Lucas O. Souza

🧠 A story about the brain, told through self-supervised learning: Grid cells, known for mapping physical spaces, also appear to track *abstract spaces* like ideas and sounds. How? 🤔 Excited to share the first of 2 papers at #NeurIPS2024! With @SarthakChandra @FieteGroup. 1/n

.@spracklen and @lucasosouza will be speaking at @AIDevWorld next Tues (10/25) at 2pm PT. They will be talking about how to accelerate AI models and optimize hardware efficiency, unlocking order-of-magnitude speedups and energy savings. numenta.com/company/events…

Excited to be presenting our (@Numenta) recent research to Stanford's SystemX Alliance next week: systemx.stanford.edu/events/seminar… Accelerating your DL models on existing CPUs, GPUs by 100X...

👉Paper Review👈 How to avoid catastrophic forgetting in multi-task and continual learning? This paper takes inspiration from biology and develops Active Dendrites: context-sensitive gating mechanisms applied to the input signal. Check out the video: youtu.be/O_dJ31T01i8

Excited to share my discussion with @samcharrington on @twimlai. We discussed neuroscience, sparsity (and yes even dendrites!) and how that can lead to orders of magnitude improvements in AI. Sam was a gracious host, and asked great questions throughout. twimlai.com/go/562

New preprint alert! We introduce Complementary Sparsity, a novel technique that significantly improves the performance of dual sparse networks on existing #hardware. Results suggest that weight + activation #sparsity can be a potent combination for scaling future #AI models.

Excited to share details of one of our new preprints: "Avoiding Catastrophe: Active Dendrites Enable Multi-Task Learning in Dynamic Environments" Work with Abhiram Iyer, Karan Grewal, @akash_velu99, @lucasosouza, @jerem_forest #tweeprint👇🏽 (1/11) arxiv.org/abs/2201.00042

A step towards the brain’s efficiency: We've achieved a 10x parameter reduction in @GoogleAI's BERT with no loss of accuracy on the GLUE benchmark. Read Ben's blog: numenta.com/blog/2021/12/1… More below 👇 (1/6)

To have true artificial general intelligence, must we first understand the brain?  This session with Jeff Hawkins of @Numenta explores a neuroscience-based approach to AI that may lead to true machine intelligence. #EmTechMIT trib.al/OJlgDqW

My colleagues and I studied recent advances in AI for this year's #AI100Report, to be released 9/16! Join me and some of the report authors and organizers to discuss key findings at either of two sessions on 9/28. Registration is free! hai.stanford.edu/events/ai100

I believe the main target of a good education should be a love of learning, confidence in one's ability to learn anything, and knowledge of how to learn effectively -- how to approach new domains and new problems. Education is the transition to autonomy. Not a molding process.

@SubutaiAhmad @Numenta I might be biased😬 but I heard it is a great place to work, with an awesome team tackling some of the most interesting open challenges in AI

We have a fantastic new job opportunity at Numenta. Come join and help manage a great research team! Please apply if you are interested, or if you know anyone qualified, please pass the news along. @Numenta numenta.com/company/career…

Our Brains@Bay Meetup turns 2 this week! Our latest blog post gives a recap on its mission, past topics, and possible future ones. If you are interested in machine algorithms inspired by neuroscience research, this meetup may be for you! numenta.com/blog/2021/07/0…

Last chance to register (for free!) for to attend the neural network sparsity workshop taking place tomorrow and Friday at sparseneural.net. Join 700 (!) registrants, 62 poster presenters, 7 spotlights, 6 invited talks, 3 panels, and 1 tutorial. See you tomorrow!

Brains@Bay on grid cells just starting! We have a great lineup today, with Kim Stachenfeld (DeepMind), Marcus Lewis (Numenta), James Whittington and Tim Behrens (University of Oxford). Looking forward to it! bit.ly/3z8uKVl

Why is sparsity important in neural networks? Read our researcher @lucasosouza's exploration into #sparsity in these blog posts below: 🧠 The Case For Sparsity in Neural Networks Part 1: Pruning: numenta.com/blog/2019/08/3… 🧠 Part 2: Dynamic Sparsity: numenta.com/blog/2020/10/3…

Our Brains@Bay Meetup group is all about fostering the study and development of machine learning algorithms heavily inspired by neuroscience research 🧠 Stay tuned for news on our next meetup. In the meantime, you can browse through past recordings here: youtube.com/playlist?list=…

If you are into neuroscience or AI, join our Clubhouse event this Friday 5pm PST. @jenzhuscott will be talking to Jeff Hawkins and @SubutaiAhmad about the brain and how the Thousand Brains Theory applies to AI. See you there! joinclubhouse.com/event/ma4oaDJ7

Thank you to everyone who joined our Brains@Bay meet up yesterday, we hope you enjoyed it! In case you missed the fireside chat between @lucasosouza and Jeff Hawkins, you can catch the replay here: youtu.be/XK-Vksr3IT8