Overview
ResNei — Research Neighbourhood – is an AI-augmented environment designed to transform how we discover, analyse, and connect ideas. At its core is the Research Discovery Engine, which constructs a living, responsive knowledge graph through the distillation of verified concepts and the dynamic linking of an evolving corpus of scientific knowledge. This graph is structured as a set of Conceptual Nexus Models (CNMs)—modular representations of connected ideas, designed to surface signals to support impactful inquiry and collaboration. Our UX design moves away from conventional passive, attention-driven systems toward a dynamic, action-intention model that centres collaboration, responsiveness, and contextual insight. A world where research becomes not just efficient, but responsive, intuitive, and interconnected. This platform aims to scale research processes through the interplay of generalists and specialists. Our platform involves a graphical concept model, enabling specialists to engage with reliable evaluation of scientific output, while democratising the otherwise intractable web of knowledge for the generalist. The co-evolution of expert-lead early adoption with generalist participation drives insight and growth: where research thrives through their collaboration. Grounded in scientific protocols and social accountability, ResNei values reciprocity, interdependence, and access. Innovation here acknowledges its real-world costs—social, ecological, and temporal—and builds not just tools, but shared capacity for meaningful discovery. Project Status: As of the second version of this publication in August 2025, this Solution Design Document outlines the foundational concepts, architecture, and guiding principles of ResNei. The project is currently in early development, with a set of working prototypes and a technical paper in preprint. This document supports the transition from conceptual design to implementation.
Overview source: Curated paper metadata.
Methods and contributions
Read the source for the full argument, qualifications, and evidence.
Findings and contributions
- The document presents ResNei as an AI-augmented research discovery and collaboration environment, and states the project is in early development with working prototypes.
- It proposes an initial force-directed concept graph, similar to a simplified Obsidian.md structure, deferring clustering and semantic zooming to later work.
- To counter misplaced trust, the design frames AI outputs as provisional suggestions and the knowledge graph as a navigational aid rather than a definitive map.
- The authors state they are now seeking stakeholders, with special attention to trans- and interdisciplinary spaces, as they move from prototyping to implementation.
Methods
- Action-Intention interaction model (no predictive intent inference yet) — Design treats explicit user actions (uploading, annotating, opening a concept map) as signals of research direction; the document states intent is not yet inferred predictively.
- Knowledge graph of Conceptual Nexus Models (Research Discovery Engine) — The proposed core engine builds a knowledge graph structured as Conceptual Nexus Models, modular representations of connected ideas from a research corpus.
- Proposed full-stack architecture (PDF.js, ElasticSearch, graph DB, Docker) — Specifies frontend/backend components including PDF.js and LaTeX rendering, NLP document processing, ElasticSearch indexing, a graph database, and Docker/CI deployment.
- Assumptions and harms/risks/amelioration tables — Tabulates nine usage assumptions and a set of potential harms (e.g. algorithmic misrepresentation, marginalisation of less-cited research) with proposed mitigations.
- Neuro-informed cognitive load design principles — Applies progressive disclosure, contextual cues, fault tolerance (undo/redo, version history), and adaptive complexity, referencing a neuroscience-informed UX table.
Summary sources: Paper metadata and evidence · Extracted source text.
Document downloads
PDFs are hosted here; supporting sources may use pinned archives.
- 2025_ResNei.pdf PDF · 1.76 MiB · extracted-text source
Citation
Citation metadata follows the unified bibliography.
Catalog details and resources
Related works
Ranked by recorded methods, topics, contributors and associated software, with contributor roles and domain-only matches labelled. Metadata overlap is not a citation or a scientific equivalence.