Caltech Logo
Thomas Fuchs
Thomas J. Fuchs

Jet Propulsion Laboratory
California Institute of Technology
4800 Oak Grove Drive
MC 198-235E
Pasadena, CA 91109

Office:++1 (818) 354-3637
Admin.:++1 (818) 354-4731
Fax:++1 (818) 393-5007
I moved to Memorial Sloan Kettering / Weill Cornell.
Please visit my new website:

I am intrigued by making sense out of vast amounts of data and the best way to achieve this is machine learning. My research focuses on developing new statistical learning algorithms and on their application in computer vision, computational pathology and medicine.

I am a research technologist at NASA's Jet Propulsion Laboratory in Pasadena where I have the unique opportunity to extend my scope of research from medicine to space exploration. In addition I have a visitor appointment at the California Institute of Technology where I occasionally give guest lectures and organize the Machine Learning Seminar.

My postdoctoral research was focused on approximate Bayesian computation and its application on likelihood-free parameter estimation for computer vision and medical imaging. This work was conducted at the computational vision lab of Pietro Perona at Caltech and the computer vision group of Larry Matthies at JPL.

In 2010 I received my PhD (Dr.Sc.) from ETH Zurich for my work in the machine learning laboratory of Joachim Buhmann. During this time I also completed the Ph.D. program on System Biology and Medicine from ETH's CC-SPMD.

I received my MSc degree (Dipl.-Ing.) from TU Graz where I majored in technical mathematics with a minor in computer science. Most of the work for my master thesis on Bayesian networks was conducted at SCR in Princeton.




Research interests:
Large Scale Machine Learning
Machine Learning I am genuinely excited about the range of possibilities the field of machine learning opens up for science in general and medicine in particular. I consider myself lucky to be able to work in this diverse and exciting line of research. Especially Computational Pathology offers a rich and challenging set of research problems for which machine learning can not only provide solutions but can also open up novel and fruitful research directions. From a theoretical point of view I am particularly interested in ensemble learning, both off-line and on-line, as well as approximate Bayesian computation for likelihood free inference. The research questions I work on range from medical imaging in pathology over survival modeling to clinical decision support.
Selected papers:
Working papers:
  • Thomas J. Fuchs and Christian L. Mueller, An approximate Bayesian computation view of parameter estimation in computer vision, Journal Article, 2013
Computational Pathology

Computational Pathology We are in the midst of a data-driven revolution in the biomedical sciences. New imaging modalities, deep sequencing and mass-spectrometry lead to an explosion of data in molecular biology, medical research and clinical practice which is not graspable for a single human expert any more. Machine learning gives us the capabilities to gain knowledge from these vast amounts of data and it hands us the tools to tackle a multitude of research questions.
The field of pathology is particularly exciting for statistical learning: (i) it is the pivotal point in the clinical workflow to enable translational and personalized medicine, (ii) the petabytes of data produced by just histopathology alone exceeds radiology 200 fold, and (iii) even in the 21st century it is vastly a manual and qualitative science and hence pathology is on the brink of dramatic change.
I am interested in the question how we can train ensemble classifiers based on low-level cues from terabytes of whole slide image data and how this information is utilized in a generative top-down model to understand tissue morphology and its impact on the survival of cancer patients. In particular I focus on novel decision forest models for image understanding and a new sampling technique for approximate Bayesian computation to estimate the parameters of spatial point processes.
These recent breakthroughs in large scale machine learning provide the enabling techniques for applications in Computational Pathology ranging from protein quantification over survival analysis to clinical decision support.
Selected papers:
Working papers:
  • Thomas J. Fuchs and Johannes Haybaeck, Approximate Bayesian computation for inferring spline based point process models of the hippocampus in temporal lobe epilepsy patients, Journal Article, 2013
  • Sara Abbasabadi, Thomas J. Fuchs, Monika Bieri, Norbert Wey, Daniela Mihic-Probst, and Joachim M. Buhmann, Automated detection and classification of melanoma micrometastases and nodal nevi on immunohistochemically stained sentinel lymph nodes, Journal Article, 2013
Biomedical Statistics
Biomedical Statistics
Selected papers:
Working papers:
  • Jay W. Shin, Thomas J. Fuchs, Weiniu Gan, Rainer Kunstfeld, Joachim Buhmann, Ulrich Mrowietz, Sascha Gerdes, and Michael Detmar, Quantification of vascular lineage-specific differentiation and in vivo (lymph)angiogenesis by a novel low-density microvascular differentiation, Journal Article, 2013
Space Exploration
Space Exploration
Left: MIB-1 stained nuclei of hepatocytes in murine liver tissue. (0.23μm/pixel)
Right: SAR images from Venus, taken by the Magellan space probe. (75m/pixel)
Selected papers:
Working papers:
  • Thomas J. Fuchs, David R. Thompson, Brian D. Bue, Julie Castillo-Rogez, Steve A. Chien, Dero Gharibian, and Kiri L. Wagstaff, Computer vision for enhanced flyby science at small bodies, Journal Article, 2013


  • [2013] Caltech Machine Learning Seminar, more ...
  • [2012] Caltech Machine Learning Seminar, more ...
  • [2010] Computational Intelligence Lab, more ...
  • [2009] Advanced Topics in Pattern Recognition, more ...
  • [2008] Zurich Center for Imaging Science and Technology (CIMST) - Summer School
  • [2007] Zurich Center for Imaging Science and Technology (CIMST) - Summer School
  • [2007] C# Programming (Languages in Depth Series), more ...
  • [2006] Introductions to Computational Science
  • [2006] Informatik I


Computer Vision - Experiment Manager

Experiment Manager

CompPath Application Suite

TMA Annotator -- TMA Estimator -- TMA Classifier -- Node Annotator
TMA Classifier        TMA Estimator

TMA Annotator        Node Annotator

Computational Pathology Online - Web Application

CP online CP online

Metabolomics Studio - R Package

Metabolomics Studio

PRR: Prediction Relevance Ranking - R Package

PRR R packagePRR R package

Random Detection Forests

C# Library, coming soon


I also try to keep my Google Scholar profile up to date.

Peer-Reviewed Articles


  1. Thomas J. Fuchs, Computational Pathology: A Machine Learning Approach, PhD Thesis, ETH Zurich, 2010
  2. Thomas J. Fuchs, Bayesian Networks Classification in Bioinformatics and Medical Decision Support, Master Thesis, Technical University Graz, 2005