@BioAI_NeuralNet

@MIT trained Neuroscientist interested #BioAI, #NeuroAI, #Neuro_Aging 🧠⚡🤖 https://xitter.cf/t.co/mrNBAVydzT 🎯 [email protected]

 Houston, Texas, United States
Joined April 2023
  • Tweets3.3K
  • Following8.7K
  • Followers16.8K
429Media

Synthetic Data for BioAI and NeuroAI: From Mechanistic Simulators to Foundation Models

Synthetic data is only as good as the model that generates it, and in biology the best generators are mechanistic simulators whose parameters mean something. Foundation models for neural recordings, s…

Article · @BioAI_NeuralNet
3
6
2
28
5,720

How Neuroscience Gave Birth to Artificial Neural Networks and AI ?

Artificial intelligence is often narrated as a story of computation of transistors, algorithms, and ever-larger matrices of numbers. But its origin are from neuroscientific principles. The earliest a…

Article · @BioAI_NeuralNet
3
6
1
30
2,126

Beyond the Connectome: Toward a Multidimensional NeuroAI

Connectomics has produced wiring diagrams of increasing scale and resolution, from the 302-neuron nervous system of Caenorhabditis elegans to dense reconstructions of the adult Drosophila brain and gr…

Article · @BioAI_NeuralNet
1
9
1
35
14,020

World Models: A brief Review

In March 2018, David Ha and Jürgen Schmidhuber published a paper titled World Models (Ha & Schmidhuber, 2018), later published at NeurIPS under the more descriptive title "Recurrent World Models Faci…

Article · @BioAI_NeuralNet
3
8
1,246
The Transformer architecture and LLMs have transformed AI, but they may be only the beginning. What next? In this review, I explore emerging models, algorithms, and architectures that could define the post-Transformer era and shape the next gen of AI.

From LLMs to World Models: The Future of Artificial Intelligence

The past decade of AI has been defined by the Transformer and its most visible product, the Large Language Model. Transformers made unprecedented scaling possible and now power reasoning, coding, mult…

Article · @BioAI_NeuralNet
14
7
1
31
3,028

World Models in Biology: From Sequences to Simulated Cells

Over the past two years a specific architectural idea from reinforcement learning and robotics the “world model,” a learned system that predicts how an environment's state changes in response to …

Article · @BioAI_NeuralNet
2
2
2
19
8,127

From Molecules to Platforms: The Evolution of AI Pharma Deals

For much of the past decade, the promise of AI-driven drug discovery revolved around a simple narrative: artificial intelligence would identify novel molecules faster and more efficiently than traditi…

Article · @BioAI_NeuralNet
3
2
14
1,359

Data Diversity Bottlenecks in BioAI: Stylistic and Semantic Constraints

Abstract

Article · @BioAI_NeuralNet
3
6
1
17
3,364

AI Can Accelerate Intelligence But Understanding Remains Human

Generative artificial intelligence (AI) systems built on large language models (LLMs) now perform many cognitive operations once considered hallmarks of human intellectual work: drafting arguments, re…

Article · @BioAI_NeuralNet
2
6
10
1,193

Global Workspace Theory as a Blueprint for Biological and Artificial Intelligence

The emergence of large language models and multimodal foundation models has transformed artificial intelligence from a collection of narrow computational tools into increasingly general-purpose system…

Article · @BioAI_NeuralNet
4
1
9
2,086

The BioAI Data Economy

The rapid emergence of biological foundation models has transformed the landscape of artificial intelligence in the life sciences. Models trained on genomic sequences, protein structures, transcriptom…

Article · @BioAI_NeuralNet
7
20
2,883

From Neurons to Neural Networks: What Biology Can Teach Artificial Intelligence

The human brain is arguably the most sophisticated information-processing system known. Despite containing approximately 86 billion neurons connected through more than 100 trillion synapses, it operat…

Article · @BioAI_NeuralNet
1
7
1
14
1,819
Load more