Projects



Brain Decoding



Train Wavenet-based group-level models on MEG data, and uncover neuroscientifically interpretable information. Explore the differences between sliding window and full-epoch models in my other repo.

EMG silent-speech BCI



A proof-of-concept real-time silent speech BCI using EMG electrodes.



Protein Structure Simulation and Searching Algorithms



Easily simulate the change of dipole moments of complex protein structures and provide a 3D UI in OpenGL. There are also several searching algorithms implemented to find protein structures that satisfy user defined logic functions. Download here.

Processing Demos



While learning processing from this book, I created a repository for my example demos. You can check two of them in the browser.

OpenGL game



A very simple platformer game in OpenGL. You can download it here.

Papers

PhD thesis: Decoding non-invasive brain activity with novel deep learning approaches

University of Oxford (2024)

Richard Csaky

[Thesis]

Generalizable gesture recognition using magnetomyography

biorxiv (2024)

Richy Yun, Richard Csaky, Debadatta Dash, Isabel Gerrard, Gabriel Gonzalez, Evan Kittle, David Taylor, Rahil Soroushmojdehi, Dominic Labanowski, Nishita Deka

[Paper]

Magnetomyography: A novel modality for non-invasive muscle sensing

biorxiv (2024)

Richy Yun, Gabriel Gonzalez, Isabel Gerrard, Richard Csaky, Debadatta Dash, Evan Kittle, Nishita Deka, Dominic Labanowski

[Paper]

Foundational GPT Model for MEG

arXiv (2024)

Richard Csaky, Mats W.J. van Es, Oiwi Parker Jones, Mark Woolrich

[Paper]

Inner Speech Decoding from EEG and MEG

Abstract @ BCI Meeting (2023)

Richard Csaky, Mats W.J. van Es, Oiwi Parker Jones, Mark Woolrich

[Abstract] [Poster] [Slides]

Interpretable many-class decoding for MEG

NeuroImage (2023)

Richard Csaky, Mats W.J. van Es, Oiwi Parker Jones, Mark Woolrich

[Paper] [Code]

Group-level Brain Decoding with Deep Learning

Human Brain Mapping (2023)

Richard Csaky, Mats W.J. van Es, Oiwi Parker Jones, Mark Woolrich

[Paper] [Poster] [Code]

The Gutenberg Dialogue Dataset

Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume (2021)

Richard Csaky, Gábor Recski

[Paper] [Code] [Demo] [Slides] [Poster]

Improving Neural Conversational Models with Entropy-Based Data Filtering

Proceedings of the 57th ACL (2019)

Richard Csaky, Patrik Purgai, Gabor Recski

[Paper] [Filtering Code] [Training Code] [Evaluation Code] [Poster (vertical)] [Poster (horizontal)] [Blog post] [Slides]

Proposal Towards a Personalized Knowledge-powered Self-play Based Ensemble Dialog System

arXiv (2019)

Richard Csaky

[Paper]

Deep Learning Based Chatbot Models

National Scientific Students' Associations Conference (2017)

Richard Csaky

[Paper] [Code]

Study of dipole-dipole coupled protein-based circuits using self-developed simulation software

Scientific Students' Associations Conference (2016)

Richard Csaky, Edvard Bayer

[Paper] [Code]

Presentations

BCI Meeting 2023

June 2023

Inner Speech Decoding from EEG and MEG

[Poster] [Slides]

Biomag 2022

August 2022

Generalizing Brain Decoding Across Subjects with Deep Learning

[Poster]

OHBM 2022

June 2022

Generalizing Brain Decoding Across Subjects with Deep Learning

[Poster]

Cortico 2022

March 2022

Generalizing Brain Decoding Across Subjects with Deep Learning

[Poster]

EACL 2021

April 2021

The Gutenberg Dialogue Dataset

[Poster]

OUBT Biohackathon

March 2021

Brainstream: machine learning driven BCI that translates thoughts into text

EurNLP 2019

October 2019

Improving Neural Conversational Models with Entropy-Based Data Filtering

[Poster]

NLP for ConvAI workshop @ ACL

August 2019

Improving Neural Conversational Models with Entropy-Based Data Filtering

[Poster]

ACL 2019

July 2019

Improving Neural Conversational Models with Entropy-Based Data Filtering

[Talk]

EEML 2019

July 2019

Improving Neural Conversational Models with Entropy-Based Data Filtering

[Poster]

RAAI 2019

June 2019

Improving Neural Conversational Models with Entropy-Based Data Filtering

[Poster]

Hungarian NLP Meetup

May 2019

Neural Chatbots

[Slides]

Experience


  • Feb 2024 - Present
    Computational Neuroscientist @ Sonera

    Developing machine learning methods for analyzing and decoding multichannel electrophysiology data.


