Program

Siks page: Advances in Information Retrieval course, 2019.

October 7th

1. Claudia Hauff, “Machine Learning for IR”; .pdf, @google;
2. Hinda Haned, “Explainable and fair machine learning”; .pdf, @github;
3. Faegheh Hasibi, “Knowledge graphs & semantic search”; .pdf;
4. Arjen P. de Vries, “Siks Query Variants Experiment” (part 1). .pdf, assignment.

October 8th

1. Djoerd Hiemstra, “Managing terabytes: A Google-sized search index”; .pdf;
2. Suzan Verberne, “Evaluation and user models”; .pdf;
3. Djoerd Hiemstra, “Siks Query Variants Experiment” (part 2); assignment;
4. Mostafa Dehghani, “Learning with Imperfect Supervision”; .pdf;
5. Dimitrios Rafailidis, “Cross-domain Recommendation”; .pdf;
6. David Graus, “Bias in recommendation”; .pdf, @slideshare.

Practical Assignment

As Siks PhD students in the information retrieval course, we will investigate the effect of query formulation on retrieval performance.

Inspiration / background reading: a recent query variations paper by Shane Culpepper; the UQV 100 dataset; classic paper on rank fusion.

Experiments using the OSIRRC jig.

Pick your team and go!

• red
• green
• yellow
• purple
• blue

Results

Results so far are summarised in the slides.

Evaluation Assignment

Two versions of the “Implementing IR Evaluation Measures” Notebook:

• MyBinder “do not feed the google” version
• Google CoLab “feed the google” version