🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Breaking News: Registration + Abstract submission for Montreal #AI & #Neuroscience (MAIN 2024) are now open ! 🎉🎉 Join us for an in-person meeting in Montreal October 22-25th ! Register now (limited space): 👉 main2024.org Abstract submission deadline: 15 Oct!

New preprint where we ask if the psychedelic-induced hallucinations can be explained by the role of dendrites in learning mechanisms in the brain. In short: classical psychedellics might hijack physiological gating mechanisms in generative learning. biorxiv.org/content/10.110…

Mila's annual supervision request process opens on October 15 to receive MSc and PhD applications for Fall 2025 admission! Join our community! More information here mila.quebec/en/prospective…

Last 24h to register for MAIN 2024 happening in Montreal next week! 👉 main2024.org Also check-out the Educational day on Oct 25th (which requires separate registration). 🧠 🤖 🥯 🐼 #Neuroscience #AI #Montreal

Marc's company is going places. If you're looking for a solid place to help build great things.... give this a look

1/5 Very excited to announce our new paper that formally defines compositionality, giving us theoretical insight into what it actually is and helping us build it into AI models. This was work with Thomas Jiralerspong, Yoshua Bengio, and Guillaume Lajoie. arxiv.org/abs/2410.14817…

In-context learnin (ICL) is one of the most exciting part of the LLM boom. Sequence models (not just LLMs) implement on-the-fly models conditionned on inputs w/o weight updates! Q: are in-context models better than «in-weights» ones? A: some times ICL is better than standard opt.

Compositional representations are a key attributes of intelligent systems that generalize well. An issue is that there is no robust way to quantify compositionality. Below is our attempt at such a quantifiable measurement. w/ Eric Elmoznino Thomas Jiralerspong Yoshua Bengio

Really excited to share our work on crosscoders, a generalization of sparse autoencoders that allows us to identify shared structure across layers or even across models, greatly simplifying our description of model representations. transformer-circuits.pub/2024/crosscode…

Why does #compneuro need new learning methods? ANN models are usually trained with Gradient Descent (GD), which violates biological realities like Dale’s law and log-normal weights. Here we describe a superior learning algorithm for comp neuro: Exponentiated Gradients (EG)! 1/12

Roman Pogodin Arna Ghosh Blake Richards Mila - Institut québécois d'IA I’m thrilled to announce I’ll be joining University of Leeds School of Computing to work on resource efficient AI and theories of biological intelligence! If you’re interested in work like this, I’ll be recruiting PhD students and postdocs. Please reach out or forward this on! 12/12

Long time coming. A very cool project that showcases the advantages of single neuron adaptation in RNNs. #PLOSCompBio: Neural networks with optimized single-neuron adaptation uncover biologically plausible regulari ... dx.plos.org/10.1371/journa… Props to Victor Geadah and co-authors

Talk by Guillaume Lajoie at the Montreal AI and Neuroscience (MAIN) Conference on credit assignment in neural networks without plasticity. #neuroscience #neuroAI #AI #compneuro youtube.com/watch?v=CvCqz5…

Recruiting phd/postdocs to mentor Undergrad Awardees!! This program brings UGs from across the 🌎 considering a career in neuro research Duties: meet students on 1st day, take them to a poster session, help them feel welcome. Apply: (very short!) form forms.gle/9hCniCA29ky59x…

As sequence models and in-context conditioning for inference are being developed to perform all kinds of ML tasks, we make systematic and tracktable evaluations to compare point v.s. distributional estimates . imo a key step to scale predictive modeling for general ML

La stimulation du nerf vague (VNS) représente une innovation prometteuse dans le traitement de maladies chroniques comme l'épilepsie ou la dépression, entre autres. Découvrez comment des chercheurs de Mila utilisent l’apprentissage automatique pour identifier en temps réel les

Recent progress on foundation-like models for neural data. As many use cases for generalizable models demand flexible online deployment, here we focus on a design enabling low latency real time use. We use hybrid SSM architecture & demonstrate various transfer learning scenarios