The first systematization of reconstruction attacks on synthetic tabular data finds that generator choice dominates privacy risk over attack choice, with differential privacy effective only at low budgets and most leakage reflecting population structure rather than memorization.
Title resolution pending
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
years
2026 6representative citing papers
IMCBench is a new benchmark for image-grounded multi-turn medical conversations that evaluates eight multimodal LLMs on safety, accuracy, and uncertainty, finding Claude Opus highest overall but safety drops for malignant and rare conditions.
Clinical narrative format beats raw JSON for LLMs up to 8B parameters on medication reconciliation but raw JSON wins at 70B scale, with omissions as the main error type.
A pipeline generates 1,732 valid FHIR bundles from clinician cases, revealing lower LLM diagnostic accuracy on structured inputs than on plain text.
Synthetic data generation exhibits disparate impact from group-specific approximation, sampling, and estimation errors; group-wise models improve both utility and parity on graphical model methods.
MedGemma 1.5 4B reports absolute gains of 11% on 3D MRI classification, 3% on 3D CT, 47% macro F1 on pathology slides, 35% IoU on anatomical localization, and 5-22% on clinical QA tasks over MedGemma 1.
citing papers explorer
-
SoK: Reconstruction Attacks on Synthetic Tabular Data (Insights from Winning the NIST CRC)
The first systematization of reconstruction attacks on synthetic tabular data finds that generator choice dominates privacy risk over attack choice, with differential privacy effective only at low budgets and most leakage reflecting population structure rather than memorization.
-
IMCBench: A benchmark for multimodal LLMs in Image-grounded Medical Conversations
IMCBench is a new benchmark for image-grounded multi-turn medical conversations that evaluates eight multimodal LLMs on safety, accuracy, and uncertainty, finding Claude Opus highest overall but safety drops for malignant and rare conditions.
-
MedCase-Structured: A Text-to-FHIR Dataset for Benchmarking Diagnostic Reasoning in Clinically Realistic EHR Settings
A pipeline generates 1,732 valid FHIR bundles from clinician cases, revealing lower LLM diagnostic accuracy on structured inputs than on plain text.
-
Disparate Impact in Synthetic Data Generation
Synthetic data generation exhibits disparate impact from group-specific approximation, sampling, and estimation errors; group-wise models improve both utility and parity on graphical model methods.
-
MedGemma 1.5 Technical Report
MedGemma 1.5 4B reports absolute gains of 11% on 3D MRI classification, 3% on 3D CT, 47% macro F1 on pathology slides, 35% IoU on anatomical localization, and 5-22% on clinical QA tasks over MedGemma 1.