Ask the corpus a question directly, or explore it through two interactive lenses — a concept map of how ideas connect, and a semantic scatter of papers by meaning.
Ask my research
Put a question to the corpus behind these visualisations — 1,047 passages from 26 works. Answers are assembled from the passages actually retrieved, and every one is listed below the answer so you can check it.
Retrieval runs on embeddings of my own papers; the wording of answers is generated. Passages are extracts, so treat the sources as the record and the summary as a convenience.
Concept map of my research portfolio — nodes are areas, papers, projects, and methods; edges trace intellectual relationships.
Each node represents a research area (navy), publication (purple), applied deployment (teal), or core methodology (amber). Dashed arrows mark idea evolution from conference papers to journal submissions.
Drag to move · Scroll to zoom · Tap a node to highlight connections
Embedding-based scatter of my work by semantic similarity. Each paper is split into ~200-word passages and embedded individually; the faint cloud shows where a paper's passages land and the solid dot is their centroid. Hover a paper to isolate its own passages. Switch to Concept Pair mode to project onto any two research axes. Requires running embed_pipeline_v2.py first.
Each solid dot is a paper and each faint dot one passage from it — position reflects semantic similarity, so related material clusters together. A widely spread cloud means the work ranges across topics; a tight one means it stays put. Switch to Concept Pair mode to project papers onto any two research axes (x = similarity to concept A, y = concept B).