The Performance of ChatGPT-4V in Interpreting Images and Tables in the Japanese Medical Licensing Exam
Bibliographic record
Abstract
The recent introduction of Chat Generative Pre-trained Transformer 4 Vision (ChatGPT-4V) has expanded the capabilities of language models to include image input features, potentially broadening their application in the medical field. This Research Letter evaluates the performance of ChatGPT-4V in interpreting clinical images and tables through the Japanese Medical Licensing Exam (JMLE). Employing the September 25, 2023, version of ChatGPT-4V, the study compared the program’s responses to the 117th JMLE against the passing criteria and the average scores of human examinees. While ChatGPT-4V surpassed the passing threshold with an 85.1% correct response rate in essential knowledge and 76.5% in other areas, it fell short in image-based (71.9%) and table-based questions (35.0%), indicating a significant gap compared to human performance. This suggests limitations in the model’s image and table interpretation, exacerbated by its lower proficiency in non-Latin characters and potential overreliance on text information. Despite its success in passing the JMLE, the study highlights the need for further development of ChatGPT-4V to enhance its reliability for medical applications, including diagnostics.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.029 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.005 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".