Malek Kamoun
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Malek Kamoun has a total of 69 co-authors. Here are the top co-authors by number of shared publications:
- Harsh Bandhey
Cedars-Sinai Health Sciences University, Los Angeles, USA
5 shared publications - Ryan J. Urbanowicz
Cedars Sinai Medical Center, West Hollywood, USA
5 shared publications - Gabriel Lipschutz-Villa
University of Pennsylvania, Philadelphia, USA
3 shared publications - Eline T. Luning Prak
University of Pennsylvania
2 shared publications - Khoi Dinh
University of Pennsylvania, Philadelphia, PA, USA
2 shared publications A. Ben Gara
Department of Animal Production, School of Higher Education in Agriculture of Mateur, Tunisia
1 shared publication
Latest Publications
- Open Access
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1175-1183
Audience-Customized Translation of Rule-Based Evidence with Large Language Models Across Multiplexer Benchmarks
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1175-1183
Open AccessAudience-Customized Translation of Rule-Based Evidence with Large Language Models Across Multiplexer Benchmarks
- Open Access
- 10 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 300-309
Phase-Alternation and Tree-Initialization to Facilitate Interpretable Rule-based Machine Learning
Interpretable machine learning is important for scientific, biomedical, and other high-stakes domains where transparency, trust, and human understanding of models are required. Recently, an evolutionary rule-based machine learning algorithm called HEROS was proposed, which separated rule from rule-set discovery each with distinct multi-objective optimization strategies yielding a two-phase, noise agnostic framework that found accurate, highly compact, and interpretable solutions on a diverse set of challenging benchmarks without requiring hyperparameter optimizations. However, HEROS is currently limited by its strictly sequential two-phase design which requires committing in advance to a fixed number of rule-discovery iterations. This work extends HEROS with a phase alternation scheme interleaving rule discovery and rule-set optimization. This enables refinement of both tasks throughout training in an 'anytime-algorithm' manner. We also introduce a tree-based rule initialization strategy to accelerate early-stage rule discovery. These extensions aim to preserve the empirical performance of the original HEROS framework under a similar computational budget while providing greater flexibility for algorithm application and problem scalability. Using the same diverse set of benchmark datasets, we evaluate and compare these HEROS extensions to the original HEROS algorithm and established rule-based learning systems including RIPPER and BioHEL. We demonstrate the advantages of this extended HEROS framework.
- 10 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 300-309
Open AccessPhase-Alternation and Tree-Initialization to Facilitate Interpretable Rule-based Machine Learning
Interpretable machine learning is important for scientific, biomedical, and other high-stakes domains where transparency, trust, and human understanding of models are required. Recently, an evolutionary rule-based machine learning algorithm called HEROS was proposed, which separated rule from rule-set discovery each with distinct multi-objective optimization strategies yielding a two-phase, noise agnostic framework that found accurate, highly compact, and interpretable solutions on a diverse set of challenging benchmarks without requiring hyperparameter optimizations. However, HEROS is currently limited by its strictly sequential two-phase design which requires committing in advance to a fixed number of rule-discovery iterations. This work extends HEROS with a phase alternation scheme interleaving rule discovery and rule-set optimization. This enables refinement of both tasks throughout training in an 'anytime-algorithm' manner. We also introduce a tree-based rule initialization strategy to accelerate early-stage rule discovery. These extensions aim to preserve the empirical performance of the original HEROS framework under a similar computational budget while providing greater flexibility for algorithm application and problem scalability. Using the same diverse set of benchmark datasets, we evaluate and compare these HEROS extensions to the original HEROS algorithm and established rule-based learning systems including RIPPER and BioHEL. We demonstrate the advantages of this extended HEROS framework.
