AI Briefing#neuroscience
The latest AI and developer news about #neuroscience, with the original source and a short summary.
The latest AI and developer news about #neuroscience, with the original source and a short summary.
Jakob Macke, Daniel Gedon, and Stefan Wahl argue that AI tools for science should return the full distribution of compatible models rather than a single optimal fit to reveal what data actually constrains.
The third lesson in the AI Quests series, 'Polar Peaks,' teaches students ages 11 to 14 data cleaning and model testing through neuroscience, with over 400,000 quests completed to date.
The system cited over 700 references compared to 11 for a single agent, with human experts preferring the distributed results 75% of the time.
The 218-million-parameter model predicts fMRI responses at 20,484 cortical points every 1.49 seconds without a scanner.
A simulation replicating the entire fruit fly brain has been revealed, demonstrating its ability to learn Super Mario Bros. 3 in real time.
TBC announced a partnership with AWS to commercialize an AI video model trained on neurons, claiming an 80% reduction in inference costs.
Google Research and HHMI Janelia researchers have released the first complete brain map, using AI to fully map the 166,000 neuronal connections in the male fruit fly brain.
It has been revealed that differences in evaluation resolution are the primary cause of the phenomenon where untrained CNNs exhibit higher similarity in V1 than trained CNNs.
The MVICAD2 model is proposed to account for both delay and dilation in order to reflect individual differences in brain signal analysis.
Zyphra has released ZUNA1.1, an open foundation model under Apache 2.0 that restores, denoises, and upsamples EEG signals from arbitrary channel layouts.
Introduces NEvo, a technique that uses an evolutionary algorithm to generate videos that maximally activate specific brain regions.
This introduces a technology that uses AI to generate videos that maximize the activity of specific brain regions and map visual selectivity.
Meta has unveiled Brain2Qwerty v2, which converts brain signals into text using a non-invasive method.
Meta has unveiled Brain2Qwerty v2, which combines non-invasive brainwave measurement (MEG) with LLMs to improve real-time sentence decoding performance.
Microsoft Research unveiled the GCT framework, which uses LLMs to explain and verify how the brain processes language.
A predictive learning mechanism operating at the neurochemical level was implemented with the Axon framework, demonstrating its performance.
Backpropagation sharply reduces brain alignment in the V1 region immediately after training, while local learning rules such as predictive coding preserve it.
By comparing visual cortex data from humans and macaques, the study analyzed the alignment performance of various learning rules.
TRIBE v2, trained on 1,000 hours and 720 subjects of fMRI data, predicts brain activity from video, audio, and language.
It asks whether AI is truly intelligent by examining the differences between LLMs and the brain.
OpenAI is participating in the seed round of Merge Labs, a brain-computer interface (BCI) research company.