CBMM

[blog] Beneficial Misalignment: Why We Shouldn't Always Align AI to Humans In the rapidly evolving field of NeuroAI, a significant amount of energy is dedicated to 'alignment', the idea that representations from artificial intelligence should converge... poggio-lab.mit.edu/blogsupdates/b…

[blog post] A Conversation with Blaise Agüera y Arcas: On Intelligence, Life, and the Future of AI What does it mean to call something intelligent - and when did this question get so hard to answer? For Blaise Agüera y Arcas, VP at Google and founder... poggio-lab.mit.edu/blogsupdates/i…

[blog post] Can a Neural Network Think Before It Speaks? Somewhere around 2022, an observation started making the rounds among researchers working with large language models: if you just asked a model... poggio-lab.mit.edu/blogsupdates/c…

[blog post] Edge of (Stochastic) Stability made simple — Part II: the mini-batch case In Part I we had one landscape and a deterministic update. Now we have a distribution of mini-batch landscapes and a stochastic update... poggio-lab.mit.edu/blogsupdates/e…

[blog post] Edge of (Stochastic) Stability made simple — Part I: A crash course on (full-batch) Edge of Stability In this part I introduce the phenomenon and what I believe are the two key mechanisms—which we’ll use as the springboard for the mini-bat... poggio-lab.mit.edu/blogsupdates/e…

[blog post] Are Transformers Just "Stochastic Parrots"? A common criticism of Large Language Models (LLMs) is that they are merely "stochastic parrots"—statistical mimics that stitch together likely patterns without genuine reasoning... poggio-lab.mit.edu/blogsupdates/a…

🧵 New paper: LLM-ERM: Sample-Efficient Program Learning via LLM-Guided Search arxiv.org/abs/2510.14331 We use reasoning LLMs to learn tasks like IsPrime from ~200 samples by proposing short programs, making both the learned function *and* the learning process interpretable 🤯

Does SGD really “seek flat minima”? We show that SGD has no intrinsic preference for flatness, even for stable linear networks—going against ~10 years of folklore. Flatness emerges iff label noise is isotropic; anisotropic noise drives SGD to arbitrarily sharp solutions. This reveals a new flattening–sharpening mechanism in late training, unrelated to standard progressive sharpening or Edge-of-Stability effects.

Reinforcement Learning (RL) has long been the dominant method for fine-tuning, powering many state-of-the-art LLMs. Methods like PPO and GRPO explore in action space. But can we instead explore directly in parameter space? YES we can. We propose a scalable framework for full-parameter fine-tuning using Evolution Strategies (ES). By skipping gradients and optimizing directly in parameter space, ES achieves more accurate, efficient, and stable fine-tuning. Paper: arxiv.org/pdf/2509.24372 Code: github.com/VsonicV/es-fin…

[blog post] Intelligence Begins with Memory: From Reflexes to Attention Why associative memory is the oldest mechanism of intelligence—and still its computational core. sites.mit.edu/poggio-lab/int…

[blog post] Most Real Numbers Do Not Exist (And Why That Matters for Intelligence) The most useful mathematical objects are the ones that aren’t real at all... poggio-lab.mit.edu/most-real-numb…

[blog post] Genericity - Where compositionality is about structure, genericity is about geometry: the shape of the optimization landscape, the presence of gradients, and the existence of stable signals that guide learning. It answers one of the deep... sites.mit.edu/poggio-lab/the…

The First Principle: Nature Builds with LEGO Bricks Why can we understand a complex world? Because it is not a random mess — it is a hierarchy of reusable parts By Tomaso Poggio & Daniel Mitropolsky sites.mit.edu/poggio-lab/the…

Modern AI is like electricity in 1800: powerful artifacts, but no unifying theory. In post #1 of our new series, The Missing Foundations of Intelligence, Tomaso Poggio & @JohnGabrieli propose two candidate principles: Sparse compositionality & genericity. sites.mit.edu/poggio-lab/the…

The Center for Brains, Minds & Machines [2013-2025]

[video] "A Theory of Appropriateness with Applications to Generative Artificial Intelligence" Joel Leibo, senior staff research scientist at Google DeepMind and professor at King's College London cbmm.mit.edu/video/theory-a…

[video] Aligning deep networks with human vision will require novel neural architectures, data diets and training algorithms Thomas Serre, Brown University cbmm.mit.edu/video/aligning…

[video] "The success of CBMM and the people who made it possible" Over the last 11+ years, the CBMM has become a place of gathering brilliant minds to discuss and solve the challenges and questions of intelligence, organic and artificial. youtu.be/5y8fB0Cy6z8

[news] Tomaso Poggio autobiography, produced by NotebookLM cbmm.mit.edu/video/tomaso-p…

[video] "Conveying Tasks to Computers: How Machine Learning Can Help" Michael Littman, Brown University cbmm.mit.edu/video/conveyin…