| name | aaai |
| description | Writing and formatting guidance for AAAI Conference on Artificial Intelligence papers. Use when drafting, revising, or formatting papers for AAAI 2026 submission or camera-ready. Covers aaai2026 LaTeX class, anonymous submission mode, page limits, reference formatting with natbib, figure/table placement, Times font requirements, and CCF-A category requirements. |
AAAI 2026 Conference Writing Skill
CCF-A | AI | Publisher: AAAI
Conference: https://aaai.org/conference/aaai/aaai-26
Template: ccf-latex-templates/AAAI/aaai2026_template.tex
Document Setup
Preamble Structure
\documentclass[letterpaper]{article}
% Switch between modes:
% Uncomment for anonymous submission:
% \def\aaaianonymous{true}
% Conditional package loading:
\ifdefined\aaaianonymous
\usepackage[submission]{aaai2026}
\else
\usepackage{aaai2026}
\fi
% REQUIRED fonts (DO NOT CHANGE):
\usepackage{times}
\usepackage{helvet}
\usepackage{courier}
\usepackage[hyphens]{url}
\usepackage{graphicx}
\urlstyle{rm}
\def\UrlFont{\rm}
% REQUIRED (no options):
\usepackage{natbib}
\usepackage{caption}
\frenchspacing
\setlength{\pdfpagewidth}{8.5in}
\setlength{\pdfpageheight}{11in}
% Required pdfinfo:
\pdfinfo{
/TemplateVersion (2026.1)
}
Required Packages
Always include these packages in the exact order shown. The template enforces Times font via times, helvet, and courier packages — this is mandatory and cannot be changed.
Submission vs Camera-Ready
| Mode | Declaration | Effect |
|---|
| Anonymous Submission | \def\aaaianonymous{true} | Hides author info, no copyright notice |
| Camera-Ready | Comment out \def\aaaianonymous | Shows authors, includes copyright |
Page Limits
- Main content: 7 pages
- References: 1 page (separate)
- Total submission: 8 pages maximum
- Camera-ready: Follow conference-specific guidelines
Anonymity Requirements
For anonymous submission mode:
- Author names and affiliations must not appear
- Use "Anonymous Submission" as sole author
- Anonymize self-citations (cite as third person)
- Clear PDF metadata with a metadata-cleaning tool
- Do not include links that reveal identity
- References to authors' previous work must be anonymized
Title and Author Formatting
% Single institution:
\author{
Author Name
}
\affiliations{
Affiliation Name\\
Address Line 1\\
email@example.com
}
% Multiple institutions with superscripts:
\author{
Author One\textsuperscript{\rm 1},
Author Two\textsuperscript{\rm 2}
}
\affiliations{
\textsuperscript{\rm 1}Affiliation 1\\
\textsuperscript{\rm 2}Affiliation 2\\
first@affil1.com, second@affil2.com
}
Use \equalcontrib for equal contribution footnotes.
Abstract and Links
\begin{abstract}
Your abstract here (no references in abstract).
\end{abstract}
% For camera-ready (only):
\begin{links}
\link{Code}{https://example.com/code}
\link{Datasets}{https://example.com/datasets}
\end{links}
Section Organization
Arrange sections in this order:
- Main content sections (numbered)
- Appendices (optional, use
\appendix then lettered sections)
- Ethical Statement (optional, unnumbered)
- Acknowledgments (optional, unnumbered)
- References (unnumbered, last section)
Recommended Section Structure for AI Papers
- Introduction — Problem statement, motivation, contributions (enumerate 3-5 key contributions)
- Related Work — Background and positioning against prior work
- Preliminaries — Mathematical notation, definitions, background concepts
- Methodology — Core algorithm/model with sufficient technical depth
- Experiments — Datasets, baselines, metrics, results, statistical analysis
- Conclusion — Summary, limitations, future work
Writing the Introduction (Critical for AAAI)
AAAI reviewers often decide within the first page. Structure your introduction as:
\section{Introduction}
% Paragraph 1: Problem and motivation (2-3 sentences)
The problem of X has become increasingly important due to Y...
% Paragraph 2: Gap in existing work (2-3 sentences)
Prior work has addressed X through approaches A, B, and C.
However, these approaches suffer from limitations including...
% Paragraph 3-4: Our approach (3-4 sentences)
In this paper, we propose a novel approach that...
Key contributions of our work include:
\begin{itemize}
\item A new formulation of X as a Y problem...
\item An efficient algorithm with O(n) complexity...
\item Extensive experiments on 5 benchmark datasets...
\end{itemize}
% Paragraph 5: Paper roadmap (optional)
The remainder of this paper is organized as follows...
