuration [1,4] or engage in an interactive session with the user (i.e., conversational search [7]) to understand the search intent and provide a better search experience.

Predicting query performance is a challenging problem due to many characteristics of queries, collections, and search systems. Existing QPPs are extracted from traditional retrieval models (e.g., BM25 or Divergence from randomness) using the pre-retrieval features based on the collection statistics or the post-retrieval features based on the top-retrieved documents [2, 3]. With the advent of language models (e.g., BERT [6]), neural IR models have been proposed and shown to have better retrieval effectiveness [8]. However, QPP on the neural retrieval model has not yet been explored [9].

This Ph.D. project aims to develop neural query performance predictors from neural IR models and combine them with existing QPPs based on traditional IR models. Experiments could be conducted on standard TREC collections (e.g., MS MARCO and TREC Deep learning tracks) to demonstrate the effectiveness of the QPPs and compare them with the state-of-the-art approaches. Another goal of this project would be to design appropriate metrics to evaluate the QPPs and apply them to Conversational search [7].

Prospective applicants are encouraged to contact the Supervisor before submitting their applications. Please consult with the Supervisor if you want to work on the broad areas of IR and NLP.

Academic qualifications

A first-class honours degree, or a distinction at master level, or equivalent achievements in Computer Science or Data Science or Software Engineering.

English language requirement

If your first language is not English, comply with the University requirements for research degree programmes in terms of English language.

Application process

Prospective applicants are encouraged to contact the supervisor, Dr. Md Zia Ullah () to discuss the content of the project and the fit with their qualifications and skills before preparing an application. 

The application must include: 

Research project outline of 2 pages (list of references excluded). The outline may provide details about

  • Background and motivation, explaining the importance of the project, should be supported also by relevant literature. You can also discuss the applications you expect for the project results.
  • Research questions or
  • Methodology: types of data to be used, approach to data collection, and data analysis methods.
  • List of references

The outline must be created solely by the applicant. Supervisors can only offer general discussions about the project idea without providing any additional support.

  • Statement no longer than 1 page describing your motivations and fit with the project.
  • Recent and complete curriculum vitae. The curriculum must include a declaration regarding the English language qualifications of the candidate.
  • Supporting documents will have to be submitted by successful candidates.
  • Two academic references (but if you have been out of education for more than three years, you may submit one academic and one professional reference), on the form can be downloaded here.

Applications can be submitted here.

Download a copy of the project details here.

Computer Science (8)

References

[1] Mothe J and Ullah MZ, Defining an Optimal Configuration Set for Selective Search Strategy – A Risk-Sensitive Approach, Proceedings of the 30th ACM International Conference on Information and Knowledge Management (CIKM 2021), 2021.
[2] Déjean S, Ionescu RT, Mothe J, and Ullah MZ, Forward and backward feature selection for query performance prediction, The 35th ACM/SIGAPP Symposium on Applied Computing (SAC), 2020.
[3] Chifu AG, Laporte L, Mothe J, and Ullah MZ, Query Performance Prediction Focused on Summarized Letor Features, The 41st International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR), 2018.
[4] Deveau R, Mothe J, Ullah MZ, Nie JY, Learning to Adaptively Rank Document Retrieval System Configurations, ACM Transactions of Information Systems (ACM TOIS), 41 pages, pp.3:1-3:41, Volume 37, Issue 1, 2019.
[5] Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina, Bert: Pre-training of deep bidirectional transformers for language understanding, arXiv preprint arXiv:1810.04805, 2018.
[6] Azad, Hiteshwar Kumar and Deepak, Akshay, Query expansion techniques for information retrieval: a survey,Information Processing and Management, 2019
[7] Faggioli et. Al., A Geometric Framework for Query Performance Prediction in Conversational Search, ACM SIGIR, 2023
[8] Zhuang et al., Rankt5: Fine-tuning T5 for text ranking with ranking losses, ACM SIGIR, 2023
[9] Faggioli et al., Query Performance Prediction for Neural IR: Are We There Yet? Advances in Information Retrieval, ECIR 2023

 About the Project