MM
M. Mahmoudi
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
2 records found
1
Predictive computational oncology models are fundamentally limited by their uni-modal input drug representations. To overcome this bottleneck, we developed DrugZip, a uniform, task-agnostic, 128-dimensional representation that compresses 25 diverse modalities from the Chemical Checker across a context of 1.2 million molecules. By using a modified autoencoder, DrugZip successfully stabilises the latent space and avoids posterior collapse from a standard variational autoencoder. We evaluated DrugZip across three downstream tasks. In drug synergy prediction, it achieved an AUC of 0.844, resisting performance collapse in unseen cell environments with a mean AUC of 0.62. In drug sensitivity prediction, DrugZip bypassed the extreme overfitting of high-dimensional baselines on unseen drugs. Finally, in cellular perturbation modelling via ChemCPA, DrugZip demonstrated representational sufficiency by matching state-of-the-art transcriptomic prediction accuracy ($R^2$ of 0.776 vs 0.792). Geometrical and information-content analyses confirm that DrugZip produces a continuous, balanced embedding space where drugs remain individually distinguishable. Ultimately, DrugZip shifts the paradigm from engineering task-specific features toward utilising a robust, generalizable, multi-modal representation for computational oncology.
...
Predictive computational oncology models are fundamentally limited by their uni-modal input drug representations. To overcome this bottleneck, we developed DrugZip, a uniform, task-agnostic, 128-dimensional representation that compresses 25 diverse modalities from the Chemical Checker across a context of 1.2 million molecules. By using a modified autoencoder, DrugZip successfully stabilises the latent space and avoids posterior collapse from a standard variational autoencoder. We evaluated DrugZip across three downstream tasks. In drug synergy prediction, it achieved an AUC of 0.844, resisting performance collapse in unseen cell environments with a mean AUC of 0.62. In drug sensitivity prediction, DrugZip bypassed the extreme overfitting of high-dimensional baselines on unseen drugs. Finally, in cellular perturbation modelling via ChemCPA, DrugZip demonstrated representational sufficiency by matching state-of-the-art transcriptomic prediction accuracy ($R^2$ of 0.776 vs 0.792). Geometrical and information-content analyses confirm that DrugZip produces a continuous, balanced embedding space where drugs remain individually distinguishable. Ultimately, DrugZip shifts the paradigm from engineering task-specific features toward utilising a robust, generalizable, multi-modal representation for computational oncology.
Recognizing facial emotions is key for social interaction, yet the subjective nature of emotion labeling poses challenges for automatic facial affect prediction. Variability in how individuals interpret emotions leads to uncertainty in training data for machine learning models. While multiple raters and interrater agreement (IRA) measures are used to address this, the extent of their use and their impact on dataset reliability is not well understood. This systematic literature review investigates the methodologies used to measure IRA in facial affect recognition datasets. Concrete eligibility and feasibility criteria were applied, and it resulted in 47 papers being retrieved from Scopus, Web of Science, IEEExplore, and ACM Digital Library. Data on affect states, affect representation schemes (ARS), and IRA methodologies used by the datasets and their corresponding papers were extracted to provide a comprehensive overview and allow a detailed analysis. Clear correlation was not found in between ARS and IRA, but the retrieved data showed that Fleiss' kappa was the most popular methodology over time but also in the recent years.
...
Recognizing facial emotions is key for social interaction, yet the subjective nature of emotion labeling poses challenges for automatic facial affect prediction. Variability in how individuals interpret emotions leads to uncertainty in training data for machine learning models. While multiple raters and interrater agreement (IRA) measures are used to address this, the extent of their use and their impact on dataset reliability is not well understood. This systematic literature review investigates the methodologies used to measure IRA in facial affect recognition datasets. Concrete eligibility and feasibility criteria were applied, and it resulted in 47 papers being retrieved from Scopus, Web of Science, IEEExplore, and ACM Digital Library. Data on affect states, affect representation schemes (ARS), and IRA methodologies used by the datasets and their corresponding papers were extracted to provide a comprehensive overview and allow a detailed analysis. Clear correlation was not found in between ARS and IRA, but the retrieved data showed that Fleiss' kappa was the most popular methodology over time but also in the recent years.