Next Article in Journal
Granzyme B in Autoimmune Skin Disease
Next Article in Special Issue
DL-TODA: A Deep Learning Tool for Omics Data Analysis
Previous Article in Journal
Biomolecules from Macroalgae—Nutritional Profile and Bioactives for Novel Food Product Development
Previous Article in Special Issue
KinFams: De-Novo Classification of Protein Kinases Using CATH Functional Units
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Biomolecules
Volume 13
Issue 2
10.3390/biom13020387
Review Report
biomolecules-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

Printed Edition

A printed edition of this Special Issue is available at MDPI Books....
share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
Article
Peer-Review Record

Using GPT-3 to Build a Lexicon of Drugs of Abuse Synonyms for Social Media Pharmacovigilance

Biomolecules 2023, 13(2), 387; https://doi.org/10.3390/biom13020387
by Kristy A. Carpenter 1 and Russ B. Altman 1,2,*
Reviewer 2:
Iain J. Cruickshank
Biomolecules 2023, 13(2), 387; https://doi.org/10.3390/biom13020387
Submission received: 11 January 2023 / Revised: 9 February 2023 / Accepted: 16 February 2023 / Published: 18 February 2023
(This article belongs to the Special Issue Biomolecular Data Science—in Honor of Professor Philip E. Bourne)

Round 1

Reviewer 1 Report

In this study, authors used GPT-3 to analyze social media text for the purpose of pharmacovigilance to solve the problem of drug abuse. The study argues and tries to prove that GPT-3 is valuable for social media pharmacovigilance. The authors made the data and the code available. I have the following comments:

1. At the end of the Introduction Section, a list of well-defined contributions must be added.

2. All abbreviations should be defined in their first use.

3. How did you define the “Widely-discussed drugs of abuse”?

4. A figure for the full pipeline is needed to show the full picture of the proposal. 

5. The parameters tuning has been done manually. How can you assure that these are the optimum parameters? The same for Google Search Depth Analysis, why did not use an automatic search optimization algorithm.

6. Did you check different formats for the queries?

7. “… For both the set of alprazolam iterations and the set of fentanyl  iterations, we observed that terms generated only once or twice were overwhelmingly manually labeled as non-synonyms, and that most, but not all, terms generated more than 15 times were manually labeled as synonym…” Is this sufficient to know that a term is synonym? How you consulted domain experts? 

8. In the “Pipeline performance”, what is the benchmark performance that you are comparing with. How can we know that these are the best results. How can we know that a term is indeed a synonym for a drug? You are using three filters to determine, but who said that the result after the three filters is correct?

9. Authors are required to share the results of the lexicon of GPT-3 synonyms for drugs of abuse.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Please see attached document for comments and suggestions.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Accept.

Thank you.

Back to TopTop