Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly
Sign up
Appearance settings

Commit 4b61630

Browse files
committed
add rl notes
1 parent 85ca8fb commit 4b61630

File tree

5 files changed

+70
-19
lines changed

5 files changed

+70
-19
lines changed

‎_includes/00_about.html

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -16,7 +16,7 @@ <h1 class="brand-heading">&nbsp;</h1>
1616
</div>
1717
<div class="coll2">
1818
<p class="intro-text" style="margin-bottom: 20px;">
19-
seeking good explanations with machine learning
19+
seeking superhuman explanations
2020
</p>
2121
<ul class="list-inline social-buttons">
2222
<li><a href="https://scholar.google.com/citations?hl=en&user=vPYz4QwAAAAJ"><i

‎_notes/ai/decisions_rl.md

Lines changed: 1 addition & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -251,8 +251,7 @@ typora-copy-images-to: ../assets
251251
- like an approximation of ADP
252252
- when we transition $s \to s',ドル update $U^\pi(s) = U^\pi (s) + \alpha \left[R(s) - U^\pi (s) + \gamma \:U^\pi (s') \right]$
253253
- $\alpha$ should decrease over time to converge
254-
- *prioritized sweeping* - prefer adjustments to states whose likely successors have just undergone a large adjustment in their own utility estimates
255-
- speeds things up
254+
- *prioritized sweeping* - prefer adjustments to states whose likely successors have just undergone a large adjustment in their own utility estimates (speeds things up)
256255

257256
## active reinforcement learning
258257

‎_notes/neuro/comp_neuro.md

Lines changed: 14 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -969,11 +969,12 @@ subtitle: Diverse notes on various topics in computational neuro, data-driven ne
969969
- [non-human primate optogenetics datasets](https://osf.io/mknfu/)
970970
- [vision dsets](https://www.visualdata.io/)
971971
- MRNet: knee MRI diagnosis
972+
- [natural scenes dataset (NSD)](https://naturalscenesdataset.org/) - vision fMRI
973+
- NSD-Imagery: A benchmark dataset for extending fMRI vision decoding methods to mental imagery ([kneeland...kay, naselaris, 2025](https://arxiv.org/abs/2506.06898)) - participants memorized a handful of image stimuli and were asked to imagine a particular one
972974
- [datalad lots of stuff](http://datalad.org/datasets.html)
973975
- calcium imaging records in mice
974976

975977
- Recordings of ten thousand neurons in visual cortex during spontaneous behaviors ([stringer et al. 2018](https://figshare.com/articles/dataset/Recordings_of_ten_thousand_neurons_in_visual_cortex_during_spontaneous_behaviors/6163622)) - 10k neuron responses to 2800 images
976-
977978
- neuropixels probes
978979
- [10k neurons visual coding](https://portal.brain-map.org/explore/circuits/visual-coding-neuropixels) from allen institute
979980
- this probe has also been used in [macaques](https://www.cell.com/neuron/pdf/S0896-6273(19)30428-3.pdf)
@@ -1024,6 +1025,11 @@ subtitle: Diverse notes on various topics in computational neuro, data-driven ne
10241025
- model frequency bands
10251026
- EEG foundation model: Learning Topology-Agnostic EEG Representations with Geometry-Aware Modeling ([yi...dongsheng li, 2023](https://openreview.net/pdf?id=hiOUySN0ub))
10261027
- Strong Prediction: Language Model Surprisal Explains Multiple N400 Effects ([michaelov...coulson, 2024](https://direct.mit.edu/nol/article/5/1/107/115605/Strong-Prediction-Language-Model-Surprisal))
1028+
- datasets
1029+
- DEAP: A Database for Emotion Analysis ;Using Physiological Signals ([koelstra...ebrahimi, 2012](https://ieeexplore.ieee.org/abstract/document/5871728)) - 32-channel system
1030+
- SEED: Investigating Critical Frequency Bands and Channels for EEG-Based Emotion Recognition with Deep Neural Networks ([zheng & lu, 2015](https://ieeexplore.ieee.org/abstract/document/7104132)) - 64-channel system
1031+
- HBN-EEG dataset ([shirazi...makeig, 2024](https://www.biorxiv.org/content/10.1101/2024.10.03.615261v2)) - EEG recordings from over 3,000 participants across six distinct cognitive tasks
1032+
10271033

10281034
## cross-subject modeling
10291035

@@ -1116,6 +1122,7 @@ subtitle: Diverse notes on various topics in computational neuro, data-driven ne
11161122
- Aligning brain functions boosts the decoding of visual semantics in novel subjects ([thual...king, 2023](https://arxiv.org/abs/2312.06467)) - align across subjects before doing decoding
11171123
- A variational autoencoder provides novel, data-driven features that explain functional brain representations in a naturalistic navigation task ([cho, zhang, & gallant, 2023](https://jov.arvojournals.org/article.aspx?articleid=2792546))
11181124
- What's the Opposite of a Face? Finding Shared Decodable Concepts and their Negations in the Brain ([efird...fyshe, 2024](https://arxiv.org/abs/2405.17663)) - build clustering shared across subjects in CLIP space
1125+
- When compared to vision, brain activity patterns measured during mental imagery have much lower signal-to-noise ratios (SNR) ([roy...kay, naselaris, 2023](https://jov.arvojournals.org/article.aspx?articleid=2792335)), vary along fewer signal dimensions ([roy...kay, naselaris, 2024](https://2024.ccneuro.org/pdf/415_Paper_authored_tiasha_ccn2024_withauthors.pdf)), and encode imagined stimuli with expanded receptive fields and lower spatial frequency preferences, especially in early visual cortex ([breedlove...naselaris, 2020](https://www.cell.com/current-biology/fulltext/S0960-9822(20)30494-2?dgcid=raven_jbs_etoc_email))
11191126
- bmi
11201127
- Accelerated learning of a noninvasive human brain-computer interface via manifold geometry ([busch...turk-brown, 2025](https://www.biorxiv.org/content/10.1101/2025.03.29.646109v1)) - train subjects to control avatar navigation through fMRI, then perturb environment and evaluate decoder
11211128

