Sunny Duan
@sunnyyduan
PhD student at MIT studying computational neuroscience
📢 Excited to share that I will be presenting our paper on Neuro-Inspired SLAM at #ICLR2025 TOMORROW! 🗓 Saturday, April 26th 10:00 - 12:30 pm 📍 Hall 3 (Poster #55) jd730.github.io/projects/FARMa…
Excited to share our @TmlrOrg (#FeaturedCertification) paper on "Grid Cell-Inspired Fragmentation and Recall for Efficient Map Building"! Inspired by grid cells' remapping pattern in neuroscience, we propose a fragmentation-and-recall framework for spatial exploration.#gridcell
Excited to share our #ICLR 2025 paper on Breaking Neural Network Scaling Laws with Modularity! Joint work with Sunshine Jiang, William Yue, @jaedong_hwang, @abhiramiyer & @FieteGroup akhilanb.github.io/breaking-scali… 1/?
Nature research paper: Global modules robustly emerge from local interactions and smooth gradients go.nature.com/3EK6y2N
Highly recommend this great paper! One of the papers that got me interested in the @FieteGroup. It's a great example of the emergence of structure from simple, biologically plausible mechanisms.
1/ Our paper appeared in @Nature today! nature.com/articles/s4158… @FieteGroup @KhonaMikail. Explains emergence of multiple grid cell modules, w/ excellent match to data! Novel mechanism for applying across vast systems from development to ecosystems. 🧵👇
1/ Super-excited to share our new work “Episodic and associative memory from spatial scaffolds in the hippocampus”, out in Nature today! nature.com/articles/s4158… @FieteGroup @sugsharma @rischaudhuri. Key insights and ideas 👇#tweeprint
This paper validates the phenomena we observed when studying latent memorized sequences! Seems like a lot of information stored in the model weights are hiding beneath the surface. x.com/sunnyyduan/sta…
"Do Unlearning Methods Remove Information from Language Model Weights?" In our new paper, we investigate this question and find that at least 88% of the evaluation accuracy on unlearned data can be recovered using current unlearning methods.
My 2nd to last #neuroscience paper will appear @unireps !! 🧠🧠 Maximizing Neural Regression Scores May Not Identify Good Models of the Brain 🧠🧠 w/ @KhonaMikail @neurostrow @BrandoHablando @sanmikoyejo Answering a puzzle 2 years in the making openreview.net/forum?id=vbtj0… 1/12
Excited to share our #IJCAI2024 paper on quantifying inductive bias in machine learning models Towards Exact Computation of Inductive Bias Joint work with William Yue, @jaedong_hwang, @abhiramiyer & @FieteGroup arxiv.org/abs/2406.15941 1/?