ICASSP 2026 Submission Ddl around the corner: What You Need to Know
-
With just a few days left before the ICASSP 2026 deadline (September 18, 2025), the community is buzzing with both excitement and anxiety. The recent announcement around participation rules, template changes, and LLM usage policies has sparked discussions across research forums. Below is a comprehensive transcription and elaboration of the article you shared, focusing on the submission and peer review dimensions that matter most to computer science researchers.
Mandatory In-Person Presentation
Perhaps the most striking regulation this year is the strict on-site presentation requirement. The official guideline states:
“All papers should be covered by an author registration, and presented in-person at the conference. Uncovered or no-show papers during the virtual conference poster presentation session will be removed from the conference proceedings.”
In simpler terms:
- Every accepted paper must have at least one author physically attend ICASSP 2026 in Barcelona.
- At least one author must register at the non-student rate before the author registration deadline.
- If no author shows up in person, the paper will be withdrawn from the proceedings — no exceptions.
This rule is partly a reaction to ICASSP 2024 in India, where absentee authors inflated costs and diminished participation quality. Rumors even suggest that on-site staff will check for proxy presenters, aiming to stop the trend of “代讲” (unauthorized substitutes).
Costs and Practical Concerns
The article highlights the financial burden researchers face:
“Flights + accommodation + conference registration… conservatively estimated to exceed 20,000 RMB, with flights especially expensive.”
This has become a talking point: while Barcelona offers sunshine and prestige, the barrier to participation may disproportionately impact students and researchers from underfunded institutions.
New Paper Kit and Formatting Rules
ICASSP 2026 introduces a new Paper Kit. The 2025 template is now invalid, and details such as image size, margins, and abstract length are explicitly regulated.
“Format issues may directly lead to rejection.”
Tips from the article:
- ICASSP uses single-blind review (author and institution names can be visible).
- Page limit concerns can be managed by switching to 9pt font (the minimum allowed).
- Each author can appear on a maximum of 9 submissions.
Researchers are urged to triple-check compliance with the official kit to avoid desk rejections.
LLM Usage: Allowed vs. Forbidden
A major development this year is the mandatory disclosure of LLM assistance. ICASSP aligns with other top-tier conferences by explicitly regulating AI tools:
Allowed:
- Polishing language
- Enhancing figures
- Accelerating code development
- Aiding idea generation
- Literature filtering
Forbidden:
- Having an LLM write the paper
- Generating unverified code
- Producing introductions, related work, or abstracts directly
“If LLM assistance is used, it must be clearly declared in the paper.”
This policy addresses ongoing debates around research integrity and the fine line between assistance and authorship outsourcing.
Why This Matters for Peer Review
ICASSP’s stricter regulations reflect a broader trend: peer review quality and conference reputation hinge on enforceable standards. Researchers face three bottlenecks simultaneously:
- Formatting pitfalls (new Paper Kit).
- Integrity checks (LLM declarations).
- Logistical barriers (mandatory in-person presence).
The result is higher stress — and a stronger need for early, rubric-aligned feedback before submission.
How CSPaper Can Help
This is where CSPaper Review becomes highly relevant. Unlike generic proofreading or grammar tools, it directly addresses the unique pain points raised by ICASSP 2026:
- Garbage in, garbage out — solved: Multiple PDF parsing solutions ensure high-fidelity extraction of text, figures, equations, and references. Clean input = reliable review.
- Conference- and track-specific reviews: CSPaper uses ICASSP’s official review template and rubric, giving feedback aligned with what real reviewers will evaluate.
- Actionable and quantifiable: Reviews are score-based (novelty, clarity, significance), offering concrete guidance instead of vague remarks.
- Continuous benchmarking: The system uses leading VLLMs (GPT, DeepSeek, Llama; Gemini and Claude soon) to ensure reliable, rubric-faithful outputs.
- Verified related work: No hallucinated citations — every suggested reference is checked and justified.
- Desk rejection assessment: Automatic checks on length, topic fit, and quality thresholds provide early warnings.
- Support for arXiv LaTeX: Beyond PDFs, CSPaper reviews LaTeX source, aiding fast and accurate literature assessment.
- Community-driven: Researchers can share feedback and best practices, turning CSPaper into a peer-support ecosystem.
In essence, CSPaper provides a safe rehearsal environment for ICASSP reviews, allowing authors to detect weaknesses, anticipate reviewer feedback, and strengthen papers before the actual submission.
Final Thoughts
“Format, LLM declaration, in-person presenter — miss any one, and it may be too late.”
For computer science researchers, ICASSP 2026 highlights both the tightening of academic standards and the rising cost of participation. Tools like CSPaper bridge the gap by simulating the review process and ensuring that submissions align with conference expectations — before it’s too late.
As deadlines loom, the best strategy is to prepare rigorously, review proactively, and use every available resource to maximize your paper’s chances.
