Nagele mit Köpfchen - Ep. 08: Spiking Neural Networks and SpikeGPT with Ridger Zhou

from | 10 June 2024 | Tech Deep Dive

Dr Johannes Nagele, Senior Principal AI Researcher and Consultant at [at], hosts the podcast "Nagele mit Köpfchen" in which he discusses the latest trends in AI with AI experts. In this episode, he talks to Ridger Zhou, a PhD student in electrical and computer engineering at the University of California, Santa Cruz.

YouTube

By loading the video you accept YouTube's privacy policy.
Learn more

load Video

AI Talks: Nail with brains | Ep. 08 | Ridger Zhou

Richard stops his work SpikeGPT before, a AI-based model that combines spiking neural networks (SNNs) with traditional artificial neural networks (ANNs)and discusses both its potential and its drawbacks. Regarding the latter, Richard notes that SNNs still cannot match the precision and performance of CNNs due to their binary nature. Nevertheless, he emphasises his assumption that SNNs could eventually outperform traditional neural networks in machine learning, as they are able to more closely mimic the functions of the human brain. In addition, how and why SpikeGPT-like models are particularly energy efficient will be discussed, as well as their promising applicability for tasks that require continuous learning. The audience will be encouraged to try out SpikeGPT for themselves.


Listen to and subscribe to the latest episode now on Apple Podcasts, Spotify, Google Podcasts and Deezer:

Click on the button below to load the content of nagele-mit-koepfchen.podigee.io.

Load content

Listen in if you want to do "Nagele mit Köpfchen" instead of pompous business blah-blah!

Author

[at] Video Transcriptor

This summary was created with our in-house video transcriber, which uses OpenAI Whisper to automatically transcribe and summarise video content.
With the ability to process both YouTube URLs and local files (up to 200 MB), the AI application creates accurate transcripts and summaries. Of particular note are the Transcription Engine 1 & 2, which allow transcribed words to be colour-coded according to model confidence - from red (unsafe) to yellow (neutral) to white (safe). This visual aid provides users with a quick assessment of transcription accuracy.

Do you also need generative AI solutions for your internal video materials such as meeting recordings or company presentations?
Then please contact us. We look forward to exchanging ideas with you.

0 Kommentare