A Representation Theory for Ranking Functions

Authors: Harsh H Pareek, Pradeep K Ravikumar

NeurIPS 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental For our experiments, we consider the information retrieval learning to rank task, where we are given a training set consisting of n queries. ... Table 1 shows some results across LETOR datasets which show improvements over the base rankers.
Researcher Affiliation Academia Harsh Pareek, Pradeep Ravikumar Department of Computer Science University of Texas at Austin {harshp,pradeepr}@cs.utexas.edu
Pseudocode No The paper describes the functional form of the ranking function (Equation 14) but does not present it as structured pseudocode or an algorithm block.
Open Source Code No The paper mentions using 'Rank Lib' and provides a link to its source, but there is no explicit statement or link indicating that the authors' own methodology code is open-source or publicly available.
Open Datasets Yes We use the LETOR 3.0 collection [23], which contains the OHSUMED dataset and the Gov collection: HP2003/04, TD2003/04, NP2003/04, which respectively correspond to the listwise Homepage Finding, Topic Distillation and Named Page Finding tasks.
Dataset Splits Yes LETOR contains 5 predefined folds with training, validation and test sets. We use these directly and report averaged results on the test set.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies Yes To obtain the scores b for the baseline pointwise ranking function, we used Ranklib v2.1-patched with its default parameter values.
Experiment Setup Yes For the ℓ2 regularization parameter, we pick a C from [0, 1e-5,1e-2, 1e-1, 1, 10, 1e2,1e3] tuning for maximum NDCG@10 on the validation set. We used gradient descent on w to fit parameters. ... and used the initial value w = 0 for our experiments.