Back to Home
AI Research Glossary
Clear definitions of 30+ terms related to AI research tools, multi-agent systems, citation verification, and academic writing. Perfect for students and researchers getting started with AI-assisted research.
A
- AI Agent
- A specialized AI system designed to perform a specific task autonomously. Unlike general-purpose AI, agents have focused capabilities and can work together in multi-agent systems. OpenDraft uses 19 specialized agents for different aspects of research paper generation. Learn more: Multi-Agent AI for Research →
- AI Hallucination
- When an AI model generates false or fabricated information that appears plausible. In academic contexts, this commonly manifests as fake citations—references to papers, authors, or journals that don't exist. Studies show ChatGPT hallucinates 30-50% of academic citations. Learn more: How to Avoid AI Hallucination →
- arXiv
- A free, open-access repository of over 2 million scientific preprints in physics, mathematics, computer science, and related fields. Papers on arXiv haven't been peer-reviewed but provide early access to cutting-edge research. OpenDraft searches arXiv for verified citations.
C
- Citation Verification
- The process of confirming that academic citations are real and accurate. This involves checking that the paper exists, the authors are correct, and the DOI or URL is valid. OpenDraft verifies all citations against Semantic Scholar, CrossRef, and arXiv databases. Learn more: AI Citation Verification →
- CrossRef
- A registration agency for scholarly metadata containing over 130 million DOI records. CrossRef allows verification of publication details including authors, titles, and publication dates. It's a primary source for citation verification in academic writing tools.
D
- DOI (Digital Object Identifier)
- A permanent, unique identifier assigned to digital objects like academic papers. DOIs (e.g., 10.1000/xyz123) provide persistent links to publications even if the URL changes. They're the gold standard for verifying that a citation is real.
E
- Elicit
- An AI research assistant that helps find relevant papers and extract key information. Elicit uses semantic search to find papers related to research questions. It's useful for systematic reviews but doesn't generate full paper drafts. Learn more: OpenDraft vs Competitors →
F
- Few-Shot Learning
- An AI technique where models learn to perform tasks from just a few examples, rather than large datasets. In research AI, this enables tools to adapt to specific citation styles or writing conventions with minimal training.
G
- Grounding
- Connecting AI outputs to verified external sources. Grounded AI systems cite real sources rather than generating plausible-sounding but potentially fabricated information. OpenDraft grounds all citations in verified academic databases.
J
- Jenni AI
- A commercial AI writing assistant designed for academic content. Jenni AI offers features like autocomplete, citation assistance, and paraphrasing. It's subscription-based, unlike open-source alternatives. Learn more: OpenDraft vs Competitors Comparison →
L
- Large Language Model (LLM)
- AI systems trained on vast amounts of text data that can generate human-like text. Examples include GPT-4, Claude, and Gemini. While powerful, LLMs can hallucinate facts and citations when used without verification systems.
- Literature Review
- A comprehensive survey of existing research on a topic. Literature reviews identify key findings, methodological approaches, and gaps in knowledge. They're essential components of theses, dissertations, and research papers. Learn more: How to Write a Literature Review with AI →
M
- Multi-Agent System
- An AI architecture where multiple specialized agents collaborate to complete complex tasks. Each agent handles a specific subtask (planning, searching, writing, etc.). OpenDraft's 19-agent system outperforms single-model approaches for research tasks. Learn more: Multi-Agent AI for Research →
O
- Open Source
- Software whose source code is freely available for inspection, modification, and redistribution. Open-source tools like OpenDraft provide transparency about how they work, which is important for academic integrity and reproducibility.
P
- Peer Review
- The evaluation of academic work by experts in the same field before publication. Peer review helps ensure research quality but can take months. Preprint servers like arXiv allow sharing work before peer review.
- Planner Agent
- In OpenDraft's multi-agent system, Planner agents break down research topics into structured subtasks. They analyze the user's topic and create a roadmap for Scout, Signal, and Scribe agents to follow. Learn more: Multi-Agent AI for Research →
- Preprint
- A version of a research paper shared before formal peer review. Preprints allow rapid dissemination of findings but haven't been validated by peer review. arXiv, bioRxiv, and medRxiv are major preprint servers.
- Prompt Engineering
- The practice of designing effective inputs (prompts) for AI models to produce desired outputs. For academic AI, good prompts specify the desired format, tone, citation style, and level of detail. Learn more: ChatGPT for Thesis Writing Tutorial →
R
- RAG (Retrieval-Augmented Generation)
- An AI technique that combines information retrieval with text generation. RAG systems search databases for relevant information before generating responses, reducing hallucinations by grounding outputs in real data.
S
- Scout Agent
- In OpenDraft's architecture, Scout agents search academic databases (Semantic Scholar, CrossRef, arXiv) to find relevant papers. They return verified metadata including titles, authors, abstracts, and DOIs. Learn more: Multi-Agent AI for Research →
- Scribe Agent
- OpenDraft's writing agents that generate content based on verified sources. Scribe agents synthesize information from Scout-retrieved papers into coherent academic prose with proper citations. Learn more: Multi-Agent AI for Research →
- Semantic Scholar
- A free AI-powered research database containing over 200 million academic papers. Semantic Scholar provides APIs for programmatic access to paper metadata, citations, and abstracts. It's OpenDraft's primary citation source.
- Semantic Search
- Search that understands meaning rather than just matching keywords. Semantic search finds papers about 'climate change effects on agriculture' even if they use different terminology. It's more effective than keyword search for research.
- Signal Agent
- In OpenDraft's system, Signal agents evaluate the relevance and quality of papers found by Scout agents. They rank papers by relevance to the research topic, helping ensure the final output cites the most appropriate sources. Learn more: Multi-Agent AI for Research →
- Systematic Review
- A structured, reproducible method for identifying and analyzing all relevant research on a topic. Systematic reviews follow explicit protocols and are common in medical and social science research.
T
- Thesis
- An extended research document presenting original scholarship, typically required for graduate degrees. Master's theses are usually 50-100 pages; PhD dissertations are often 200+ pages with substantial original research. Learn more: Complete Guide to AI Thesis Writing →
- Token
- The basic unit of text that AI models process. A token is roughly 4 characters or 0.75 words in English. API pricing and context limits are often measured in tokens (e.g., GPT-4's 128K token context window).
- Turnitin
- A plagiarism detection service widely used by universities. Turnitin compares submitted work against its database of academic papers, websites, and previous submissions. AI-generated content may not trigger traditional plagiarism detection.
V
- Vector Database
- A database optimized for storing and searching high-dimensional vectors (numerical representations of text). Vector databases enable semantic search by finding documents with similar meaning, not just matching keywords.
Z
- Zero-Shot Learning
- An AI capability to perform tasks without specific training examples. Modern LLMs can follow instructions for tasks they weren't explicitly trained on, enabling flexible academic writing assistance.