PhD student at @DartmouthPBS. Understanding how biological and artificial networks learn and represent.
Hanover, NH
Joined July 2013
- Tweets134
- Following32
- Followers177
- Likes353
Pinned Tweet
What determines the content of replay? A key challenge for current theories is that on tasks that feature unbalanced experience, rats paradoxically replay the less-experienced trajectory (e.g. Carey et al. 2019, Gillespie et al. 2021). Check out our new preprint on the topic!
Hung-Tu Chen retweeted
all LLMs stumble on co-referenced updates → strong context interference.
We stream N updates to the same key and ask for the current value. With 46 concudxrrent keys, accuracy collapses across all tested SOTA
ICML 2025 workshop Paper: arxiv.org/abs/2506.08184
Hung-Tu Chen retweeted
🌟 Cai Lab @Nature paper alert! In new work led by @mysteriousjoe_, we find that rest periods after learning not only stabilize new memories BUT ALSO integrate new memories with older ones from days past! (1/9)
Read it here: nature.com/articles/s41586-0…
Hung-Tu Chen retweeted
I am recruiting PhD students for 2025! ⏰✨📝
My lab @BBSutdallas @UT_Dallas studies human cognitive neuroscience 🧠🧠🧠 in high-level perception and cognition using neuroimaging, naturalistic stimuli, computational methods, and AI models 🤖🤖🤖 in both typical and neuropsychological populations 💻💻💻!
🏷️Lab: labs.utdallas.edu/coconeurol…
📔PhD (DDL Dec 1): bbs.utdallas.edu/academics/g…
✅Application details: labs.utdallas.edu/coconeurol…
Hung-Tu Chen retweeted
Activity in the M2 predicts upcoming choices. But what does the activity represent? A gaze-centered plan? A world-centered goal? A specific movement? We asked rats to plan responses to different targets from different start position to find out! 1/5
jneurosci.org/content/44/37/…
Hung-Tu Chen retweeted
Proud to have the entire @FINNLab_neuro at #CCN2024! Here's where you can catch us:
First up, Katie O'Nell @katieoneuro has a talk in tomorrow morning's session (Thurs 9:30am) about a new way to quantify what makes conversation good (hint: think pink 🩷🎀💅)
Hung-Tu Chen retweeted
Excited to present new work w/ @AnnaSchapiro at #CCN2024! On Thursday, I will present results showing that, through meta-learning, properties resembling distinctions across the neocortex-hippocampus hierarchy emerge in ANNs [Poster B89] 2024.ccneuro.org/poster/?id=…
Excited to present this work at #CCN2024! On Friday, I will give a talk (Session IV) and present a poster (C93) titled: Paradoxical replay maintains balanced and robust representations of task structure (with @mattmizumi). Happy to chat more if you are around!
biorxiv.org/content/10.1101/… with @mattmizumi. We capitalized on prior computational work describing differences in learned task representations under “rich” vs. “lazy” learning (Flesch et al. 2022 etc.) to explore the relationship between experience, task representations, and replay.
H/t to Sagiv et al. (2024) and Zhou et al. (2024) who provide different, likely complementary perspectives on paradoxical replay, and show how it might fit into larger, more powerful theories of replay. Also thank everyone who provided feedbacks!
Hung-Tu Chen retweeted
How can we efficiently find task representations online that work well for the ongoing task while leveraging relations with other tasks? Our simulations show that episodic memory can help!
URL: shorturl.at/tT246
Here’s a summary:
(1/n)
Hung-Tu Chen retweeted
📢 We’re recruiting neuroscience postdocs! 🧠
The Cai Lab is on a mission to unravel the neural processes behind memory stability and flexibility across the lifetime. If you’re interested in working in these areas, please send an intro and CV to [email protected].
Hung-Tu Chen retweeted
Come check out our work on memory-updating in hippocampus. UU12, 13, 14 @SfNtweets. @htchen_neuro @austinbaggetta @mysteriousjoe_ @MiniscopeTeam @SinaiBrain – at Washington, DC
Hung-Tu Chen retweeted
Do you want to do a postdoc in London at the fabulous @SWC_Neuro with me and @AnnDuan2 ?
Do you have expertise in rodent behavior, neurophysiology and/or imaging? Interested in economics? decision-making? Multiplayer games? Let's chat in DC #sfn #neuroscience #postdoc
Hung-Tu Chen retweeted
Lab's first paper is accepted at ICLR!
How do you quantify whether two neural nets have similar representations *in the presence of noise*? We provide an answer that is able to capture important noise correlation structure.
Full tweet thread to come arxiv.org/abs/2211.11665