MVPy
A fast, GPU-enabled multivariate analysis toolkit for neuroscience and beyond, designed around a scikit-learn-inspired API.
HI, THIS IS
Fabian Schneider
AI / ML engineer · neuroscientist
I build models and tools for complex human data, with a focus on representation learning, bio- and psychophysical data, and reliable scientific software.
01. About
I am a computational neuroscientist and machine learning engineer working across AI,
neural representations, and scientific software.
My research explores how the human brain represents and applies knowledge, how modern
machine learning methods can aid our understanding of the brain, and how insights into
the brain may aid machine learning in return. I am particularly interested in representation
learning, human variability, and neural coding.
Beyond research, I build tools that make multivariate analysis and dynamical systems more accessible, performant, and reproducible. I enjoy working on technically demanding problems that require rigorous thinking and innovative methods.
learning breakout
gen
01
score
00
best
00
evolving spiking network
←
→
02. Projects
Personalising language embeddings reveals structured individual differences in meaning
We show that semantic embeddings capture shared human semantic geometry, but not the substantial individual variation in human semantics. We demonstrate that finetuning based on only ten minutes of data can boost embedding-human fit by 40%.
Sensory sharpening and prediction errors unify competing models of predictive processing in human speech comprehension
We show that semantic expectations about speakers shift auditory perception towards expected sounds, with delayed processing of semantic error signals.
03. Toolkit
AI / ML
PyTorch
Hugging Face
scikit-learn
spaCy
Languages
Python
Julia
Go
C++
Neuroscience
- EEGElectroencephalography
- MEGMagnetoencephalograhy
- MVPMultivariate pattern analysis
Engineering
Git
Testing / CI
Containers / environments
Sphinx
04. Experience
PhD Researcher
Research on individual semantic representations, predictive processing, and computational modelling.
Research assistant
Research on efficacy of interventions for systolic blood pressure in low- to middle-income countries.
Research assistant
Research on viability of different neural coding schemes on neuromorphic hardware.
Research assistant
Research on spiking neural network models of auditory word recognition using tripod models.
MSc Cognitive Neuroscience
Training across cognitive neuroscience, computational methods, and scientific communication.
Erasmus+
Training across biopsychology and statistics.
BA Language, Literature, Culture
Training across linguistics, cultural studies, and scientific communication.
05. Contact
