Lukas Muttenthaler
PhD student in Machine Learning at the Department of Software Engineering and Theoretical Computer Science, TU Berlin, working on using human inductive biases to align and improve neural network representations, with a focus on low data regimes.
Research Profile
I am a PhD student in Machine Learning at TU Berlin and the Berlin Institute for the Foundations of Learning and Data (BIFOLD). I am also a guest researcher in the ViCCo Group at the Max Planck Institute for Human Cognitive and Brain Sciences. I am mainly advised by Klaus-Robert Müller and co-supervised by Martin Hebart. I am also part of a Research Collaboration with Google DeepMind, where I am advised by Simon Kornblith. Previously, I was a MSc student in IT & Cognition / Computer Science of Isabelle Augenstein and Johannes Bjerva at the University of Copenhagen where I mostly worked on Question Answering and Machine Translation.
My research mainly revolves around representation learning. In particular, I am trying to both understand the factors that influence the degree of alignment between human and neural network representations and figure out how we can leverage human inductive biases or insights from cognitive science most fruitfully for (robust) machine learning. Have a look at my Google Scholar for more information about my work. Feel free to reach out to me, if you believe our research intentions are aligned and you are keen to collaborate on a project.
Academic Background
-
2021- | PhD, Machine Learning, Machine Learning Group & BIFOLD, TU Berlin, Germany
-
2021- | Guest researcher, Computational Cognitive Neuroscience ViCCo Group, Max Planck Insitute for Human Cognitive and Brain Sciences, Germany
-
2020-2021 | Research Assistant (RA), Computational Cognitive Neuroscience ViCCo Group, Max Planck Insitute for Human Cognitive and Brain Sciences, Germany
-
2018-2020 | M.Sc. IT & Cognition Department of Computer Science, University of Copenhagen, Denmark – Thesis: Subjective Question Answering: Deciphering the inner workings of Transformers in the realm of subjectivity
-
2015-2018 | B.Sc. Cognitive Psychology Faculty of Psychology, University of Vienna, Austria
Links
News
2023
- Paper on interpretable object dimensions in DNNs accepted to CCN 2023!
- One workshop proposal on representational alignment submitted to NeurIPS 2023!
- Two papers submitted to NeurIPS 2023!
- I am presenting our paper on human alignment of neural network representations virtually at ICLR 2023!
- Paper on human alignment of neural network representations accepted to ICLR 2023 in Kigali, Rwanda!
2022
- Paper accepted to the SVRHM workshop @ NeurIPS 2022!
- One paper submitted to ICLR 2023!
- I gave my first ever Machine Learning lecture at TU Berlin!
- Our paper on Variational Interpretable Concept Embeddings (VICE) was accepted to NeurIPS 2022!
- I received a Google Research Collabs grant with Simon Kornblith to work on human alignment (80.000 USD research funding + 10.000 USD Google Cloud credits)! I will be commencing the collaboration as part of the Google Research Collabs Cohort 2022/23 in November!
- I was awarded 1,000 USD as part of the Google Cloud Platform research credits programme for working on long-tail classification with Thomas Unterthiner!
- One paper submitted to NeurIPS 2022!
- Abstract on interpretable object dimensions in DNNs accepted as an oral to VSS 2022!
- I was accepted into the Berlin Institute for the Foundations of Learning and Data (BIFOLD) to continue my PhD studies at TU Berlin with a fully-funded position!
2021
- Paper on my Python toolbox thingsvision for extracting features from state-of-the-art computer vision models accepted to Frontiers in Neuroinformatics!
- As of July 1st, I am a PhD student in the Machine Learning Group at TU Berlin, working on cognitively inspired Machine Learning! Excited to continue my journey on the intersection of ML and Cognitive Science; this time with a stronger emphasis on ML rather than on Cognitive Science.
2020
- Paper on Unsupervised QA Evaluation of Transformers with my previous MSc. supervisors Isabelle Augenstein and Johannes Bjerva accepted to BlackBoxNLP @ EMNLP 2020!
- In August I started as a full-time Research Assistant (RA) in the Vision and Computational Cognition Group at the Max Planck Institute for Human Cognitive and Brain Sciences in Leipzig, Germany. Excited to start working at the intersection of Machine Learning and Cognitive Neuroscience!
- In June I successfully defended my MSc thesis on Subjective Question Answering and graduated with distinction from the University of Copenhagen.
2019
- Our paper on Assisted Declarative Process Creation from Natural Language Descriptions won the Best Demonstration Award at EDOC 2019!
- Paper on Assisted Declarative Process Creation from Natural Language Descriptions accepted to the 23rd IEEE International Enterprise Distributed Object Computing Conference (EDOC 2019)!
- I started working on NLP for law documents as a Research Assitant (RA) in the Department of Computer Science at the University of Copenhagen.
- Our algorithm that won the PAN autorship attribution competition at CLEF 2019 was featured in the CLEF 2019 Working Notes!
- Our team won the PAN authorship attribution competition at CLEF 2019 in Lugano, Switzerland!
- Paper on Visual Working Memory from my time as a research intern during my BSc accepted to Frontiers in Psychology!
- Paper on Amblyopia from my time as a research intern during my BSc accepted to the Journal of European Psychology Students (JEPS)!
2018
- I was addmitted to the IT & Cognition MSc programme at the University of Copenhagen. Excited to start working on Natural Language Processing in Copenhagen! I will be starting my studies in September.