Recent content in Neuromorphic computation and intelligence on Jens Egholm Pedersen
A short talk for the Acoustic Technology presentation day at DTU Electro, aimed at companies and fellow researchers, on what neuromorphic computing can offer signal processing in acoustics — and beyond. Slides are available here.
A guest lecture for 342 Energy Efficient Computing course at Stanford University, co-presented with Steven Abreu (University of Groningen). The talk covers (1) the representation relation between physical systems and abstract models (Horsman, 2014), (2) computational classes and structure-preserving maps (Yanofsky, 2022), (3) source-independent linking and semantic retention for compiler…
A guest lecture for 27020 Bioengineering (Polytechnical Foundation) at DTU on brain-inspired computing with spiking neural networks. The lecture introduces BSc students to how the brain computes with spikes and why this matters for engineering, covering (1) how the brain computes, (2) spiking neural networks and neuromorphic hardware, (3) event cameras with a live demo, (4) covariant…
This is a companion exercise for the guest lecture on Brain-Inspired Computing with Spiking Neural Networks in 27020 Bioengineering. Leaky Integrate-and-Fire Neuron A LIF neuron integrates incoming current and fires a spike when its membrane voltage crosses a threshold. Between spikes, the voltage leaks back towards rest — just like a biological neuron. The continuous-time equation for a LIF…
A talk for the Advanced Topics in Hearing Research seminar at DTU Department of Health Technology on brain-inspired event-based signal processing and neuromorphic computing. The talk covers three parts: (1) the case for event-driven processing, drawing parallels between biological sensory systems and event cameras, (2) covariant signal processing with spiking neurons, including spatio-temporal…
A talk for the REAL group (Robotics, Evolution and Artificial Life) at the IT University of Copenhagen on event-driven sensing and neuromorphic computation for embodied intelligence. The talk covers three themes: (1) the case for neuromorphic computing, motivated by the energy efficiency and speed of event cameras and neuromorphic hardware, (2) mathematical foundations for spiking neural networks,…
A talk for researchers who use computation but may not think of themselves as software developers. The argument: software is not a detail of computational science, it is the science. Starting from Kenneth Craik’s model of cognition, I trace what it means for code to carry (or corrupt) knowledge—and what a healthy lifecycle looks like in practice. Topics covered include reproducibility,…
I was invited to give student talk on at the Open Neuromorphic student talk about spiking neural receptive fields. My talk expands on our work on spatio-temporal covariance for leaky integrators and leaky integrate-and-fire neurons published at Nature Communications in 2025. I motivated the problem about neuromorphic computing, talked about the necessity for covariance guarantees, and discussed…
I defended my PhD at the KTH Royal Institute of Technology! I’m finally #PhDone! Slides are available here.
Neuromorphic computing exploits the laws of physics to perform computations, similar to the human brain. If we can “lower” the computation into physics, we achieve extreme energy gains, up to 27-35 orders of magnitude. So, why aren’t we doing that? Presently, we lack theories to guide efficient implementations. We can build the circuits, but we don’t know how to combine…
I was invited to give an overview of neuromorphics and talk about my research to the Acoustic Technology group at the Danish Technical University. Slides are available here.
I was invited to talk about dynamical systems for control at the Telluride neuromorphic cognition engineering workshop. It was a tremendous pleasure and honor to talk at the workshop with a ridiculously cool line-up from world-leading researchers. Slides are available here and a recording can be provided upon request.
TL;DR I claim that modern science is synonymous with open source software. This post explains why, why it matters, and what you can (and should) do next. Why do you care about (open source) software? - Everyone I spend a lot of my time working on software. I have been asked why software matters more times than I can remember. Software is, people say, not science. It’s a time sink, something…
I was invited to talk about neuromorphic computing at the Faroese tech festival, Tonik. Tonik is a fantastic mix of tech, business, and art and I thoroughly enjoyed interacting with the driven and creative minds there! The talk is popular-scientific in nature: I give negative and positive motivations (sticks and carrots) for researching and applying neuromorphics. And I talk about technology that…
I was invited to talk about neuromorphic computing from first principles at the Stategic Alliance for Neuromorphic Computing and Engineering (STANCE), organized by the Frauenhofer Institute for Industrial Mathematics. In the talk, I give an overview of computational definitions for digital, analog, and neuromorphic hardware. I motivate NIR, present some existing open-source efforts in the…
Are we heading towards a power wall? $10^{-12}$
Anders Høeg Nissen interviewed me for the AI Denmark podcast where we had great fun discussing neuromorphic computing and how the technology can help AI. Anders asked some profound and poignant questions that really made me think. I hope you enjoy the episode as much as I did! The episode is available on the AI Denmark website and on Spotify. Note that the podcast is in Danish.
