How can a neural network be like the brain?
NeuroAI is the budding research field at the intersection of neuroscience and artificial intelligence. One of the core concepts used in the field is that artificial neural networks can act as good models of the brain. For example, it’s often claimed that convolutional neural networks can account for the response of the ventral visual stream…
How much energy does ChatGPT use?
I received this intriguing question from Daria Marchenko, who is a contributor on Moteur de Recherche, on Canadian public radio. I went into a deep rabbit hole trying to find a good answer–especially since she also interviewed the wonderful Sasha Luccioni, an expert on ML and climate change. Most of the estimates I’ve found online…
Does GPT-4 have common sense?
There’s been remarkable progress in Masked Language Modeling (MLM) and Reinforcement Learning from Human Feedback (RLHF), taking us much closer to generally useful AI than I could have imagined. GPT-4 exhibits impressive capabilities such as passing the bar exam, navigating virtual mazes, and even demonstrating what seems like a theory of mind. This brings us…
Denoising diffusion models for neuroscience
2022 was the year of generative AI models: DALL-E 2, MidJourney, Stable Diffusion, and Imagen all showed that it’s possible to generate grounded, photorealistic images. These generative AIs are instances of conditional denoising diffusion probabilistic models, or DDPMs. Despite these flashy applications, DDPMs have thus far had little impact on neuroscience. I want to change…
2022 in review: neuroAI comes of age
2022 has come and gone and we’ve learned a lot about how the brain is – or isn’t – like an artificial neural network. There isn’t a single journal or venue that focuses exclusively on neuroAI so the literature is spread across preprint servers, conference proceedings and the occasional prestige journal. As an end-of-year treat,…
How science startups actually work
It’s tough being an academic these days. Between the tiny number of permanent positions and the systemic exploitation of cheap labour, people are looking for a way out. Students, postdocs, and even PIs are asking themselves: should I join a startup? Should I make my own startup? It sounds seductive—change the world! Make money! Keep…
Large language models will change science
It takes dedication to keep up with the scientific literature. 2344 papers were accepted at last year’s NeurIPS conference. Who has time to read all of that? It’s difficult to see the forest from the trees as ever more research is published. What if every single one of us had infinite access to world experts…
What’s the endgame of neuroAI?
This blog post was adapted into an article for a16z’s Future. It’s been 60 years since Hubel and Wiesel first started unlocking the mysteries of the visual system. Proceeding one neuron at a time, they discovered the fundamental building blocks of vision, the simple and complex cells. Yet for a long time, neurons in high-level…
Unsupervised models of the brain
We’re in a golden age of merging AI and neuroscience. No longer tied to conventional publication venues with year-long turnaround times, our field is moving at record speed. 2021 saw a Cambrian explosion of research into unsupervised learning to explain brain representations, which is teaching us about how the brain might have evolved for sensing.…
The Good Research Code Handbook
When I was in grad school, I wrote a lot of code which ended up biting me later. It wasn’t until I got a job as a software engineer at Google that I discovered better ways of organizing code. Writing reusable, easy-to-read code is a core skill to have when you’re spending half of your…
Loading…
Something went wrong. Please refresh the page and/or try again.
xcorr: AI & neuro 