  • Apr 2023 - Present
    Technical Specialist @ BCI Stealth Startup

    I am a former co-founder and current specialist at a BCI stealth startup.


  • Jul 2023 - Dec 2023
    Scientific Consultant @ Sonera

    Developed machine learning methods for analysis and decoding of multichannel electrophysiology data.


  • Oct 2020 - Dec 2023
    PhD in Machine Learning + Neuroscience @ University of Oxford

    My research at Oxford is threefold.
    First, I successfully designed and executed inner speech EEG and MEG experiments. I collected over 40 hours of data, which I analysed using signal processing, machine learning, and innovative interpretability and visualization methods.

    Second, I pioneered the use of a new noninvasive device, optically-pumped magnetometers (OPM). I successfully demonstrated the effectiveness of language decoding in OPMs in a first-of-its-kind experiment.

    Finally, my research in deep learning methods has led to significant advancements in group-level decoding of brain data, resulting in two publications and presentations at multiple international conferences. For example, I adapted the conditional Wavenet model to generate and decode brain data, through both discriminative and generative frameworks. My model increased the multi-subject decoding accuracy of visual images from MEG data, from 15% to 50%. I also adapted the GPT2 model for multichannel MEG forecasting and showed that generated data resembles real data in a number of metrics. Generated data can be used as pretraining for a decoder to improve performance.


  • Feb 2020 - Jun 2020
    M.S. Erasmus Semester @ KU Leuven

    I participated in the Erasmus program at KU Leuven, taking 4 courses: Bioinformatics, Brain Computer Interfaces, Topics in Behavioural Neuroscience, Artificial Neural Networks and Deep Learning.


  • Sep 2018 - Jun 2020
    Computer Science Engineering M.S. @ Budapest University of Technology

    Excellent with Highest Honours, 4.73/5.00 degree grade. During my Master's I worked on dialogue modeling research, detailed in my Natural Language Processing Researcher experience description. I presented at 6 conferences and meetups and I was awarded the prestigious National Excellence Scholarship. I also attended EEML, a machine learning summer school in Romania.


  • Feb 2018 - Oct 2019
    Natural Language Processing Researcher @ Budapest University of Technology

    As a researcher during my Master's, I developed one of the first Transformer-based chatbots in 2017. This project and a comprehensive review of 100+ publications received over 400 GitHub stars and earned me a national research competition win. Despite having no prior Python or deep learning experience, I achieved these accomplishments in under a year. My research included the development of new methods and datasets for improving chatbots, resulting in multiple actively used GitHub repositories (10-100 stars) and a publication at the Association for Computational Linguistics, the leading conference in the field.

    I also developed an online deep learning chatbot, utilized by hundreds. I modified and trained GPT-2 on a dataset containing 14.8 million utterances in 7 languages, that I created from books. I integrated the models into the backend of my website with a response time of less than 1 second, capable of handling tens of simultaneous users. These accomplishments also earned me a national excellence scholarship and a salary from the university, a rare achievement for a Master's student.


  • Jul 2017 - Aug 2018
    Software Engineer @ Bosch

    As a data-driven problem solver at Bosch, I successfully applied deep learning techniques to improve parking space segmentation through my self-initiated project. By developing an OpenGL-based user interface in both C++ and Python, I was able to manipulate parking spots in real-time, overlaid on a car's camera feed. I collected and labelled 10,000 images with the help of a test driver and modified the YOLO computer vision model achieving impressive results. These efforts led to the allocation of a dedicated team and additional funding for the project by upper management. Throughout the project, I navigated cross-functional (economic, social, engineering, software) challenges with ease, collaborating with engineering groups to drive success.


  • Sep 2014 - Jan 2018
    Mechatronics Engineering B.S. @ Budapest University of Technology

    Excellent with Highest Honours, 4.79/5.00 degree grade. During my B.S. I took part in a research project and implemented a molecular circuit simulation platform in C++, winning 2nd place at a university conference.

Resume [download]