Researcher Influence Prediction using Academic Genealogy Graph (ResIP)
This repository accompanies the paper:
Researcher influence prediction (ResIP) using academic genealogy network
Dhananjay Kumar, Plaban Kumar Bhowmick, Jiaul H. Paik
Journal of Informetrics, Volume 17, Issue 2, Article 101392 (2023)
https://doi.org/10.1016/j.joi.2023.101392
Overview
In academia, researchers join a research community over time and contribute to the advancement of a field in a variety of ways. One of the most established ways to contribute is by passing on knowledge to future generations through academic advising. While many academic scholars go on to have significant influence, others fail to make a lasting impact. Academic influence, in this sense, refers to a researcher's ability to pass on their "academic gene" to future researchers.
This work proposes the task of Researcher Influence Prediction (ResIP) — predicting a researcher's future influence in an academic field by modelling the growth of their academic lineage over time. Researcher influence prediction has implications for several academic outcomes, including funding, awards, career progression, collaboration, and identifying prolific researchers.
Approach
To address ResIP, a number of end-to-end deep learning architectures are proposed. Each architecture:
- Takes as input the lineage graph of a researcher at a given time point.
- Predicts the growth of that researcher's academic family at future time points.
- Uses an encoder that captures both the temporal and structural information of the input lineage graph.
- Uses decoders tuned to the nature of the desired output — either a single future point or a full future sequence.
Data
The proposed models are trained, validated, and compared against strong baselines using datasets built from a subset of researchers drawn from the Mathematics Genealogy Project (MGP).
Citation
If you use this work, please cite:
@article{kumar2023resip,
title = {Researcher influence prediction ({ResIP}) using academic genealogy network},
author = {Kumar, Dhananjay and Bhowmick, Plaban Kumar and Paik, Jiaul H.},
journal = {Journal of Informetrics},
volume = {17},
number = {2},
pages = {101392},
year = {2023},
doi = {10.1016/j.joi.2023.101392}
}