Rethinking Retrieval-Augmented Generation for Medicine: A Large-Scale, Systematic Expert Evaluation and Practical Insights

Large language models (LLMs) are transforming the landscape of medicine, yet two fundamental challenges persist: keeping up with rapidly evolving medical knowledge and providing verifiable, evidence-grounded reasoning. Retrieval-augmented generation (RAG) has been widely adopted to address these limitations by supplementing model outputs with retrieved evidence. However, whether RAG reliably achieves these goals remains unclear. Here, we present the most comprehensive expert evaluation of RAG in medicine to date. Eighteen medical experts contributed a total of 80,502 annotations, assessing 800 model outputs generated by GPT-4o and Llama-3.1-8B across 200 real-world patient and USMLE-style queries. We systematically decomposed the RAG pipeline into three components: (i) evidence retrieval (relevance of retrieved passages), (ii) evidence selection (accuracy of evidence usage), and (iii) response generation (factuality and completeness of outputs). Contrary to expectation, standard RAG often degraded performance: only 22% of top-16 passages were relevant, evidence selection remained weak (precision 41-43%, recall 27-49%), and factuality and completeness dropped by up to 6% and 5%, respectively, compared with non-RAG variants. Retrieval and evidence selection remain key failure points for the model, contributing to the overall performance drop. We further show that simple yet effective strategies, including evidence filtering and query reformulation, substantially mitigate these issues, improving performance on MedMCQA and MedXpertQA by up to 12% and 8.2%, respectively. These findings call for re-examining RAG's role in medicine and highlight the importance of stage-aware evaluation and deliberate system design for reliable medical LLM applications.

EZ-Sort: Efficient Pairwise Comparison via Zero-Shot CLIP-Based Pre-Ordering and Human-in-the-Loop Sorting

Pairwise comparison is often favored over absolute rating or ordinal classification in subjective or difficult annotation tasks due to its improved reliability. However, exhaustive comparisons require a massive number of annotations (O(n^2)). Recent work has greatly reduced the annotation burden (O(n log n)) by actively sampling pairwise comparisons using a sorting algorithm. We further improve annotation efficiency by (1) roughly pre-ordering items using the Contrastive Language-Image Pre-training (CLIP) model hierarchically without training, and (2) replacing easy, obvious human comparisons with automated comparisons. The proposed EZ-Sort first produces a CLIP-based zero-shot pre-ordering, then initializes bucket-aware Elo scores, and finally runs an uncertainty-guided human-in-the-loop MergeSort. Validation was conducted using various datasets: face-age estimation (FGNET), historical image chronology (DHCI), and retinal image quality assessment (EyePACS). It showed that EZ-Sort reduced human annotation cost by 90.5% compared to exhaustive pairwise comparisons and by 19.8% compared to prior work (when n = 100), while improving or maintaining inter-rater reliability. These results demonstrate that combining CLIP-based priors with uncertainty-aware sampling yields an efficient and scalable solution for pairwise ranking.

On the Energy Distribution of the Galactic Center Excess' Sources

The Galactic Center Excess (GCE) remains one of the defining mysteries uncovered by the Fermi $\gamma$-ray Space Telescope. Although it may yet herald the discovery of annihilating dark matter, weighing against that conclusion are analyses showing the spatial structure of the emission appears more consistent with a population of dim point sources. Technical limitations have restricted prior analyses to studying the point-source hypothesis purely spatially. All spectral information that could help disentangle the GCE from the complex and uncertain astrophysical emission was discarded. We demonstrate that a neural network-aided simulation-based inference approach can overcome such limitations and thereby confront the point source explanation of the GCE with spatial and spectral data. The addition is profound: energy information drives the putative point sources to be significantly dimmer, indicating either the GCE is truly diffuse in nature or made of an exceptionally large number of sources. Quantitatively, for our best fit background model, the excess is essentially consistent with Poisson emission as predicted by dark matter. If the excess is instead due to point sources, our median prediction is ${\cal O}(10^5)$ sources in the Galactic Center, or more than 35,000 sources at 90% confidence, both significantly larger than the hundreds of sources preferred by earlier point-source analyses of the GCE.