Karsten Roth
One-Page Resume (pdf)   @confusezius   @karsten-roth   @Confusezius   Karsten Roth

  TL;DR

PhD Student - ELLIS & IMPRS-IS

Co-Supervised by Prof. Zeynep Akata (University of Tuebingen) and Prof. Oriol Vinyals (Deepmind). Working on multimodal representation learning | Generalization in ML | XAI.


Previous: Research Intern at MILA/Vector/AWS

Various research internships at MILA (Montreal), Vector (Toronto) and Amazon AWS (Tuebingen) working on topics ranging from Self-Supervised Learning, Deep Metric Learning, Few-Shot Learning to Medical Deep Learning and Anomaly Detection.


  News

June 2021

Preprint out on Industrial Anomaly Detection - work done at Amazon AWS. Link


May 2021

Two papers submitted to NeurIPS 2021.


May 2021

One paper accepted to ICML 2021 on Self-Distillation for Deep Metric Learning: Link


May 2021

Started my PhD as part of ELLIS & IMPRS-IS co-supervised by Zeynep Akata (University of Tuebingen) and Oriol Vinyals (Deepmind).


March 2021

One paper submitted to ICCV 2021.


March 2021

Completed a successful 5 month research internship at AWS Research Lablet Amazon in Tuebingen.


Dec. 2020

Two accepted workshop submissions at NeurIPS 2020 on continual few-shot learning for multilabel medical applications and uniform priors for Meta-learning. Links soon!


March 2021

Started a 5 month research internship at AWS Research Lablet Amazon in Tuebingen supervised by Peter Gehler and Thomas Brox.


Sept. 2020

New preprint on Knowledge Distillation in Deep Metric Learning. Arxiv: Link, Github: Link


July 2020

One paper accepted to TPAMI on Improved Generalization in Deep Metric Learning: Link


July 2020

One paper accepted to ECCV 2020 on Multi-Feature Deep Metric Learning: Link


June 2020

One paper accepted to International Journal for Numerical Methods in Biomedical Engineering on diffuse domain method for needle insertion simulations. Link


June 2020

One paper accepted to ICML 2020 on comparability and generalization in Deep Metric Learning. Arxiv: Link, Github: Link


May 2020

Started a four month research internship at the Vector Institute under supervision of Marzyeh Ghassemi.


Apr. 2020

Selected as participant to attend the MLSS 2020 in Tuebingen (p < 180/1300)


Mar. 2020

One paper submitted to ECCV 2020 on Multi-Feature Deep Metric Learning.


Feb. 2020

One paper accepted to CVPR 2020 on Deep Metric Learning and Reinforcement Learning. Arxiv: Link, Github: Link


Feb. 2020

One paper submitted to ICML 2020 examining driving factors for generalization in Deep Metric Learning.


Feb. 2020

Selected as reviewer for MICCAI 2020.


Jan. 2020

Selected as challenge reviewer for MIDL 2020.


Jan. 2020

One paper accepted to ISBI 2020 on Mask Mining for Liver Lesion Segmentation. Arxiv: Link


Nov. 2019

One paper submitted to CVPR 2020 on Adaptive Deep Metric Learning.


Oct. 2019

One paper submitted to TPAMI on Improved Generalization for Deep Metric Learning.


Oct. 2019

One paper accepted to the NeurIPS 2019 Med-Neurips Workshop Track on error mining for medical semantic segmentation. Arxiv: Link


Sept. 2019

Started a six month research internship at the Montreal Institute for Learning Algorithms (MILA) under supervision of Joseph Paul Cohen and Yoshua Bengio


Sept. 2019

One paper submitted to Numerical Methods in Biomedical Engineering on domain methods for needle insertion simulations.


July 2019

One paper accepted to ICCV 2019 on Deep Metric Learning with Interclass Properties. Arxiv: Link, Github: Link


May 2019

One paper submitted to NeurIPS 2019 on Deep Metric Learning.


April 2019

One paper submitted to MICCAI 2019 on Image Segmentation.


March 2019

One paper submitted to ICCV 2019 on Deep Metric Learning.


