Map Citation Networks with AI-Powered Mind Maps
Use AI mind maps to trace citation chains, identify seminal papers, and discover intellectual lineages. Branch forward and backward through reference networks.
Explore the reference map
A real citation network around Kahneman and Tversky's prospect theory, zoomed out to the whole lineage. Zoom in on any paper to read what it argued.
How it works
Citation networks are the hidden skeleton of academic knowledge. Every paper sits within a web of influences, extensions, and critiques. Understanding where a paper comes from and where it led reveals more than reading the paper alone. Mind maps make these networks navigable.
-
Start with a seed paper or concept. Enter a seminal paper, a key finding, or a theoretical concept. The AI maps the immediate citation context: foundational works it builds on, contemporaneous related papers, and major works that cite it.
-
Branch backward into intellectual roots. Expand the “cited by this paper” direction. Trace the foundational assumptions, the methodological lineage, and the debates that motivated the work. Each backward branch reveals the intellectual soil the paper grew from.
-
Branch forward into influence. Expand the “papers that cite this” direction. Map how the ideas were extended, applied to new domains, challenged, or synthesized with other frameworks. Forward branches show the intellectual legacy and current frontier.
-
Trace lateral connections. The most interesting citations cross disciplinary boundaries. Branch into unexpected connections: a psychology paper cited in economics, a physics method adopted in biology. These lateral branches often contain the most innovative work.
Why branching matters for citation mapping
Citation networks are inherently branching structures. A single paper might cite 40 references and be cited by 400 subsequent works. Listing these in a spreadsheet gives you a flat index. Mapping them as branches gives you the intellectual topology — which lineages are dense and active, which have gone dormant, where two previously separate branches converged.
Branching also captures the temporal dimension of citation networks. A forward branch from a 1979 paper might show a burst of activity in the 1990s, a quiet period in the 2000s, and a revival in the 2020s driven by new methods. This temporal pattern tells you something about the field’s dynamics that no single paper reveals. The mind map makes these patterns visible by letting you expand any time period and see what changed.
Example
You’re tracing the intellectual lineage of transformer architectures in NLP. Starting from “Attention Is All You Need” (2017), you branch backward to Bahdanau attention (2014) and sequence-to-sequence models (2014). Forward branches split into BERT-style masked language models and GPT-style autoregressive models. You discover that the scaling laws branch — studying how performance changes with model size — emerged surprisingly late (2020) and drew on much older work in neural scaling from the physics community. This cross-disciplinary connection suggests an underexplored area where physics intuitions might further inform architecture design.
Pair this workflow with literature survey to build a comprehensive understanding of a field, or use systematic review when you need to move from exploratory citation tracing to formal evidence synthesis. For understanding how research ideas become applied innovations, deep dive explores how research ideas connect and where gaps exist.