Suppressing Prior-Comparison Hallucinations in Radiology Report Generation via Semantically Decoupled Latent Steering

• URL: http://arxiv.org/abs/2602.23676v1
• Date: Fri, 27 Feb 2026 04:49:01 GMT
• Title: Suppressing Prior-Comparison Hallucinations in Radiology Report Generation via Semantically Decoupled Latent Steering
• Authors: Ao Li, Rui Liu, Mingjie Li, Sheng Liu, Lei Wang, Xiaodan Liang, Lina Yao, Xiaojun Chang, Lei Xing,
• Abstract summary: We develop a training-free, inference-time control framework termed Semantically Decoupled Latent Steering.<n>Our approach constructs a semantic-free intervention vector via large language model (LLM)-driven semantic decomposition.<n>We show that our approach significantly reduces the probability of historical hallucinations.
• Score: 94.37535002230504
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Automated radiology report generation using vision-language models (VLMs) is limited by the risk of prior-comparison hallucination, where the model generates historical findings unsupported by the current study. We address this challenge with a training-free, inference-time control framework termed Semantically Decoupled Latent Steering (SDLS). Unlike generic activation steering, which often suffers from semantic entanglement, our approach constructs a semantic-free intervention vector via large language model (LLM)-driven semantic decomposition followed by $QR$-based orthogonalization. This orthogonalization step is critical. It leverages geometric constraints to filter out the clinical semantics often entangled in standard principal component analysis (PCA) directions, ensuring that the steering vector targets only the ``historical comparison" axis. We validate our method on the BiomedGPT foundation model, demonstrating that it overcomes the trade-off between hallucination suppression and clinical accuracy. Extensive experiments on MIMIC-CXR, and zero-shot transfer evaluation on CheXpert Plus and IU-Xray, demonstrate the robustness of our approach. Quantitative evaluations on MIMIC-CXR show that our approach significantly reduces the probability of historical hallucinations (FilBERT score decreases from 0.2373 to 0.1889) and improves clinical label fidelity (CheXpert macro-F1 increases from 0.2242 to 0.3208). Supplementary evaluations confirm that the structural integrity of the clinical narrative is maintained.

Related papers

• Cross-Linguistic Persona-Driven Data Synthesis for Robust Multimodal Cognitive Decline Detection [20.599682298329213]
  We introduce SynCog, a novel framework integrating controllable zero-shot multimodal data synthesis with Chain-of-Thought deduction fine-tuning.<n>This generative paradigm enables the rapid, zero-shot expansion of clinical corpora across diverse languages.<n>Experiments on the ADReSS and ADReSSo benchmarks demonstrate that augmenting limited clinical data with synthetic phenotypes yields competitive diagnostic performance.
  arXiv  Detail & Related papers  (2026-02-08T14:10:05Z)
• SurvKAN: A Fully Parametric Survival Model Based on Kolmogorov-Arnold Networks [7.352227733654751]
  We introduce SurvKAN, a fully parametric, time-continuous survival model based on Kolmogorov-Arnold Networks (KANs)<n>SurvKAN treats time as an explicit input to a KAN that directly predicts the log-hazard function, enabling end-to-end training on the full survival likelihood.
  arXiv  Detail & Related papers  (2026-02-02T14:49:14Z)
• ClinNet: Evidential Ordinal Regression with Bilateral Asymmetry and Prototype Memory for Knee Osteoarthritis Grading [3.337151338735509]
  Knee osteoarthritis (KOA) grading based on radiographic images is a critical yet challenging task.<n>In this work, we propose ClinNet, a novel trustworthy framework that addresses KOA grading as evidential ordinal regression problem.
  arXiv  Detail & Related papers  (2026-01-24T05:49:41Z)
• Conformal Lesion Segmentation for 3D Medical Images [82.92159832699583]
  We propose a risk-constrained framework that calibrates data-driven thresholds via conformalization to ensure the test-time FNR remains below a target tolerance.<n>We validate the statistical soundness and predictive performance of CLS on six 3D-LS datasets across five backbone models, and conclude with actionable insights for deploying risk-aware segmentation in clinical practice.
  arXiv  Detail & Related papers  (2025-10-19T08:21:00Z)
• Measuring Sample Quality with Copula Discrepancies [0.0]
  Copula Discrepancy (CD) is a principled and computationally efficient diagnostic for dependence structure.<n>Our theoretical framework provides the first structure-aware diagnostic specifically designed for the era of approximate inference.<n>With computational overhead orders of magnitude lower than existing Stein discrepancies, the CD provides both immediate practical value for MCMC practitioners and a theoretical foundation for the next generation of structure-aware sample quality assessment.
  arXiv  Detail & Related papers  (2025-07-29T02:11:45Z)
• Diffusion with a Linguistic Compass: Steering the Generation of Clinically Plausible Future sMRI Representations for Early MCI Conversion Prediction [13.937881108738042]
  We propose a diffusion-based framework that synthesizes clinically plausible future sMRI representations directly from baseline data.<n>Experiments on ADNI and AIBL cohorts show that MCI-Diff outperforms state-of-the-art baselines.
  arXiv  Detail & Related papers  (2025-06-05T07:01:05Z)
• CRTRE: Causal Rule Generation with Target Trial Emulation Framework [47.2836994469923]
  We introduce a novel method called causal rule generation with target trial emulation framework (CRTRE) CRTRE applies randomize trial design principles to estimate the causal effect of association rules. We then incorporate such association rules for the downstream applications such as prediction of disease onsets.
  arXiv  Detail & Related papers  (2024-11-10T02:40:06Z)
• Improving Multiple Sclerosis Lesion Segmentation Across Clinical Sites: A Federated Learning Approach with Noise-Resilient Training [75.40980802817349]
  Deep learning models have shown promise for automatically segmenting MS lesions, but the scarcity of accurately annotated data hinders progress in this area. We introduce a Decoupled Hard Label Correction (DHLC) strategy that considers the imbalanced distribution and fuzzy boundaries of MS lesions. We also introduce a Centrally Enhanced Label Correction (CELC) strategy, which leverages the aggregated central model as a correction teacher for all sites.
  arXiv  Detail & Related papers  (2023-08-31T00:36:10Z)
• Learning to diagnose cirrhosis from radiological and histological labels with joint self and weakly-supervised pretraining strategies [62.840338941861134]
  We propose to leverage transfer learning from large datasets annotated by radiologists, to predict the histological score available on a small annex dataset. We compare different pretraining methods, namely weakly-supervised and self-supervised ones, to improve the prediction of the cirrhosis. This method outperforms the baseline classification of the METAVIR score, reaching an AUC of 0.84 and a balanced accuracy of 0.75.
  arXiv  Detail & Related papers  (2023-02-16T17:06:23Z)
• Bayesian prognostic covariate adjustment [59.75318183140857]
  Historical data about disease outcomes can be integrated into the analysis of clinical trials in many ways. We build on existing literature that uses prognostic scores from a predictive model to increase the efficiency of treatment effect estimates.
  arXiv  Detail & Related papers  (2020-12-24T05:19:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.