On the Value of Target Data in Transfer Learning

Authors: Steve Hanneke, Samory Kpotufe

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical Finally, the procedures described in this work remain of a theoretical nature, but yield new insights into how various practical decisions in transfer can be made near-optimally in a data-driven fashion.
Researcher Affiliation Academia Steve Hanneke Toyota Technological Institute at Chicago steve.hanneke@gmail.com Samory Kpotufe Columbia University, Statistics skk2175@columbia.edu
Pseudocode Yes Algorithm 1: Minimize ˆRSP (h) subject to ˆRSQ(h) ˆRSQ(ˆh SQ) c ˆPSQ(h 6= ˆh SQ)An Q + c An Q (6)
Open Source Code No The paper describes theoretical algorithms and does not provide any links or statements about the availability of source code for these algorithms.
Open Datasets No The paper is theoretical and does not mention using or providing access to any specific datasets for training or other empirical evaluation.
Dataset Splits No The paper is theoretical and does not conduct experiments, therefore no training, validation, or test dataset splits are specified.
Hardware Specification No The paper does not conduct experiments and therefore does not provide any details about hardware specifications.
Software Dependencies No The paper does not mention any specific software dependencies with version numbers for reproducibility, as it focuses on theoretical contributions.
Experiment Setup No The paper is theoretical and does not include details about an experimental setup, such as hyperparameters or training configurations.