Figures and Tables
- Use
.jpg, .png, or .pdf only (no .eps, .ps, .gif)
- Place at top of page when possible
- Caption under the figure, above the table
- 9-point minimum for caption text, 10-point roman
- Use
booktabs for professional tables (no vertical rules)
- Number figures/tables sequentially
\begin{figure}[t]
\centering
\includegraphics[width=0.9\linewidth]{figure}
\caption{Figure caption here.}
\label{fig:example}
\end{figure}
% Professional table with booktabs:
\begin{table}[t]
\caption{Main results on benchmark datasets. Best results in \textbf{bold}.}
\label{tab:results}
\centering
\begin{tabular}{lccc}
\toprule
Dataset & Method A & Method B & Ours \\
\midrule
Dataset1 & 85.2 & 87.4 & \textbf{91.3} \\
Dataset2 & 72.1 & 75.8 & \textbf{79.2} \\
\bottomrule
\end{tabular}
\end{table}
References (natbib)
The aaai2026.sty automatically sets the bibliography style. Use natbib commands:
% In preamble:
\usepackage{natbib} % NO options
% In document:
\bibliography{your_bibfile}
% Citation commands:
\citet{key} % (Author Year)
\shortcite{key} % (Year)
\citeauthor{key} % Author
\citeyear{key} % Year
For self-citations in anonymous mode: cite in third person ("Smith et al. showed..." not "we showed").
Bibliography Best Practices
% Good bibliography entry:
@inproceedings{smith2024example,
author = {Smith, John and Doe, Jane},
title = {A Novel Approach to X},
booktitle = {Proceedings of AAAI},
year = {2024},
pages = {1234--1245},
publisher = {AAAI}
}
% For anonymous submission, cite as:
Smith et al. showed that... % Instead of We showed...
Disallowed Commands and Packages
Forbidden packages:
authblk, balance, color, CJK, float, flushend, fontenc
fullpage, geometry, grffile, hyperref, navigator
multicol, nameref, savetrees, setspace, stfloats
tabu, titlesec, tocbibind, ulem, wrapfig
Forbidden commands:
\nocopyright, \addtolength, \balance, \baselinestretch
\clearpage, \columnsep, \newpage, \pagebreak
\pagestyle, \tiny, \vspace{-...}, \vskip{-...}
Typography Rules
- Font: Times Roman or Nimbus (required)
- Size: 10-point text with 12-point leading
- Margins: 0.75" sides, 0.75" top (non-first pages), 1.25" bottom
- Columns: Two-column format (required)
- No color text — color only in figures (CMYK, not RGB)
- No Type 3 fonts in any portion including graphics
AI-Specific Writing Conventions
Mathematical Notation
Define all notation clearly before use:
\section{Preliminaries}
Let $\mathcal{X} = \{x_1, \dots, x_n\}$ denote the set of input instances...
We use bold lowercase letters $\mathbf{x}$ for vectors and uppercase
$\mathbf{X}$ for matrices. $\|\mathbf{x}\|_2$ denotes the Euclidean norm.
Algorithm Description
\begin{algorithm}[tb]
\caption{The Proposed Algorithm}
\label{alg:proposed}
\begin{algorithmic}[1]
\REQUIRE Dataset $\mathcal{D}$, learning rate $\eta$, epochs $T$
\ENSURE Trained model parameters $\theta$
\STATE Initialize $\theta_0$ randomly
\FOR{$t = 1$ to $T$}
\STATE Sample mini-batch $\mathcal{B} \sim \mathcal{D}$
\STATE Compute gradient $\nabla \mathcal{L}(\theta_{t-1}; \mathcal{B})$
\STATE Update $\theta_t \leftarrow \theta_{t-1} - \eta \nabla \mathcal{L}$
\ENDFOR
\RETURN $\theta_T$
\end{algorithmic}
\end{algorithm}
Experiment Reporting Standards
AAAI expects rigorous experimental evaluation:
\section{Experiments}
\subsection{Setup}
We evaluate on four benchmark datasets:
\begin{itemize}
\item \textbf{Dataset A}: X instances, Y features, source: URL
\item \textbf{Dataset B}: ...
\end{itemize}
Metrics: Accuracy, F1-score (macro-averaged), and AUC-ROC.
We compare against: (1) baseline method, (2) state-of-the-art method A,
(3) state-of-the-art method B. All experiments run 5 times with different
random seeds; we report mean $\pm$ standard deviation.
\subsection{Main Results}
Results are summarized in \autoref{tab:results}.
Our method outperforms all baselines with statistically significant
improvements (paired t-test, $p < 0.01$).
\subsection{Ablation Study}
We ablate each component of our approach in \autoref{tab:ablation}.
Removing component X degrades performance by Y\%, confirming its importance.
Rebuttal Preparation
AAAI uses author responses during review:
\section*{Response to Reviewers}
% Address each concern systematically
\textbf{Regarding Reviewer 1's concern about X:}
We thank the reviewer for this comment. To clarify...
[Provide additional evidence, analysis, or explanation]
\textbf{Regarding the comparison with Y:}
We added new experiments comparing with Y in the supplementary material.
Results show our method remains superior even under the suggested settings.
Rebuttal Do's and Don'ts
DO:
- Thank reviewers for constructive feedback
- Address every concern systematically
- Provide new evidence when asked
- Be polite and professional
- Acknowledge valid criticisms
DON'T:
- Be defensive or argumentative
- Dismiss concerns without explanation
- Make unsubstantiated claims
- Argue about scores
Camera-Ready Specifics
After acceptance:
- Comment out
\def\aaaianonymous{true}
- Add complete author/affiliation information
- Include links environment with code/dataset URLs
- Sign and submit AAAI copyright form
- Submit as single
.tex file (no \input for sections)
- Include
.bib file(s) separately
Camera-Ready Checklist
Checklist Before Submission