  Publications

ECCV 2020
DiVA: Diverse Visual Feature Aggregation for Deep Metric Learning
Timo Milbich*, Karsten Roth*, Homanga Bharadhwaj, Samarth Sinha, Yoshua Bengio, Bjoern Ommer, Joseph P. Cohen
TPAMI 2020
Sharing Matters for Generalization in Deep Metric Learning
Timo Milbich*, Karsten Roth*, Biagio Brattoli, Bjoern Ommer
ICML 2020
Revisiting Training Strategies and Generalization Performance in Deep Metric Learning
Karsten Roth*, Timo Milbich*, Samarth Sinha, Prateek Gupta, Bjoern Ommer, Joseph P. Cohen
CVPR 2020
PADS: Policy-Adapted Sampling for Visual Similarity Learning
Karsten Roth*, Timo Milbich*, Bjoern Ommer
IJBE 2020
Diffuse Domain Method for Needle Insertion Simulation
Katharina I. Jerg, René Phillip Austermühl, Karsten Roth, Jonas Große Sundrup, Guido Kanschat, Jürgen W. Hesser, Lisa Wittmayer
CUREUS 2020
Predicting COVID-19 Pneumonia Severity on Chest X-ray with Deep Learning
Joseph Paul Cohen, Lan Dao, Paul Morrison, Karsten Roth, Yoshua Bengio, Beiyi Shen, Almas Abbasi, Mahsa Hoshmand-Kochi, Marzyeh Ghassemi, Haifang Li, Tim Q Duong
ICCV 2019
MIC: Mining Interclass Characteristics for Improved Metric Learning
Karsten Roth*, Biagio Brattoli*, Bjoern Ommer
MICCAI 2017
The Liver Tumor Segmentation Benchmark (LiTS)
Patrick Bilic, ..., Karsten Roth, ..., Bjoern H. Menze
Phys. A.R. 2014
Efficient preparation and detection of microwave dressed-state qubits and qutrits with trapped ions
Joe Randall, ..., Karsten Roth, Winfried Hensinger

  Research

Remote Research Intern @Vector
2020/05 - present Prof. Marzyeh Ghassemi
Working on Few-Shot and Zero-Shot fundamental research and applications to the medical domain.
Research Intern @MILA
2019/09 - 2020/03 Dr. Joseph P. Cohen, Prof. Yoshua Bengio
Research intern at MILA under supervision of Joseph Paul Cohen and Yoshua Bengio. Worked on unsupervised representation learning and transfer learning for 3D CT data. Concurrently worked on fundamental Deep Metric Learning, with two publications in ICML 2020 and ECCV 2020.
Research Assistant @HCI
2018/10 - 2019/06 Biagio Brattoli, Timo Milbich, Bjoern Ommer
Worked on feature mining and training policies for improved generalization in Deep Metric Learning. Resulted in three publications at ICCV 2019, TPAMI 2020 and CVPR 2020.
Research Assistant @CIID
2017/09 - 2019/04 Philipp Klein, PD Alessia Ruggieri, Prof. Fred Hamprecht
Worked at the Center for Integrative Infectious Disease Research in collaboration with the Heidelberg Collaboratory for Image Processing (HCI). Focus on tracking and data colocalization for Hepatitis-C infected cells to investigate oscillatory stress behaviour. Publication currently in progress.
Research Assistant @UMM
2017/09 - 2019/04 Dr. Tomasz Konopczynski, Prof. Juergen Hesser
Worked at the Experimental Oncology Group at the University Hospital Mannheim. Worked on Medical Image Analysis, specifically Semantic Segmentation of Liver and Liver Lesions in CT data.
Intern @IQT
2013/08 - 2014/04 Sebastian Weidt, David Murgia, Prof. Winfried Hensinger
Internship at the Ion Quantum Trapping group (IQT) at Sussex University. Working on various Quality-of-Life duties (circuit board generation, soldering, building switches), and frequency modulation systems.

  Repositories

Deep-Metric-Learning-Baselines
PyTorch Implementation for Deep Metric Learning Pipelines.
  Python   353   64
Revisiting_Deep_Metric_Learning_PyTorch
(ICML 2020) This repo contains code for our paper "Revisiting Training Strategies and Generalization Performance in Deep Metric Learning" (https://arxiv.org/abs/2002.08473) to facilitate consistent research in this field.
  Python   72   10
ICCV2019_MIC
Implementation for our ICCV 2019 paper: MIC: Mining Interclass Characteristics for Improved Metric Learning.
  Python   49   5
CVPR_2020PADS
Code for PADS: Policy-Adapted Sampling for Visual Similarity Learning (CVPR 2020)
  Python   46   6
unet-lits-2d-pipeline
Liver Lesion Segmentation with 2D Unets.
  Python   34   11