@@ -1148,6 +1155,12 @@ subtitle: Diverse notes on various topics in computational neuro, data-driven ne
11481155
- Shared computational principles for language processing in humans and deep language models ([goldstein...hasson, 2022](https://www.nature.com/articles/s41593-022-01026-4)) - predict ECoG responses to podcasts from DL embeddings
11491156

11501157

1158+
# brain foundation models
1159+
1160+
- Brain Foundation Models: A Survey on Advancements in Neural Signal Processing and Brain Discovery ([zhou, liu...wen, 2025](https://arxiv.org/pdf/2503.00580))
1161+
- Brain-JEPA: Brain Dynamics Foundation Model with Gradient Positioning and Spatiotemporal Masking ([dong...zhou, 2024](https://proceedings.neurips.cc/paper_files/paper/2024/hash/9c3828adf1500f5de3c56f6550dfe43c-Abstract-Conference.html)) - fMRI modeling that uses positional embedding matrix based on brain gradient positioning + temporal encoding matrix using sine/cosine for temporal positioning
1162+
- Brant: Foundation Model for Intracranial Neural Signal ([zhang...li, 2023](https://proceedings.neurips.cc/paper_files/paper/2023/hash/535915d26859036410b0533804cee788-Abstract-Conference.html)) - predict iEEG with learnable position encoding
1163+
- LaBraM: Large Brain Model for Learning Generic Representations with Tremendous EEG Data in BCI ([jiang, zhou, lu, 2024](https://arxiv.org/abs/2405.18765)) - predict EEG with learnable temporal & spatial encoding matrix
11511164

11521165
# advanced topics
11531166

‎_notes/research_ovws/ovw_interp.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -424,7 +424,7 @@ Symbolic regression learns a symbolic expression for a function (e.g. a mathemat
424424
- Demystifying Black-box Models with Symbolic Metamodels ([alaa, van der schaar, 2019](https://papers.nips.cc/paper/9308-demystifying-black-box-models-with-symbolic-metamodels.pdf)) - distill black-box model with Meijer G-functions (rather than pre-specifying some forms, as is done with symbolic regression)
425425
- Symbolic Metamodels for Interpreting Black-boxes Using Primitive Functions ([abroshan...khalili, 2023](https://arxiv.org/abs/2302.04791)) - use GP approach
426426
- Neural Symbolic Regression using Control Variables ([chu...shao, 2023](https://arxiv.org/abs/2306.04718))
427-
- Discovering Symbolic Models from Deep Learning with Inductive Biases ([cranmer...ho, 2020](https://arxiv.org/abs/2006.11287)) - focused on GNNs
427+
- Discovering Symbolic Models from Deep Learning with Inductive Biases ([cranmer...ho, 2020](https://arxiv.org/abs/2006.11287)) - focused on GNNs, extracts equations from model weights
428428
- Providing Post-Hoc Symbolic Explanations for Sequential Decision-Making Problems with Black Box Simulators ([sreedharan et al. 2020](https://arxiv.org/abs/2002.01080))
429429
- neural networks
430430
- 2-step symbolic regr: first generate equation skeleton, then optimize constants with GD
@@ -433,7 +433,7 @@ Symbolic regression learns a symbolic expression for a function (e.g. a mathemat
433433
- Deep symbolic regression ([petersen...kim, 2021](https://arxiv.org/pdf/1912.04871.pdf)) - RL-based
434434
- End-to-End symbolic regression (still use final refinement step)
435435
- AI Feynman: A physics-inspired method for symbolic regression ([udresku & tegmark, 2020](https://www.science.org/doi/10.1126/sciadv.aay2631)) - use a loop with many if-then checks to decompose the equations
436-
- End-to-end symbolic regression with transformers ([kamienny...charton, 2022](https://arxiv.org/abs/2204.10532))
436+
- End-to-end symbolic regression with transformers ([kamienny...charton, 2022](https://arxiv.org/abs/2204.10532)) - explicitly train transformer from scratch to do the task
437437
- SymFormer ([vastl...babuska, 2022](https://arxiv.org/abs/2205.15764))
438438
- Deep Generative Symbolic Regression ([holt...van der schaar, 2023](https://openreview.net/forum?id=o7koEEMA1bR)) - use RL
439439
- Building and Evaluating Interpretable Models using Symbolic Regression and Generalized Additive Models ([sharif, 2017](https://openreview.net/pdf?id=BkgyvQzmW))

0 commit comments

Comments
(0)

AltStyle によって変換されたページ (->オリジナル) /