arXiv subjects

Victor Garcia

Publications and source records attributed to Victor Garcia.

6 recordsLinked to original sources

arXiv2025

Performance Assessment Strategies for Language Model Applications in Healthcare

Language models (LMs) represent an emerging paradigm within artificial intelligence, with applications throughout the medical enterprise. A comprehensive understanding of the clinical task and awareness of the variability in performance when implemented in actual clinical environments lays the foundation for the LM application assessment. Presently, a prevalent method for evaluating the performance of these generative models relies on quantitative benchmarks. Such benchmarks have limitations and may suffer from train-to-the-test overfitting, optimizing performance for a specified test set at the cost of generalizability across other tasks and data distributions. Evaluation strategies leveraging human expertise and utilizing cost-effective computational models as evaluators are gaining interest. We discuss current state-of-the-art methodologies for assessing the performance of LM applications in healthcare and medical devices.

cs.LG↗

arXiv2025

Hallucinations in medical devices

Computer methods in medical devices are frequently imperfect and are known to produce errors in clinical or diagnostic tasks. However, when deep learning and data-based approaches yield output that exhibit errors, the devices are frequently said to hallucinate. Drawing from theoretical developments and empirical studies in multiple medical device areas, we introduce a practical and universal definition that denotes hallucinations as a type of error that is plausible and can be either impactful or benign to the task at hand. The definition aims at facilitating the evaluation of medical devices that suffer from hallucinations across product areas. Using examples from imaging and non-imaging applications, we explore how the proposed definition relates to evaluation methodologies and discuss existing approaches for minimizing the prevalence of hallucinations.

eess.IV↗

arXiv2022

As easy as APC: overcoming missing data and class imbalance in time series with self-supervised learning

High levels of missing data and strong class imbalance are ubiquitous challenges that are often presented simultaneously in real-world time series data. Existing methods approach these problems separately, frequently making significant assumptions about the underlying data generation process in order to lessen the impact of missing information. In this work, we instead demonstrate how a general self-supervised training method, namely Autoregressive Predictive Coding (APC), can be leveraged to overcome both missing data and class imbalance simultaneously without strong assumptions. Specifically, on a synthetic dataset, we show that standard baselines are substantially improved upon through the use of APC, yielding the greatest gains in the combined setting of high missingness and severe class imbalance. We further apply APC on two real-world medical time-series datasets, and show that APC improves the classification performance in all settings, ultimately achieving state-of-the-art AUPRC results on the Physionet benchmark.

cs.LG↗

arXiv2018

Few-Shot Learning with Graph Neural Networks

We propose to study the problem of few-shot learning with the prism of inference on a partially observed graphical model, constructed from a collection of input images whose label can be either observed or not. By assimilating generic message-passing inference algorithms with their neural-network counterparts, we define a graph neural network architecture that generalizes several of the recently proposed few-shot learning models. Besides providing improved numerical performance, our framework is easily extended to variants of few-shot learning, such as semi-supervised or active learning, demonstrating the ability of graph-based models to operate well on 'relational' tasks.

stat.ML↗

arXiv2014

The effect of interference on the CD8+ T cell escape rates in HIV

In early HIV infection, the virus population escapes from multiple CD8+ cell responses. The later an escape mutation emerges, the slower it outgrows its competition, i. e. the escape rate is lower. This pattern could indicate that the strength of the CD8+ cell responses is waning, or that later viral escape mutants carry a larger fitness cost. In this paper, we investigate whether the pattern of decreasing escape rate could also be caused by genetic interference among different escape strains. To this end, we developed a mathematical multi-epitope model of HIV dynamics, which incorporates stochastic effects, recombination and mutation. We used cumulative linkage disequilibrium measures to quantify the amount of interference. We found that nearly synchronous, similarly strong immune responses in two-locus systems enhance the generation of genetic interference. This effect, combined with densely spaced sampling times at the beginning of infection, leads to decreasing successive escape rate estimates, even when there were no selection differences among alleles. These predictions are supported by experimental data from one HIV-infected patient. Thus, interference could explain why later escapes are slower. Considering escape mutations in isolation, neglecting their genetic linkage, conceals the underlying haplotype dynamics and can affect the estimation of the selective pressure exerted by CD8+ cells. In systems in which multiple escape mutations appear, the occurrence of interference dynamics should be assessed by measuring the linkage between different escape mutations.

q-bio.PE↗

arXiv2011

Testing an agent-based model of bacterial cell motility: How nutrient concentration affects speed distribution

We revisit a recently proposed agent-based model of active biological motion and compare its predictions with own experimental findings for the speed distribution of bacterial cells, \emph{Salmonella typhimurium}. Agents move according to a stochastic dynamics and use energy stored in an internal depot for metabolism and active motion. We discuss different assumptions of how the conversion from internal to kinetic energy $d(v)$ may depend on the actual speed, to conclude that $d_{2}v^ξ$ with either $ξ=2$ or $1<ξ<2$ are promising hypotheses. To test these, we compare the model's prediction with the speed distribution of bacteria which were obtained in media of different nutrient concentration and at different times. We find that both hypotheses are in line with the experimental observations, with $ξ$ between 1.67 and 2.0. Regarding the influence of a higher nutrient concentration, we conclude that the take-up of energy by bacterial cells is indeed increased. But this energy is not used to increase the speed, with 40$μ$m/s as the most probable value of the speed distribution, but is rather spend on metabolism and growth.

q-bio.CB↗