At this yr’s Worldwide Convention on Machine Studying (ICML2025), Jaeho Kim, Yunseok Lee and Seulki Lee received an excellent place paper award for his or her work Place: The AI Convention Peer Evaluate Disaster Calls for Creator Suggestions and Reviewer Rewards. We hear from Jaeho concerning the issues they had been attempting to handle, and their proposed creator suggestions mechanism and reviewer reward system.
May you say one thing about the issue that you just deal with in your place paper?
Our place paper addresses the issues plaguing present AI convention peer evaluate techniques, whereas additionally elevating questions concerning the future route of peer evaluate.
The upcoming drawback with the present peer evaluate system in AI conferences is the exponential development in paper submissions pushed by growing curiosity in AI. To place this with numbers, NeurIPS acquired over 30,000 submissions this yr, whereas ICLR noticed a 59.8% improve in submissions in only one yr. This large improve in submissions has created a elementary mismatch: whereas paper submissions develop exponentially, the pool of certified reviewers has not stored tempo.
Submissions to among the main AI conferences over the previous few years.
This imbalance has extreme penalties. Nearly all of papers are not receiving enough evaluate high quality, undermining peer evaluate’s important operate as a gatekeeper of scientific information. When the evaluate course of fails, inappropriate papers and flawed analysis can slip via, probably polluting the scientific document.
Contemplating AI’s profound societal affect, this breakdown in high quality management poses dangers that stretch far past academia. Poor analysis that enters the scientific discourse can mislead future work, affect coverage selections, and finally hinder real information development. Our place paper focuses on this vital query and proposes strategies on how we are able to improve the standard of evaluate, thus main to raised dissemination of information.
What do you argue for within the place paper?
Our place paper proposes two main modifications to sort out the present peer evaluate disaster: an creator suggestions mechanism and a reviewer reward system.
First, the creator suggestions system allows authors to formally consider the standard of critiques they obtain. This technique permits authors to evaluate reviewers’ comprehension of their work, determine potential indicators of LLM-generated content material, and set up primary safeguards in opposition to unfair, biased, or superficial critiques. Importantly, this isn’t about penalizing reviewers, however fairly creating minimal accountability to guard authors from the small minority of reviewers who might not meet skilled requirements.
Second, our reviewer incentive system supplies each fast and long-term skilled worth for high quality reviewing. For brief-term motivation, creator analysis scores decide eligibility for digital badges (akin to “Prime 10% Reviewer” recognition) that may be displayed on educational profiles like OpenReview and Google Scholar. For long-term profession affect, we suggest novel metrics like a “reviewer affect rating” – primarily an h-index calculated from the next citations of papers a reviewer has evaluated. This treats reviewers as contributors to the papers they assist enhance and validates their position in advancing scientific information.
May you inform us extra about your proposal for this new two-way peer evaluate methodology?
Our proposed two-way peer evaluate system makes one key change to the present course of: we cut up evaluate launch into two phases.
The authors’ proposed modification to the peer-review system.
At present, authors submit papers, reviewers write full critiques, and all critiques are launched directly. In our system, authors first obtain solely the impartial sections – the abstract, strengths, and questions on their paper. Authors then present suggestions on whether or not reviewers correctly understood their work. Solely after this suggestions can we launch the second half containing weaknesses and rankings.
This method presents three predominant advantages. First, it’s sensible – we don’t want to vary present timelines or evaluate templates. The second section may be launched instantly after the authors give suggestions. Second, it protects authors from irresponsible critiques since reviewers know their work might be evaluated. Third, since reviewers sometimes evaluate a number of papers, we are able to monitor their suggestions scores to assist space chairs determine (ir)accountable reviewers.
The important thing perception is that authors know their very own work finest and might shortly spot when a reviewer hasn’t correctly engaged with their paper.
May you discuss concerning the concrete reward system that you just counsel within the paper?
We suggest each short-term and long-term rewards to handle reviewer motivation, which naturally declines over time regardless of beginning enthusiastically.
Quick-term: Digital badges displayed on reviewers’ educational profiles, awarded based mostly on creator suggestions scores. The aim is making reviewer contributions extra seen. Whereas some conferences record prime reviewers on their web sites, these lists are laborious to search out. Our badges can be prominently displayed on profiles and will even be printed on convention identify tags.Instance of a badge that might seem on profiles.
Lengthy-term: Numerical metrics to quantify reviewer affect at AI conferences. We advise monitoring measures like an h-index for reviewed papers. These metrics might be included in educational portfolios, just like how we at present monitor publication affect.
The core thought is creating tangible profession advantages for reviewers whereas establishing peer evaluate as knowledgeable educational service that rewards each authors and reviewers.
What do you suppose might be among the professionals and cons of implementing this technique?
The advantages of our system are threefold. First, it’s a very sensible resolution. Our method doesn’t change present evaluate schedules or evaluate burdens, making it simple to include into present techniques. Second, it encourages reviewers to behave extra responsibly, understanding their work might be evaluated. We emphasize that almost all reviewers already act professionally – nonetheless, even a small variety of irresponsible reviewers can critically injury the peer evaluate system. Third, with enough scale, creator suggestions scores will make conferences extra sustainable. Space chairs can have higher details about reviewer high quality, enabling them to make extra knowledgeable selections about paper acceptance.
Nonetheless, there may be sturdy potential for gaming by reviewers. Reviewers may optimize for rewards by giving overly constructive critiques. Measures to counteract these issues are undoubtedly wanted. We’re at present exploring options to handle this subject.
Are there any concluding ideas you’d like so as to add concerning the potential future
of conferences and peer-review?
One rising development we’ve noticed is the growing dialogue of LLMs in peer evaluate. Whereas we consider present LLMs have a number of weaknesses (e.g., immediate injection, shallow critiques), we additionally suppose they may ultimately surpass people. When that occurs, we are going to face a elementary dilemma: if LLMs present higher critiques, why ought to people be reviewing? Simply because the speedy rise of LLMs caught us unprepared and created chaos, we can not afford a repeat. We must always begin making ready for this query as quickly as attainable.
About Jaeho
|
Jaeho Kim is a Postdoctoral Researcher at Korea College with Professor Changhee Lee. He acquired his Ph.D. from UNIST beneath the supervision of Professor Seulki Lee. His predominant analysis focuses on time sequence studying, notably growing basis fashions that generate artificial and human-guided time sequence information to scale back computational and information prices. He additionally contributes to bettering the peer evaluate course of at main AI conferences, along with his work acknowledged by the ICML 2025 Excellent Place Paper Award. |
Learn the work in full
Place: The AI Convention Peer Evaluate Disaster Calls for Creator Suggestions and Reviewer Rewards, Jaeho Kim, Yunseok Lee, Seulki Lee.
AIhub
is a non-profit devoted to connecting the AI group to the general public by offering free, high-quality info in AI.
AIhub
is a non-profit devoted to connecting the AI group to the general public by offering free, high-quality info in AI.