- Open Access
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 407-415
Rule-based Machine Learning: Separating Rule and Rule-Set Pareto-Optimization for Interpretable Noise-Agnostic Modeling
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 407-415
Open AccessRule-based Machine Learning: Separating Rule and Rule-Set Pareto-Optimization for Interpretable Noise-Agnostic Modeling
- 21 March 2024
- Published by Springer Nature
- p. 225-239
Evolutionary Feature-Binning with Adaptive Burden Thresholding for Biomedical Risk Stratification
- 21 March 2024
- Published by Springer Nature
- p. 225-239
Evolutionary Feature-Binning with Adaptive Burden Thresholding for Biomedical Risk Stratification
- Open Access
- 15 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 1846-1854
Scikit-FIBERS: An 'OR'-Rule Discovery Evolutionary Algorithm for Risk Stratification in Right-Censored Survival Analyses
In 'time-to-event' problems, such as survival analysis in biomedical research, investigators seek to identify the factors that impact/predict not only whether an event occurred (e.g. death), but also when. Further, in the domain of kidney transplantation, there has been recent interest in determining whether higher resolution features, i.e. amino-acid mismatches (AA-MMs) between donors and recipients, can improve risk stratification for transplant failure and identify subsets of AA-MM positions for evaluating risk in future transplants. Motivated by these problems, the FIBERS algorithm was previously developed, which applies a genetic algorithm to learn a population of feature bins (i.e. OR-rules) for predicting right-censored time-to-event survival outcomes. Here, instances are categorized as high-risk if an AA-MM is present for any AA-MM position specified in a given rule, and low-risk if not. In the present study we focus further on the FIBERS algorithm; (1) implementing it as an accessible scikit-learn compatible machine learning package, and (2) applying a variety of simulated right-censored survival datasets to validate scikit-FIBERS efficacy and examine the limitations of its performance to guide ongoing development. We present preliminary results demonstrating the efficacy of scikit-FIBERS, and the limitations of both the algorithm and survival data simulator to guide future work.
- 15 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 1846-1854
Open AccessScikit-FIBERS: An 'OR'-Rule Discovery Evolutionary Algorithm for Risk Stratification in Right-Censored Survival Analyses
In 'time-to-event' problems, such as survival analysis in biomedical research, investigators seek to identify the factors that impact/predict not only whether an event occurred (e.g. death), but also when. Further, in the domain of kidney transplantation, there has been recent interest in determining whether higher resolution features, i.e. amino-acid mismatches (AA-MMs) between donors and recipients, can improve risk stratification for transplant failure and identify subsets of AA-MM positions for evaluating risk in future transplants. Motivated by these problems, the FIBERS algorithm was previously developed, which applies a genetic algorithm to learn a population of feature bins (i.e. OR-rules) for predicting right-censored time-to-event survival outcomes. Here, instances are categorized as high-risk if an AA-MM is present for any AA-MM position specified in a given rule, and low-risk if not. In the present study we focus further on the FIBERS algorithm; (1) implementing it as an accessible scikit-learn compatible machine learning package, and (2) applying a variety of simulated right-censored survival datasets to validate scikit-FIBERS efficacy and examine the limitations of its performance to guide ongoing development. We present preliminary results demonstrating the efficacy of scikit-FIBERS, and the limitations of both the algorithm and survival data simulator to guide future work.
- 23 February 2023
- Published by Elsevier in Human Immunology
- Vol. 84 (4) , 278-285
COVID-19 infection and vaccination rarely impact HLA antibody profile in waitlisted renal transplant candidates- a multicenter cohort
- 23 February 2023
- Published by Elsevier in Human Immunology
- Vol. 84 (4) , 278-285
COVID-19 infection and vaccination rarely impact HLA antibody profile in waitlisted renal transplant candidates- a multicenter cohort
- Open Access
- 1 December 2013
- Published by Science Alert in Pakistan Journal of Biological Sciences
- Vol. 16 (24) , 2031-2035
Kappa Casein Gene Polymorphism in Local Tunisian Goats
Kappa Casein Gene Polymorphism in Local Tunisian Goats
- 1 December 2013
- Published by Science Alert in Pakistan Journal of Biological Sciences
- Vol. 16 (24) , 2031-2035
Open AccessKappa Casein Gene Polymorphism in Local Tunisian Goats
Kappa Casein Gene Polymorphism in Local Tunisian Goats
- Open Access
- 16 March 2013
- Published by Springer Nature in Journal of Cardiothoracic Surgery
- Vol. 8 (1) , 48
Profound hyperacute cardiac allograft rejection rescue with biventricular mechanical circulatory support and plasmapheresis, intravenous immunoglobulin, and rituximab therapy
Hyperacute rejection is a rare but potentially catastrophic complication after cardiac transplantation. We describe an unusual case of hyperacute rejection due to preformed anti-donor antibodies despite a negative preoperative panel-reactive antibody (PRA) screen. An excellent outcome was achieved in this case and our strategy involving the use of CentriMag ventricular assist devices (VADs) for biventricular support during treatment with rituximab, intravenous immunoglobulin (IVIG), and plasmapheresis is illustrated.