I was invited to give a webinar at the academic labour union for academic professionals in Denmark, DM. I presented the high-level concept of neuromorphic computing, which technologies exist today, and how it can be used to solve real-world problems. The presentation was not recorded, but the slideshow can be found here.
I was invited to talk at a conference for leading IT managers in Denmark, Dansk IT Lederfokus. I discussed the future of AI and neuromorphic computing, and how these technologies can be used to solve real-world problems. The presentation was not recorded, but the slideshow can be found here.
The Danish Association of Masters and PhDs interviewed me about neuromorphic computing and future prospects. The article can be read here.
I was invited to the Brains and Machines podcast podcast to discuss my work on computational abstractions for neuromorphics, event-based vision, and NIR. I really enjoyed the conversation with Giulia d’Angelo, and I hope you will too! The episode can be found at https://www.eetimes.com/podcasts/next-gen-neuromorphic-researchers-look-to-future/
The event based dataset can control arbitrary affine transformations and outputs very sparse data, perfect for training event-driven neural network models. We released a library for generating event-based vision data with a host of features. Our work can be used to generate data for training and testing event-based vision algorithms, under controlled conditions, and with a high degree of…
I presented at the OffDig conference in Aarhus, Denmark, on the topic of neuromorphic computing and its applications in event-based vision and neuromorphic intelligence. The presentation was not recorded, but the slideshow can be found here.
I recently gave a seminar summarizing parts of my research. The seminar marks the 80% progress towards my PhD. The video can be found here: https://play.kth.se/embed/secure/iframe/entryId/0_o452yw54/uiConfId/23453971/pbc/347466/st/0
Together with three of co-authors of the NIR paper, Jason Eshragian, Bernhard Vogginger, and Felix Bauer, we published a hands-on tutorial of the Neuromorphic Intermediate Representation (NIR) on YouTube. We first outline and motivate NIR in a historical context and explain why NIR offers one solution to the problem of interoperability between neuromorphic hardware and software. Felix, Bernhard…
I presented the Neuromorphic Intermediate Representation at the Spiking Neural networks as Universal Function Approximators (SNUFA) workshop. The slides can be found here.
I gave an overview of neuromorphic engineering for the Nordic Neuromorphs journal club. The slides can be found here.
We published a paper on Translation and Scale Invariance for Event-Based Object tracking for the Neuro-Inspired Computational Elements Conference (NICE) held in San Antonio, Texas. The paper contributes Spiking neural networks rivals artificial neural networks when operating on event-based data A differentiable method to convert spikes into coordinates in $\mathbb{R}^2$ Real-time object tracking…
AEStream sends data from input sources to output sinks. We published a paper on the technologies behind the event-based processing library, AEStream. Specifically, we used coroutines to demonstrate that we can achieve at least 2x speedups on conventional CPUs and GPUs, compared to parallelization approaches that does not use coroutines. The work contributes A plug-and-play library that streams…
I recently held a seminar demonstrating 50% completion of my PhD thesis on Neuromorphic computing for closed-loop systems. The ~60 minute presentation (including Q&A) can be found here: https://cloud.jepedersen.dk/s/W3DPjTewgtGWiBr
A presentation on the need for neuromorphic abstractions given to the Open Neuromorphic community. The slides can be found here.
A video tutorial on neuromorphic real-time programming using Norse and AEStream. The video can be seen here and the slides can be found here.
An essay on the invention of $\lambda$ calculus, as a PDF.
The search for building and understanding intelligent machines has been under way for thousands of years and appear fundamental to the human condition [@Russel_Norvig_2010]. Deep learning models have recently been heavily dominating this search for artificial intelligence (AI), as demonstrated by the go-playing Alpha Go [@Silver2017], the narrating GPT-3 [@Brown2020], and many more. This is an odd…
During the conference on Spiking Neural networks as Universal Function Approximators (SNUFA), I gave a talk about the work I have done together with Christian Pehle on a PyTorch-based simulator for biological spiking neural networks.
I recently held a seminar demonstrating 30% completion of my PhD thesis on Neuromorphic computing for closed-loop systems. The ~60 minute presentation (including Q&A) can be found below:
This page is meant as a guide and quickstart for students at KTH. Emphasis on the word guide; critical thinking is your friend in every aspect in life. Never, never, never confuse well-meaning advise with the truth. Getting started No matter what type of pressure you’re put under to deliver a product for a company, good grades for your mom, money for your family etc., what you have to…
I gave a talk on brain-inspired control systems for the Human Brain Project (HBP) student community. The slides can be found here.
This new website will feature the research, thoughts, and notes of Jens Egholm Pedersen Welcome!
Portrait of Jens Egholm Pedersen. Jens Egholm Pedersen is a doctoral student at KTH Royal Institute of Technology working to model and construct neuromorphic control systems. His research explores the intersection of neuroscience and computer science, developing brain-inspired computing systems that process information with remarkable efficiency and adaptability. Research Focus Jens’s…