Konstantinos (Kostas) Sechidis

Advanced Methodology and Data Science Group, Novartis
Honorary Research Fellow in Machine Learning and Robotics, University of Manchester
LinkedIn Twitter Github Email ORCID Google Scholar

Hello, my name is Konstantinos (for short, Kostas) Sechidis and I am a machine learning researcher with experience in developing, enhancing and delivering novel statistical and machine learning methods tailored to healthcare analytics. I work in the Advanced Methodology and Data Science group of Novartis and specialize in exploratory analysis for biomarker discovery, digital biomarker development, and assessing treatment effect heterogeneity in clinical trials. I did my PhD in Machine Learning in the University of Manchester and subsequently I worked as a postdoctoral researcher with AstraZeneca and Roche. I am honorary research fellow in Machine Learning and Robotics in the University of Manchester, while I am also a member of the editorial board of Machine Learning Journal (MLJ) and the chair of the technical committee on Statistical Pattern Recognition Techniques of the International Association for Pattern Recognition (IAPR).

Disclaimer: this is my personal page, the content is my own responsibility and it is not connected to/supported by any entity with which I have been, am now, or will be affiliated.

News

April 2025: Our latest paper on arXiv illustrates the implementations of WATCH in the context of parametric regression models, highlighting the practical and flexible use of score-residual based methods for identifying treatment effect modifiers. A big thanks to my Novartis co-authors (Cong Zhang, Sophie Sun, Yao Chen, Björn Bornkamp) and Torsten Hothorn for a great collobartion.

March 2025: I am happy to present our recent work on assessing treatment effect heterogeneity in clinical trials at various events, including the AI in Healthcare seminars at the University of Manchester, DAGstat 2025, and the ISBS webinar series.

February 2025: In our new paper Using Individualized Treatment Effects to Assess Treatment Effect Heterogeneity we introduce methodologies for assessing heterogeneity using individualized treatment effects as a basis. These methods can be used to implement our recently proposed WATCH workflow for binary and continuous endpoints. A big thanks to my Novartis co-authors (Cong Zhang, Sophie Sun, Yao Chen, Björn Bornkamp) and Asher Spector for a great collobartion.

January 2025: Laura’s work on using low-precision conditional mutual information (CMI) for feature selection is published in Pattern Recognition.

November 2024: I am pleased to announce my appointment as the chair of the technical committee on Statistical Pattern Recognition Techniques of (IAPR). Together with Maura Pintor, I look forward to promoting interaction and collaboration among researchers engaged in statistical pattern recognition and machine learning.

September 2024: I attended the 2024 EFSPI regulatory statistics workshop to present a short topic on the quality standards of exploratory analysis. My presentation and the discussion with the panel of regulators are in youtube. I also presented two posters describing our work with the Subgroup Analysis SIG and the Biomarkers SIG, of which I am a member.

August 2024: Frank Bretz and I organized an online webinar with the Basel Biometric Society (BBS) on “Controlling the chances of false discoveries in exploratory analysis of clinical trials”. The workshop took place 29th of August and you can find the material here

June 2024: I presented WATCH in the PSI (Statisticians in the Pharmaceutical Industry) 2024 conference.

May 2024: In our new paper we introduce the WATCH: A Workflow to Assess Treatment Effect Heterogeneity in Drug Development for Clinical Trial Sponsors. This work goes beyond pure data analysis considerations and approaches the problem of assessing heterogeneity comprehensively, encompassing all critical steps from the initial problem definition, through data processing and analysis, to the incorporation of external evidence and best scientific knowledge, and the final communication of the findings. UPDATE: The paper is now published in Pharmaceutical Statistics.

March 2024: A new paper, All that Glitters Is not Gold: Type-I Error Controlled Variable Selection from Clinical Trial Data, published in Clinical Pharmacology and Therapeutics (CPT). Furthermore, an R package that implements the methods described is available in GitHub: knockofftools package.

April 2024: Together with Mark Baillie, Frank Bretz and Prashanti Goswami we organise the Data science thinking: making an impact workshop in AMLD 2024.

April 2024: There is a PhD opportunity by Gavin Brown in the UKRI AI Center for Doctoral Training (CDT) in Decision Making for Complex Systems (jointly run between the University of Manchester and the University of Cambridge). The project will focus on the areas of causality and information geometry, with the main objective to study the statistical properties of various causal effect measures and understand/reduce their inherent uncertainty. Frank Bretz and I will also co-supervise/advise the student.

Nov 2023: Gave an invited presentation in the European Statistical Forum. This year’s topic was: Statistical Methodology in Precision Medicine and the role of Artificial Intelligence.

Sep 2023: Present our work on controlled biomarker discovery in the 5th Conference of the Central European Network (CEN).

Jan 2023: I have been appointed vice-chair of the technical committee on Statistical Pattern Recognition Techniques of the International Association for Pattern Recognition (IAPR).

Dec 2022: Our paper on benchmarking methods for characterising treatment effect heterogeneity in clinical trials published in Biometrical Journal. If you are interested on simulating datasets of heterogeneous treatment effects you can check our benchtm package.

Sep 2021: Lasse Hansen’s work on assessing depression using speech emotion recognition systems is available in biorxiv. For those interested, Lasse’s has a twitter thread that summarizes some of the main points of the paper.

Aug 2021: On September 13th, together with colleagues from industry and academia, we organise the second edition of PharML workshop, colocated with ECML-PKDD 2021. Check the exciting program here.

Jul 2021: The paper on using knockoffs for controlled predictive biomarker identification has just been published in Statistics in Medicine. For those interested, check out an earlier work from our group (Advanced Methodology and Data Science in Novartis) where the sequential knockoffs algorithm was introduced.

Apr 2021: The paper of my work in Roche is published and open access in Artificial Intelligence in Medicine.

Oct 2020: I am now member of the editorial board of the Machine Learning Journal (MLJ). I also joined the Treatment Effect Heterogeneity special interest group, which is sponsored by the European Federation of Statisticians in the Pharmaceutical Industry (EFSPI) and the Statisticians in the Pharmaceutical Industry (PSI) organisation.

Sept 2020: On the September 18th I will present on the Statistical Learning workshop organised by the Data Mining and Machine Learning group of the University of Geneva. The aim of the workshop is to bring together the research communities of statistics and machine learning to foster a discussion between the two fields and develop research synergies. For more details and registration see here.

June 2020: I joined Novartis’ Advanced Methodology and Data Science (AMDS) group, where I focus on developing novel machine learning methods with the aim of improving drug development in multiple projects. Furthermore, in collaboration with data scientists and biostatisticians, we work to ensure that state-of-the-art statistical and machine learning methods are used at the trial and project level.

Key research interests

  • Feature selection
  • Information theory
  • Biomarker discovery for personalised healthcare
  • Digital biomarker discovery
  • Multi-target learning