ExplainabilityTransparencyTrustworthinesscs.AI
Trends in Explainable AI (XAI) Literature
Alon Jacovi
January 13, 2023
arXiv.org
21 citations
3 influential
Key Takeaways
-
XAI research is becoming increasingly fragmented but more discoverable: The field historically suffered from decentralized terminology and scattered publication venues, but researchers have recently converged around common keywords, making it easier to systematically identify and track XAI papers across the 5,199-paper collection studied.
-
Explainable AI is expanding beyond computer science: The research shows relative growth in XAI papers from non-CS scientific fields, indicating that explainability concerns are becoming important across diverse disciplines beyond traditional AI research communities.
-
Collaboration across fields is accelerating: The paper documents increasing cross-field collaborative authorship and citation activity, suggesting that XAI is becoming a genuinely multidisciplinary effort rather than remaining siloed within computer science departments.
-
XAI research influence extends beyond its own community: The collection can identify papers that are influential outside their immediate field, highlighting how explainability concerns are shaping research agendas in other scientific domains and suggesting XAI principles are gaining broader relevance for responsible AI implementation.
Abstract
The XAI literature is decentralized, both in terminology and in publication venues, but recent years saw the community converge around keywords that make it possible to more reliably discover papers automatically. We use keyword search using the SemanticScholar API and manual curation to collect a well-formatted and reasonably comprehensive set of 5199 XAI papers, available at https://github.com/alonjacovi/XAI-Scholar . We use this collection to clarify and visualize trends about the size and scope of the literature, citation trends, cross-field trends, and collaboration trends. Overall, XAI is becoming increasingly multidisciplinary, with relative growth in papers belonging to increasingly diverse (non-CS) scientific fields, increasing cross-field collaborative authorship, increasing cross-field citation activity. The collection can additionally be used as a paper discovery engine, by retrieving XAI literature which is cited according to specific constraints (for example, papers that are influential outside of their field, or influential to non-XAI research).