steve abreu

prof_pic.jpg

I am a PhD student at the AI department of the University of Groningen, working in the MINDS group of Prof. Herbert Jaeger. I am affiliated with CogniGron and partially funded by Post-Digital. During my PhD, I was a visiting researcher at ETH Zurich and University of Gent and a research intern at Google and Intel Labs.

In research I keep a dual focus on advancing machine learning while exploring new compute paradigms and hardware for AI, especially those taking inspiration from the brain. I enjoy cross-disciplinary research between AI, computer science, mathematics, neuroscience, physics, and cognitive science.

Within ML, I am interested in efficiency (e.g., model compression, low-power hardware), continual learning (e.g. using brain-inspired neuromodulation or plasticity rules), meta-learning and automated machine learning, among other things.

To make progress towards well-aligned and interpretable AI, I work on representation engineering and mechanistic interpretability of large language models.

I also work on physical and brain-inspired computing, developing theories that align computation with physics, in order to make better use of neuromorphic chips, photonic devices, and other physical computing systems. I aim to develop computational abstractions in physical substrates (e.g. through the neuromorphic intermediate representation, NIR), new hardware-compatible efficient learning algorithms, and principled ways of programming novel AI hardware.

I am also a hobby photographer, an enthusiastic coffee brewer, an avid motorcyclist, I like to read, travel, and go to music and art events.

news

May 2025 Presented “Large Language Models on a Tiny Power Budget” at ISCAS 2025 (Slides here)
Apr 2025 Presented four workshop papers at ICLR in Singapore!
Dec 2024 Joris presented our paper on conceptor steering at the MINT workshop at NeurIPS in Vancouver!
Jul 2024 Spent two weeks at Telluride in the Language and Thought topic area, presented our work on quantized SSMs at ICML in Vienna, and our neuromorphic programming paper at ICONS in Virginia!
May 2024 Starting my research internship at Intel’s Neuromorphic Computing Lab, working on neuromorphic transformer-like architectures based on state space models.
Dec 2023 Starting my research internship at Google in Waterloo (Canada), working on efficient and adaptive AR user interfaces with multimodal LLMs.
Oct 2023 Attending NNPC in Hannover. Presenting a poster on neuromorphic programming and abstractions.
Jun 2023 Spent three weeks in Telluride to work on neuromorphic abstractions, representation learning in RL, and other exciting stuff!
Jun 2023 Presenting our neuromorphic cytometry paper at CVPR (event-based vision workshop) in Vancouver.
Apr 2023 Attending CHI in Hamburg. Giving a talk on interfacing with bio-inspired computer hardware at the workshop on Bio-Digital Interfaces for Human-Computer Interaction.
Apr 2023 Gave a talk about neuromorphic computing to AI students in Groningen: “Computing and engineering principles from the brain”. Slides here.
Dec 2022 Spent 3 months at the University of Gent in the Photonics Research Group, working on photonic computing and neuromorphic flow cytometry (see publications, CVPR-W ‘23).
Jun 2022 Spent a week learning and discussing about machine learning and photonics in Como.
Jun 2022 Spent three months at the Institute of Neuroinformatics (ETH/UZH in Zurich), working on mixed-signal neuromorphic hardware.
May 2022 Talk on physical reservoir computing to the CogniGron research center (slides).
Feb 2022 Presentation on my PhD topic to the CogniGron center of cognitive materials (slides).

selected publications

  1. conceptorsteering.png
    Steering Large Language Models using Conceptors: Improving Addition-Based Activation Engineering
    Joris Postmus, and Steven Abreu
    In Neural Information Processing Systems (NeurIPS) 2024 - MINT Workshop, 2024
    Presented at NeurIPS workshop MINT
  2. qs5.png
    Q-S5: Towards Quantized State Space Models
    Steven Abreu*, Jens E. Pedersen*, Kade M. Heckel*, and 1 more author
    In International Conference on Machine Learning (ICML) 2024 - NGSM workshop, 2024
    Presented at ICML workshop NGSM
  3. mambaptq.png
    Mamba-PTQ: Outlier Channels in Recurrent Large Language Models
    Alessandro Pierro*, and Steven Abreu*
    In International Conference on Machine Learning (ICML) 2024 - ES-FOMO-II workshop, 2024
    Presented at ICML workshop ES-FOMO-II
  4. nir.png
    Neuromorphic Intermediate Representation: A Unified Instruction Set for Interoperable Brain-Inspired Computing
    Jens E. Pedersen*Steven Abreu*, Matthias Jobst, and 12 more authors
    Nature Communications (in press), 2024
  5. parseego4d.png
    PARSE-Ego4D: Personal Action Recommendation Suggestions for Egocentric Videos
    Steven Abreu, Tiffany D. Do, Karan Ahuja, and 4 more authors
    under review, 2024
  6. photonics.png
    A photonics perspective on computing with physical substrates
    S. Abreu, I. Boikov, M. Goldmann, and 25 more authors
    Reviews in Physics, 2024
  7. icons2024.png
    Neuromorphic Programming: Emerging Directions for Brain-Inspired Hardware
    Steven Abreu, and Jens E. Pedersen
    In Proceedings of the International Conference of Neuromorphic Systems, 2024
  8. dvsflow.gif
    Flow Cytometry With Event-Based Vision and Spiking Neuromorphic Hardware
    Steven Abreu, Muhammed Gouda, Alessio Lugnan, and 1 more author
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2023
  9. reservoir.png
    Hands-on reservoir computing: a tutorial for practical implementation
    Matteo Cucchi*Steven Abreu*, Giuseppe Ciccone, and 2 more authors
    Neuromorphic Computing and Engineering, 2022
    Accepted manuscript
  10. jub.png
    Automated architecture design for deep neural networks
    Steven Abreu
    ArXiv, Aug 2019