Papers for

clinical data teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Robust method improves speech based Alzheimer's detection accuracy

Robust Rank Aggregation for Multimodal Speech-Based Alzheimer's Disease Detection

Abstract: Speech-based Alzheimer's disease (AD) detection has recently benefited from multimodal foundation-model representations that integrate complementary acoustic and linguistic information. However, conventional probability averaging over these complementary classifiers is unreliable, because their posterior probabilities exhibit mismatched scales: identical values may reflect different confidence levels across models. We propose a robust rank aggregation framework that aggregates normalized prediction ranks instead of posterior probabilities. Each subject is scored by its percentile within a fixed training-cohort distribution of out-of-fold predictions; since rank ordering is invariant to monotonic transformations, this avoids probability-scale mismatch while preserving classifier confidence ordering. A confidence-gated Random Forest further corrects residual errors using clinically interpretable linguistic features, overriding the rank prediction only when the two disagree and the RF is highly confident, without additional deep model training or explicit posterior-probability calibration. On ADReSS2020 and ADReSSo2021, the method achieves accuracies of 95.83% and 90.14%, respectively, comparing favorably with previously reported results.

Wed 9 SeptSound

The gist

Detecting Alzheimer's disease through speech can use both sounds and words people say. The authors noticed that simply averaging probabilities from different models for this task can be misleading because each model's confidence is measured on different scales. They created a method that ranks predictions instead of using their raw scores, making the combined result more reliable. Their approach also uses a decision tree that checks when language-based features strongly disagree with ranks, improving accuracy without extra complex training. This method showed high accuracy on two standard Alzheimer’s speech datasets.

Open → 2609.09948v1

Tefm reduces tokens and improves trust in structured data models

TEFM: Token-Efficient Faithful Modeling for Structured Data

Abstract: In this paper, we solve two fundamental obstacles in applying LLMs to critical domains: token efficiency and faithfulness. To address both constraints jointly, we present TEFM (Token-Efficient Faithful Modeling), a framework designed for structured data analysis in critical domains. TEFM achieves token efficiency by compressing lengthy structured observations into compact Behavioral Code tokens, dramatically reducing token consumption with minimal information loss. Moreover, TEFM enables faithful rationalization through a dual-fidelity objective that jointly optimizes code-level reconstruction and prediction-level fidelity, identifying minimal sufficient feature subsets grounded in input data. Comprehensive experiments across various domain datasets and model backbones (Qwen3, Gemma-2, Phi-4) show that TEFM achieves competitive classification accuracy with dramatic token reduction (approximately 1\% token retention in clinical and 2\% in security domains) while producing faithful rationales.

Wed 9 SeptComputation and LanguageMachine Learning

The gist

Using large language models with complex data can be expensive and hard to trust. The authors developed TEFM, which shrinks data into small, meaningful codes so models use far fewer tokens without losing much information. Their approach also helps explain predictions better by focusing on the most important features. Tests show TEFM keeps accuracy high while using only a tiny fraction of tokens, especially in health and security data.

Open → 2609.09552v1

Uncertainty matters when checking fairness of machine learning models

The Role of Uncertainty in Assessing the Fairness of Machine Learning Models

Abstract: Machine learning models are widely used in clinical applications, social media, law enforcement and critical infrastructure. Verifying whether their outputs are biased against disadvantaged groups or individuals is crucial to ensuring they are fair and allowing their use in such settings. A rigorous risk assessment of possible fairness violations requires quantifying the uncertainty associated with selecting and estimating such models. Yet, this is rarely done in the literature, which focuses on identifying a single model with a suitable trade-off between predictive accuracy and fairness. In this paper, we move beyond point estimation and discuss frequentist and Bayesian approaches to uncertainty quantification for fair machine learning, with practical examples and implications for simulated and real data.

Mon 7 SeptMachine Learning

The gist

Machine learning models are used in important areas like healthcare and law enforcement, where fairness is crucial. The paper explains that checking if these models are fair often ignores how uncertain we are about the model choices themselves. The authors discuss methods to measure this uncertainty using statistical and Bayesian approaches. This helps give a clearer picture of how confident we can be that models treat people fairly, based on both simulated and real data examples.

Open → 2609.07959v1