KDD 2025 2nd-round Review Results: How Did Your Paper Do?

[cocktailfreedom]

KDD reviews are coming out, and most community-shared scores cluster around 2.6 to 3.6 on average. Common patterns are like 333xx, 433xx, or 422xx, showing that many papers are seen as average in novelty and technical quality. Even submissions authors are proud of are mostly getting 3s, with only a few 4s or 5s. The overall vibe: “low scores are normal, let’s just hope for kind reviewers and make the most of rebuttal.”

In short, KDD remains tough, and scores are modest across the board.

[magicparrots]

A data point:

GNN work, got

Novelty: 3, 2, 2, 3, 2
Technical Quality: 2, 2, 2, 2, 2
Confidence: 3, 4, 3, 4, 4

Need to rebuttal? anyone knows more? 2 weeks challenge ahead!

[Sylvia]

I made a comparison of KDD 2024 vs. KDD 2025 scoring/reviewing system. Here you go!

Scoring Dimensions and Their Scales

Scoring Dimension	KDD 2024	KDD 2025	Change
Relevance	1–4	1–4	No change
Novelty	1–5	1–4	Reduced
Technical Quality	1–5	1–4	Reduced
Presentation Quality	1–5	1–4	Reduced
Reproducibility	1–5	1–4	Reduced
Reviewer Confidence	1–5	1–4	Reduced

Note: The reduction from a 5-point to a 4-point scale compresses the neutral midpoint, encouraging reviewers to take a clearer stance on each dimension.

---

Review Form Structure Changes

Review Element	KDD 2024	KDD 2025	Change
Paper Summary, Strengths, Weaknesses	Required (Free-form)	Required (Free-form)	No change
Questions for Rebuttal	Optional / General	Required: Numbered, specific	New requirement
Resubmission Flag	Not included	"Resubmission" + "Repeat Reviewer"	New
Ethics Review Flag	Yes / No	Yes / No	No change
LLM Usage Disclosure	Not asked	Mandatory	New

---

Emphasis in KDD 2025

Rebuttal Process:

• Authors benefit from clearly numbered, targeted reviewer questions.
• Reviewers are expected to provide actionable feedback.

Transparency:

• Reviewers must disclose any use of Large Language Models (LLMs).
• Tracks resubmission history and reviewer continuity.

Reproducibility:

• Still emphasized, with refined grading from "insufficient" to "excellent" support materials.

---

A summary table

Area	KDD 2024	KDD 2025	Key Difference
Scoring Scale	1–5 (most categories)	1–4 (all categories)	️ Compressed scale
Review Structure	Free-form + ratings	Structured + specific queries	More actionable
Rebuttal Support	Optional	Mandatory, numbered	Enforced
LLM Disclosure	Not applicable	Required	New
Resubmission Tracking	Not tracked	Explicitly included	New

[lelecao]

My reproducibility score hurt a lot because of my source code link does not work any more. I was using LimeWire + ShortURL. Real bad service!

Next time, I will use CSPaper!!

https://cspaper.org/category/10/anonymous-sharing-supplementary-materials

Here is an example:

https://cspaper.org/topic/38/kdd2025-2nd-tgn-adapted-anonymous-source-code-for-review-only

[river]

KDD community stats

https://papercopilot.com/statistics/kdd-statistics/kdd-2025-statistics/

[river]

@lelecao I feel you, been there too!

[river]

I made a summary of data points from KDD 2025 1st round results:

Novelty Scores	Technical Quality Scores	Confidence Scores	Rebuttal Outcome	Final Decision	Notes
3 3 3 3 3 3	4 3 3 2 3 2	–	Addressed issues	Accepted	"Rebuttal 一波三折太难了"
2 2 3 2 2	3 3 2 2 3	3 3 3 3 3	Submitted	Rejected	"是不是可以直接跑路了"
4 3 3 1	4 4 2 2	–	Explained issues	Rejected	"Large variance across reviewers; no score changes post-rebuttal"
3 3 3	3 3 2	–	Unsure	🟡 Unknown	"Still considering rebuttal; not sure if it's worth the effort"
3 3 3 3 3 3	3 3 3 3 3 2	–	Minor clarifications	Accepted	"Final scores unchanged but accepted after positive AC decision"
3 4 3 3 3 3	2 2 3 2 2 3	–	Clarified results	Rejected	"Novelty OK, but TQ too weak; didn't convince reviewers"
3 3 3 3 3 3	3 3 3 3 3 3	3 3 3 3 3	Submitted	Accepted	"Strong consensus; one of the smoother cases"
3 3 3	3 3 2	–	No rebuttal	Rejected	"No rebuttal submitted; borderline scores"
3 3 2 2	3 3 2 2	–	Rebuttal sent	Rejected	"Reviewers did not change their opinion"
3 3 3 3 3 3	3 3 3 3 2 2	–	Rebuttal helped	Accepted	"Accepted despite one weaker reviewer"
3 3 3 3	3 3 3 3	3 3 3 3	Rebuttal sent	🟡 Unknown	"In limbo; waiting for final decision"
3 3 3 3	2 2 2 2	–	Not convincing	Rejected	"Work deemed not ‘KDD-level’ despite rebuttal"
3 3 3 3 3 3	3 3 3 3 3 3	3 3 3 3 3	Submitted	Accepted	"Perfectly consistent reviewers; smooth acceptance"
3 3 3 2	3 3 2 2	–	Rebuttal failed	Rejected	"Low technical quality and variance led to rejection"

Note: Data sourced from community discussions on Zhihu, Reddit, and OpenReview threads. Subject to sample bias.

[lelecao]

@magicparrots

So sorry to hear that — sounds like a solid paper.

For my case,
One reviewer gave two 2s just because they didn’t see the value of improving efficiency or where it would be useful, even though that’s the whole point of many ML contributions. Another reviewer didn’t understand the paper and asked for line-by-line comments on pseudocode. That’s just disheartening.

Also noticed each review response is limited to 2500 characters. Does anyone know if we can reply in multiple stacked comments?

[Kevin Cris]

https://www.zhihu.com/question/12035973262/answers/updated
some data points from Chinese researcher community

[Hu8kKo34]

Anyone knows the likelihood of an NLP (LLM agent and its evaluation on many public datasets) work accepted to KDD, either main or applied data science track?