Overview#
Bench 2026 is committed to maintaining a rigorous, fair, and transparent peer-review process to ensure the quality and integrity of research contributions in Evaluation Science and Engineering.
The review process follows a strict double-blind peer-review policy. Reviewers are expected to provide objective, constructive, and confidential evaluations based solely on the scientific and technical merits of submissions.
Under the continuous-submission model of Bench 2026, reviews are conducted on a continuous basis throughout the review cycle.
Review Process#
Double-Blind Review#
Bench 2026 adopts a strict double-blind review process.
During the review and discussion process:
- Reviewers do not know the identities of other reviewers;
- Reviewers should not attempt to identify authors;
- Reviewer discussions should focus exclusively on technical arguments and evidence.
The anonymity of reviewers is intended to encourage open discussion and ensure that decisions are based on the quality of arguments rather than seniority, reputation, or authority.
Continuous Review Model#
Bench 2026 adopts a continuous-review model.
Review assignments are made throughout the review cycle based on:
- Reviewer expertise;
- Research relevance;
- Reviewer availability.
Reviewers should complete assigned reviews within the requested timeline to ensure efficient and timely communication of research results.
Review Criteria#
Reviewers should evaluate submissions according to the following criteria.
Scientific and Technical Contribution#
Reviewers should assess:
- Significance of the research problem;
- Novelty of the proposed ideas, methods, systems, datasets, or practices;
- Technical correctness and rigor;
- Quality of supporting evidence.
Evaluation Methodology#
Because Bench focuses on Evaluation Science and Engineering, reviewers should pay particular attention to:
- Soundness of evaluation methodology;
- Validity of benchmarks, datasets, metrics, and measurements;
- Fairness and reproducibility of comparisons;
- Appropriateness of experimental design;
- Reliability of reported conclusions.
Impact and Practical Value#
Reviewers should consider:
- Potential impact on research or engineering practice;
- Generalizability of the proposed approach;
- Contribution to evaluation theory, methodology, or practice.
Related Work and Novelty Assessment#
When reviewers identify that a submission substantially overlaps with prior work:
- They should provide appropriate citations to the relevant prior work;
- Their assessment should be based on objective comparison.
When requesting comparisons with concurrent work or non-archival preliminary work, reviewers should not use the absence of such comparisons as the sole reason for lowering evaluation scores.
Reviewer Responsibilities#
Provide Constructive Feedback##
Reviewers should provide useful and actionable feedback.
Reviews should be:
- Professional;
- Respectful;
- Objective;
- Constructive.
Reviewers should identify both strengths and weaknesses and provide suggestions that may help authors improve their work.
Personal criticism or dismissive comments are inappropriate.
Review the Submitted Version#
Reviewers should evaluate the current submitted manuscript.
If a reviewer has previously reviewed an earlier version of the work:
- The review should be based on the current submission;
- Improvements and changes should be considered objectively.
Write Reviews Independently#
Reviewers must compose their own reviews and provide independent assessments.
Reviewers should not:
- Ask others to write reviews on their behalf;
- Share review assignments without permission;
- Obtain external reviews without consulting the Program Chairs.
If reviewers involve students as part of formal research training, reviewers remain fully responsible for the submitted review and must ensure confidentiality.
Confidentiality and Ethical Requirements#
All submissions, reviews, and discussions are strictly confidential.
Reviewers must:
- Not share manuscripts with others;
- Not distribute submitted materials;
- Not disclose review discussions;
- Not reveal paper identities or review outcomes;
- Not use unpublished information obtained through review.
Reviewers should not:
- Download or retain unnecessary materials from the review system;
- Discuss submitted papers outside officially authorized review channels;
- Reveal author identities even if they recognize the authors.
Reviewer confidentiality remains in effect until the work is publicly available.
Conflicts of Interest#
Reviewers should not review papers where a conflict of interest exists.
Potential conflicts include:
- Recent collaboration with authors;
- Direct academic or professional relationships;
- Institutional conflicts;
- Personal relationships that may affect objectivity.
If a reviewer recognizes authors or suspects a conflict of interest, they should contact the Program Chairs and request reassignment if necessary.
Reviewers should not actively search for author identities.
Reviewer Discussions and Decisions#
Reviewer discussions should remain focused on technical arguments and evidence.
Reviewers should:
- Respect different opinions;
- Evaluate arguments based on merit;
- Avoid attempts to influence others through authority or reputation.
Review recommendations should reflect both:
- Absolute merit of the submission;
- Relative assessment among reviewed submissions when ranking is required.
The final decision is made by the Program Committee based on reviewer feedback and overall program considerations.
Reviewer Workload#
Bench 2026 appreciates the contributions of all reviewers.
Under the continuous-review model:
- Program Committee members participate in reviews throughout the 2026 review cycle;
- The expected workload remains comparable to previous Bench conferences;
- Each Program Committee member is expected to review no more than three submissions on average during the review cycle.
Reporting Ethical Issues#
Reviewers should promptly contact the Program Chairs if they encounter:
- Potential plagiarism;
- Duplicate submission;
- Author attempts to identify reviewers;
- Attempts to influence review outcomes improperly;
- Other ethical concerns.
Acknowledgments#
The reviewer ethics guidelines of Bench 2026 are adapted from established practices of leading international conferences, including MICRO, ASPLOS, and ISCA, with adjustments for the continuous-submission and publication model of Bench 2026.