An official website of the United States government
BACKGROUND Expert feedback lays the foundation of rigorous research. However, the rapid growth of scholarly production challenges the conventional scienti c feedback mechanisms. High-quality peer reviews are increasingly dif cult to obtain. METHODS We created an automated pipeline using Generative Pretrained Transformer 4 (GPT-4) to provide comments on scienti c papers. We evaluated the quality of GPT-4’s feedback through two large-scale studies. We rst quantitatively compared GPT-4’s gen- erated feedback with human peer reviewers’ feedback in general scienti c papers from 15 Nature family journals (3096 papers in total) and the International Conference on Learning Representations (ICLR) machine learning conference (1709 papers). To speci - cally assess GPT-4’s performance on biomedical papers, we also analyzed a subset of 425 health sciences papers from the Nature portfolio and a random sample of 666 sub- missions to eLife. Additionally, we conducted a prospective user study with 308 research- ers from 110 institutions in the elds of arti cial intelligence and computational biology to understand how researchers perceive feedback generated by our system on their own papers. RESULTS The overlap in the points raised by GPT-4 and by human reviewers (average overlap of 30.85% for Nature journals and 39.23% for ICLR) is comparable with the over- lap between two human reviewers (average overlap of 28.58% for Nature journals and 35.25% for ICLR). Results on eLife and a subset of health sciences papers as categorized by the Nature portfolio show similar patterns. In our prospective user study, more than half (57.4%) of the users found GPT-4–generated feedback helpful/very helpful, and 82.4% found it more bene cial than feedback from at least some human reviewers. We also identify several limitations of large language model (LLM)–generated feedback. CONCLUSIONS Through both retrospective and prospec- tive evaluation, we nd substantial overlap between LLM and human feedback as well as positive user perceptions regarding the usefulness of LLM feedback. Although human expert review should continue to be the foundation of the scienti c process, LLM feedback could bene t researchers, especially when timely expert feedback is not available and in earlier stages of manuscript preparation. (Funded by the Chan–Zuckerberg Initiative and the Stanford Interdisciplin- ary Graduate Fellowship.)
more »
« less
Online Learning Using Only Peer Prediction
This paper considers a variant of the classical online learning problem with expert predictions. Our model’s differences and challenges are due to lacking any direct feedback on the loss each expert incurs at each time step $$t$$. We propose an approach that uses peer prediction and identify conditions where it succeeds. Our techniques revolve around a carefully designed peer score function $s()$$ that scores experts’ predictions based on the peer consensus. We show a sufficient condition, that we call \emph{peer calibration}, under which standard online learning algorithms using loss feedback computed by the carefully crafted $$s()$ have bounded regret with respect to the unrevealed ground truth values. We then demonstrate how suitable $s()$ functions can be derived for different assumptions and models.
more »
« less
- Award ID(s):
- 2007951
- PAR ID:
- 10287777
- Editor(s):
- Chiappa, Silvia; Calandra, Roberto
- Date Published:
- Journal Name:
- Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics
- Volume:
- 108
- Page Range / eLocation ID:
- 2032--2042
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This paper studies learning-augmented decentralized online convex optimization in a networked multi-agent system, a challenging setting that has remained under-explored. We first consider a linear learning-augmented decentralized online algorithm (LADO-Lin) that combines a machine learning (ML) policy with a baseline expert policy in a linear manner. We show that, while LADO-Lin can exploit the potential of ML predictions to improve the average cost performance, it cannot have guaranteed worst-case performance. To address this limitation, we propose a novel online algorithm (LADO) that adaptively combines the ML policy and expert policy to safeguard the ML predictions to achieve strong competitiveness guarantees. We also prove the average cost bound for LADO, revealing the tradeoff between average performance and worst-case robustness and demonstrating the advantage of training the ML policy by explicitly considering the robustness requirement. Finally, we run an experiment on decentralized battery management. Our results highlight the potential of ML augmentation to improve the average performance as well as the guaranteed worst-case performance of LADO.more » « less
-
We study online convex optimization with switching costs, a practically important but also extremely challenging problem due to the lack of complete offline information. By tapping into the power of machine learning (ML) based optimizers, ML-augmented online algorithms (also referred to as expert calibration in this paper) have been emerging as state of the art, with provable worst-case performance guarantees. Nonetheless, by using the standard practice of training an ML model as a standalone optimizer and plugging it into an ML-augmented algorithm, the average cost performance can be highly unsatisfactory. In order to address the how to learn challenge, we propose EC-L2O (expert-calibrated learning to optimize), which trains an ML-based optimizer by explicitly taking into account the downstream expert calibrator. To accomplish this, we propose a new differentiable expert calibrator that generalizes regularized online balanced descent and offers a provably better competitive ratio than pure ML predictions when the prediction error is large. For training, our loss function is a weighted sum of two different losses --- one minimizing the average ML prediction error for better robustness, and the other one minimizing the post-calibration average cost. We also provide theoretical analysis for EC-L2O, highlighting that expert calibration can be even beneficial for the average cost performance and that the high-percentile tail ratio of the cost achieved by EC-L2O to that of the offline optimal oracle (i.e., tail cost ratio) can be bounded. Finally, we test EC-L2O by running simulations for sustainable datacenter demand response. Our results demonstrate that EC-L2O can empirically achieve a lower average cost as well as a lower competitive ratio than the existing baseline algorithms.more » « less
-
The emphasis on conceptual learning and the development of adaptive instructional design are both emerging areas in science and engineering education. Instructors are writing their own conceptual questions to promote active learning during class and utilizing pools of these questions in assessments. For adaptive assessment strategies, these questions need to be rated based on difficulty level (DL). Historically DL has been determined from the performance of a suitable number of students. The research study reported here investigates whether instructors can save time by predicting DL of newly made conceptual questions without the need for student data. In this paper, we report on the development of one component in an adaptive learning module for materials science – specifically on the topic of crystallography. The summative assessment element consists of five DL scales and 15 conceptual questions This adaptive assessment directs students based on their previous performances and the DL of the questions. Our five expert participants are faculty members who have taught the introductory Materials Science course multiple times. They provided predictions for how many students would answer each question correctly during a two-step process. First, predictions were made individually without an answer key. Second, experts had the opportunity to revise their predictions after being provided an answer key in a group discussion. We compared expert predictions with actual student performance using results from over 400 students spanning multiple courses and terms. We found no clear correlation between expert predictions of the DL and the measured DL from students. Some evidence shows that discussion during the second step made expert predictions closer to student performance. We suggest that, in determining the DL for conceptual questions, using predictions of the DL by experts who have taught the course is not a valid route. The findings in this paper can be applied to assessments in both in-person, hybrid, and online settings and is applicable to subject matter beyond materials science.more » « less
-
Given a pre-trained classifier and multiple human experts, we investigate the task of online classification where model predictions are provided for free but querying humans incurs a cost. In this practical but under-explored setting, oracle ground truth is not available. Instead, the prediction target is defined as the consensus vote of all experts. Given that querying full consensus can be costly, we propose a general framework for online Bayesian consensus estimation, leveraging properties of the multivariate hypergeometric distribution. Based on this framework, we propose a family of methods that dynamically estimate expert consensus from partial feedback by producing a posterior over expert and model beliefs. Analyzing this posterior induces an interpretable trade-off between querying cost and classification performance. We demonstrate the efficacy of our framework against a variety of baselines on CIFAR-10H and ImageNet-16H, two large-scale crowdsourced datasets.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
