Tokenized Graph Transformer (PyTorch)
https://github.com/jw9730/tokengt
This site is to serve as my note-book and to effectively communicate with my students and collaborators. Every now and then, a blog may be of interest to other researchers or teachers. Views in this blog are my own. All rights of research results and findings on this blog are reserved. See also http://youtube.com/c/hongqin @hongqin
https://arxiv.org/abs/2406.10833
digital patient
https://www.wiley.com/en-us/The+Digital+Patient%3A+Advancing+Healthcare%2C+Research%2C+and+Education-p-9781118952757
generative AI and digital patient
https://www.evms.edu/community/community_training/sentara_center_for_simulation_immersive_learning_a/
NIH Brain Initiative data sharing archives
NIH Brain Initiative data sharing
https://braininitiative.nih.gov/research/data-science-and-informatics

The NIH BRAIN Initiative is committed to advancing data science research, promoting effective data sharing and reuse, and building the informatics infrastructure that is needed to effectively archive, integrate, interpret, and visualize data across all areas of BRAIN Initiative research.
With broad data-sharing requirements in effect, the scale of data stored in BRAIN Initiative archives is rapidly expanding. This presents diverse challenges and opportunities for data science and informatics projects that promote data coordination, integration, interoperability, and reuse. Importantly, advancing the data ecosystems around the BRAIN Initiative—and neuroscience more broadly—will accelerate the building and testing of theories and models of brain function, leading to a deeper understanding of the brain.
The NIH Data Sharing and Management (DMS) policy (NOT-OD-21-013) recently went into effect for applications—including BRAIN Initiative applications— submitted on/after January 25, 2023. Learn more by visiting the NIH Scientific Data Sharing page.
The BRAIN Initiative's data sharing policy (NOT-MH-19-010) remains in effect, but new applications should follow DMS requirements where they differ. For example, data plans should be provided in the "Other plan(s)" section rather than as a Resource Sharing Plan in the application.
The use of BRAIN Initiative archives complies with DMS policy requirements regarding data preservation, sharing timelines, repository selection, validation, and standards.
Related Information
these work show brain processing sound and music with relevant to neural network modeling, --Hong
The Perception of Harmonic Sounds, Malinda McPherson
Time-dependent discrimination advantages for harmonic sounds suggest efficient coding for memory
https://www.pnas.org/doi/abs/10.1073/pnas.2008956117
Commonality and variation in mental representations of music revealed by a cross-cultural comparison of rhythm priors in 15 countries
https://www.nature.com/articles/s41562-023-01800-9
Decoding the dynamic representation of musical pitch from human brain activity
https://www.nature.com/articles/s41598-018-19222-3
"LLMs don't memorize things. They *grok*. They pattern match. But they can't replicate their input data."
It's an argument I hear daily.
It's a nice idea, well-reinforced and repeated by some specific groups.
The problem is that it's just not true.
Base models trained on next token loss minimization have been designed to memorize from the start.
It's not even a newly-discovered phenomenon, with results from Google dating back to 2022 (presented at ICLR 2023). As Carlini et al. report in their abstract:
"Memorization significantly grows as we increase (1) the capacity of a model, (2) the number of times an example has been duplicated, and (3) the number of tokens of context used to prompt the model."
It's never been a secret or a surprise that bigger models inevitably memorize popular content.
So next time someone tells you that models can't memorize, help teach them.
While not all models memorize all input data, this ability emerges surprisingly quickly and reliably as models scale or examples appear more frequently.
Here are a few more resources on the topic:
1. Emergent and Predictable Memorization in Large Language Models. Biderman et al., 2023. https://lnkd.in/dPthWVw8
2. Quantifying Memorization Across Neural Language Models. Carlini et al., 2022. https://lnkd.in/dZc_XDMH
3. Patronus CopyrightCatcher. https://lnkd.in/dWcGBM6k
4. leeky, open source library. https://lnkd.in/dvFPng3j