- 16 March 2013
- Published by Springer Nature in Journal of Cardiothoracic Surgery
- Vol. 8 (1) , 48
Open AccessProfound hyperacute cardiac allograft rejection rescue with biventricular mechanical circulatory support and plasmapheresis, intravenous immunoglobulin, and rituximab therapy
Hyperacute rejection is a rare but potentially catastrophic complication after cardiac transplantation. We describe an unusual case of hyperacute rejection due to preformed anti-donor antibodies despite a negative preoperative panel-reactive antibody (PRA) screen. An excellent outcome was achieved in this case and our strategy involving the use of CentriMag ventricular assist devices (VADs) for biventricular support during treatment with rituximab, intravenous immunoglobulin (IVIG), and plasmapheresis is illustrated.
- Open Access
- 2 May 2009
- Published by Oxford University Press (OUP) in American Journal of Clinical Pathology
- Vol. 131 (5) , 618-622
2008 ACLPS Panel Discussion on Resident Education in Clinical Pathology
A curriculum in clinical pathology (CP) was developed under the auspices of the Academy of Clinical Laboratory Physicians and Scientists (ACLPS) in 2
- 2 May 2009
- Published by Oxford University Press (OUP) in American Journal of Clinical Pathology
- Vol. 131 (5) , 618-622
Open Access2008 ACLPS Panel Discussion on Resident Education in Clinical Pathology
A curriculum in clinical pathology (CP) was developed under the auspices of the Academy of Clinical Laboratory Physicians and Scientists (ACLPS) in 2
- 1 September 2002
- Published by Wiley in Tissue Antigens
- Vol. 60 (3) , 262-265
A nonsense mutation in Exon 3 results in the HLA‐B null allele B*5127N
A new HLA‐B null allele has been identified within the B*51 group by combined serological and molecular typing of an Italian Caucasoid family. Serological data indicated that the proband typed homozygous for A2 and B60. Confirmatory typing using sequence specific oligonucleotide hybridization (SSPOH) detected a second B allele within the B*51 group. Allele specific typing (SSP) for B*51 subtypes, including the known B*5111N allele, was performed, and typing results were consistent with B*5101, suggesting the presence of a new null variant. Cloning and sequencing of this allele identified a B*5101 variant with a nonsense mutation in exon 3. This new null allele has been designated B*5127N. The combined use of serologic and DNA‐based typing methods facilitates the identification of null and low‐expression alleles. An overview of null alleles of class I HLA is presented.
- 1 September 2002
- Published by Wiley in Tissue Antigens
- Vol. 60 (3) , 262-265
A nonsense mutation in Exon 3 results in the HLA‐B null allele B*5127N
A new HLA‐B null allele has been identified within the B*51 group by combined serological and molecular typing of an Italian Caucasoid family. Serological data indicated that the proband typed homozygous for A2 and B60. Confirmatory typing using sequence specific oligonucleotide hybridization (SSPOH) detected a second B allele within the B*51 group. Allele specific typing (SSP) for B*51 subtypes, including the known B*5111N allele, was performed, and typing results were consistent with B*5101, suggesting the presence of a new null variant. Cloning and sequencing of this allele identified a B*5101 variant with a nonsense mutation in exon 3. This new null allele has been designated B*5127N. The combined use of serologic and DNA‐based typing methods facilitates the identification of null and low‐expression alleles. An overview of null alleles of class I HLA is presented.