Christoffer Löffler

I am a profesor asociado at the school of informatics of the Pontificia Universidad Católica de Valparaíso (PUCV) in Chile.

My research is on machine learning with few labeled data and broadly on time series. I also contributed to optical positioning, complex event processing, and large-scale virtual reality.

Recently, I got my Dr.-Ing. at Björn Eskofier's MaD Lab, while working at Fraunhofer IIS in the ADA Lovelace Center as a senior scientist. I also did my Bachelors and Masters at Friedrich-Alexander University Erlangen-Nürnberg (FAU), where I was advised by Christopher Mutschler, back then researching at Michael Philippsen's Programming Systems lab. These and other collaborations led to a Dijkstra number of four and an Erdös number of fivefour!

Email  /  CV  /  Google Scholar  /  ResearchGate  /  ORCID  /  Github  /  Twitter  /  LinkedIn  /  Youtube  /  Blog

profile photo
Research

I'm interested in machine learning with few labeled data (like active learning) and building interactive systems for the real-world.

Learning with Limited Labelled Data
Christoffer Löffler, Rasmus Hvingelby, Jann Goschenhofer
Mutschler, C., Münzenmayer, C., Uhlmann, N., Martin, A. (eds) Unlocking Artificial Intelligence. Springer, Cham.
pdf / DOI: 10.1007/978-3-031-64832-8_4

We explore how semi-supervised and active learning address the challenge of training machine learning models with limited labeled data by leveraging unlabeled data and optimizing annotation efforts, while comparing their principles, strengths, and future potential.

Sequence-based Learning
Christoffer Löffler, Felix Ott, Jonathan Ott, Maximilian P. Oppelt, Tobias Feigl
Mutschler, C., Münzenmayer, C., Uhlmann, N., Martin, A. (eds) Unlocking Artificial Intelligence. Springer, Cham.
pdf / DOI: 10.1007/978-3-031-64832-8_2

We cover various applications, data types, key architectures, and methods such as deep metric learning and time series similarity learning, while also addressing interpretability in terms of safety, fairness, and non-discrimination.

Autonomous Parameter Balance in Population-Based Approaches: A Self-Adaptive Learning-Based Strategy
Emanuel Vega, José Lemus-Romani, Ricardo Soto, Broderick Crawford, Christoffer Löffler, Javier Peña, El-Gazhali Talbi
Biomimetics 2024, 9(2), 82;
pdf / DOI: 10.3390/biomimetics9020082

We augment population-based meta-heuristics with a learning component to bias the exploration direction of the search.

Active Deep Learning of Representations for Similarity Search
Christoffer Löffler, supervisors Björn M. Eskofier and Ute Schmidt.
Dissertation, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Technische Fakultät 10/2023.
pdf / DOI: 10.25593/open-fau-42

This dissertation provides streamlined information retrieval, cost-effective annotation through Deep Active Learning, and efficient learning of similarity functions. Three key publications contribute to the thesis: a metric learner for enhanced information retrieval, an imitation learning approach (IALE) optimizing Deep Active Learning, and a method combining fine-tuning with active learning, facilitating adaptive similarity search for unstructured data and cost-effective annotation of complex datasets.

Don't Get Me Wrong: How to apply Deep Visual Interpretations to Time Series
Christoffer Löffler, Wei-Cheng Lai, Björn M. Eskofier, Dario Zanca, Lukas Schmidt, Christopher Mutschler
preprint, 09/2023
code / preprint

A framework of six orthogonal metrics for gradient- or perturbation-based post-hoc visual interpretation methods. Its designed for time series classification and segmentation tasks.

Active Learning of Ordinal Embeddings: A User Study on Football Data
Christoffer Löffler, Kion Fallah, Stefano Fenu, Dario Zanca, Bjoern Eskofier, Christopher J. Rozell, Christopher Mutschler
Transactions on Machine Learning Research (TMLR), 04/2023
TMLR infinite conf / pdf / OpenReview / video / project / code

This work improves information retrieval in a football trajectory dataset using deep metric learning and an entropy-based active learning method, and analyzes the effectiveness of sampling heuristics through a user study.

IALE: Imitating Active Learner Ensembles
Christoffer Löffler, Christopher Mutschler
Journal Track @ Neural Information Processing Systems (NeurIPS), New Orleans, LA, 11/2022
pdf / poster / video / code

IALE uses imitation learning to select informative data samples in deep active learning by imitating the best-performing expert heuristic, and outperforms state-of-the-art imitation learners and heuristics on well-known image datasets.

IALE: Imitating Active Learner Ensembles
Christoffer Löffler, Christopher Mutschler
Journal of Machine Learning Research 23, 02/2022
code / pdf

We propose an imitation learning approach that learns a policy for active learning from an ensemble of deep active learners.

Deep Siamese Metric Learning: A Highly Scalable Approach to Searching Unordered Sets of Trajectories
Christoffer Löffler, Luca Reeb, Daniel Dzibela, Robert Marzilger, Nicolas Witt, Björn M. Eskofier, Christopher Mutschler
ACM Transactions on Intelligent Systems and Technology, Special Issue on Intelligent Trajectory Data Analytics, 02/2022.
pdf / project / DOI: 10.1145/3465057

We enable interactive search in unordered sets of trajectories, with a focus on team sports.

Recipes for Post-training Quantization of Deep Neural Networks
Ashutosh Mishra, Christoffer Löffler, Axel Plinge
EMC^2: Workshop on Energy Efficient Machine Learning and Cognitive Computing, 2020
Axel's video / Ashutosh's video / pdf / project

We show that post-training quantization (done greedily) benefits from an optimal global bit-width and evaluate this on VGG, ResNet, UNet and our tool tracking FCN.

A Sense of Quality for Augmented Reality Assisted Process Guidance
Anes Redzepagic, Christoffer Löffler, Tobias Feigl, Christopher Mutschler
IEEE Intl. Symposium on Mixed and Augmented Reality (ISMAR), poster track, 2020
video / project / DOI: 10.1109/ISMAR-Adjunct51615.2020.00046

We combine inertial sensors, mounted on work tools, with AR-headsets to enrich modern assistance systems with a sense of process quality, powered by machine learning.

Automated Quality Assurance for Hand-held Tools via Embedded Classification and AutoML
Christoffer Löffler, Christian Nickel, Christopher Sobel, Daniel Dzibela, Jonathan Braat, Benjamin Gruhler, Philipp Woller, Nicolas Witt, Christopher Mutschler
European Conf. on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD), demo track, 2020
demo video / video presentation / project / pdf / DOI: 10.1007/978-3-030-67670-4_33

We describe an AutoML system for our custom hardware and classify multivariate data using deep and shallow methods.

ViPR: Visual-Odometry-aided Pose Regression for 6DoF Camera Localization
Felix Ott, Tobias Feigl, Christoffer Löffler, Christopher Mutschler,
Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition Workshops, workshop track, 2020
video / pdf / DOI: 10.1109/CVPRW50498.2020.00029

Learning to fuse absolute poses (6 degrees of freedom) with optical flow (e.g., FlowNet) to improve a mobile agent's self positioning.

Localization Limitations of ARCore, ARKit, and Hololens in Dynamic Large-scale Industry Environments.
Tobias Feigl, Andreas Porada, Steve Steiner, Christoffer Löffler, Christopher Mutschler, Michael Philippsen
15th Intl. Conf. on Computer Graphics Theory and Applications (GRAPP), 2020
pdf / DOI: 10.5220/0008989903070318

An evaluation of the big AR systems for real use-cases.

Evaluation criteria for inside-out indoor positioning systems based on machine learning
Christoffer Löffler, Sascha Riechel, Janina Fischer, Christopher Mutschler
IEEE Intl. Conf. on Indoor Positioning and Indoor Navigation (IPIN), 2018
project / warehouse dataset / pdf / poster / slides / DOI: 10.1109/IPIN.2018.8533862

Using reference positioning systems, we record a multi-camera dataset with exact labels and propose criteria for evaluating indoor positioning.

Optical Camera Communication for Active Marker Identification in Camera-based Positioning Systems
Lorenz Gorse, Christoffer Löffler, Christopher Mutschler, Michael Philippsen
IEEE 15th Workshop on Positioning, Navigation and Communications (WPNC), 2018
pdf / DOI: 10.1109/WPNC.2018.8555846

How to build a cheap but reliable optical positioning system with Raspberry Pi with active LED markers and continuously identify them.

Approximative event processing on sensor data streams
Christoffer Löffler, Christopher Mutschler, Michael Philippsen
Proc. of the 9th ACM Intl. Conf. on Distributed Event-Based Systems (DEBS), poster track, best demo/poster award, 2015
pdf / DOI: 10.1145/2675743.2776767

Event-Based Systems (EBS) can efficiently analyze large streams of sensor data in near-realtime. But they struggle with noise or incompleteness that is seen in the unprecedented amount of data generated by the Internet of Things

Predictive load management in smart grid environments
Christopher Mutschler, Christoffer Löffler, Nicolas Witt, Thorsten Edelhäußer, Michael Philippsen
Proc. of the 8th ACM Intl. Conf. on Distributed Event-Based Systems (DEBS), 2014
pdf / DOI: 10.1145/2611286.2611330

We've won the DEBS 2014 Grand Challenge with our hidden Markov model approach.

Simulating the energy management on smartphones using hybrid modeling techniques
Ibrahim Alagöz, Christoffer Löffler, Vitali Schneider, Reinhard German
GI/ITG Intl. Conf. on Measurement, Modelling, and Evaluation of Computing Systems and Dependability and Fault Tolerance (MMB & DFT), best student paper award, 2014
pdf / DOI: 10.1007/978-3-319-05359-2_15

We model a smartphone playing back music and its energy management.

Evolutionary algorithms that use runtime migration of detector processes to reduce latency in event-based systems
IEEE NASA/ESA Conf. on Adaptive Hardware and Systems (AHS), 2013
Christoffer Löffler, Christopher Mutschler, Michael Philippsen
pdf / DOI: 10.1109/AHS.2013.6604223

When running a distributed low-latency event processing system, you may want to optimize latency - heuristically.

Patents
Method to determine a present position of an object, positioning system, tracker and computer program
Stephan Otto, Tobias Feigl, Christian Daxer, Alexander Bruckmann, Christoffer Löffler, Christopher Mutschler, Marc Faßbinder
US20200371226A1. Active since 30th of May 2023.
Google Patents

Positioning system with combined optical and a radio-based determination of a position of a tracker and a tracker with an active light source.

Apparatus, method and computer program for improving the performance of an event-based distributed analysis system
Christopher Mutschler, Christoffer Löffler
DE102012112253B4. Active on 25th of April 2019, now expired.
Google Patents

Method for improving the performance of an event-based distributed analysis system that uses meta heuristics.

Teaching
Lecture "Inteligencia Artificial" (AI), primer semestre 2024
Lecture "Aprendizaje Automático" (ML), primer semestre 2024
DRL seminar, deep reinforcement learning, summer 2019
Machine Learning for Timeseries (project), winter 2019-2021
Girls/Youth and Technology, summer 2019

SemML, seminar on machine learning, winter 2015-2022
Seminar Event Processing, summer 2016

This website is forked from here, thank you Jon Barron.