Geeling Chau

Congratulations to @YuanSui123 & @GeelingC for rocking out at the @evermind Genesis competition! 🚀 We are building some amazing AI Agents for Science tech at @caltech. @YueLabCaltech going strong! 💪

Excited to share our new work on cross-brain transfer of speech and handwriting BCIs! Some studies have required 10+ days of training data for peak BCI performance. Can pooling neural data from other users speed things up? 1/6 biorxiv.org/content/10.648…

Ecstatic to announce Merge Labs! 🧠🚀 I’m honored to work alongside my co-founders—Tyson Aflalo, @mikhailshapiro, @sandroherbig, @alexblania, and @sama on bridging biological and artificial intelligence. They’re *the* team to take on a mission this ambitious. We’re grateful to have partners like @OpenAI who share a long-term view on what it takes to build foundational technology. If you want to work on hard, foundational neurotech problems, come build with us.

A team of scientists led by Caltech and Cedars-Sinai has developed a new artificial intelligence framework that can accurately, quickly, and efficiently create virtual models of brain neurons. caltech.edu/about/news/mod…

NeurIPS 2025 Paper links: 1️⃣ arxiv.org/abs/2511.13022 2️⃣ arxiv.org/abs/2511.11940 3️⃣ openreview.net/forum?id=CdP8Y… If you're excited about how scaling models can help enable robust and generalizable wearable and neuro technology, let's connect + chat!! 🧠🦾

I'll be at #NeurIPS2025 presenting w/: 1️⃣ Eshani Patel and @yisongyue on generalizing models to input preprocessing 2️⃣ Chris Sandino et al., on new temporal reps of EEG 3️⃣ @shivashriganesh et al., on new iEEG training frameworks 📍Brain & Body FM on Sat 😍 🧠? Let's chat!! 🔗⬇️

Most of @YueLabCaltech will be at #NeurIPS2025! Check out papers by the group & collaborators spanning: 🔬 AI for Science — closing the loop from experiment design to inverse problems 🎯 End-to-end conformal risk control 🌊 New diffusion methods with theoretical guarantees 🤖 Advances in LLM reasoning & formal verification We're also developing a unified framework for scientific inverse problems — using foundation models to go from raw telescope/microscope/sensor data → scientific insights. Key results: -- Reconstructed M87 black hole from real EHT data (led by @klbouman) -- Plug-and-play priors + forward models across many domains - Latest release is InverseBench : devzhk.github.io/InverseBench/ (w/ @Hongkai_Zheng, @BingliangZhang, @WendaChu32619, @ZihuiRayWu, @zross_) Same algorithms also work for steering protein generative models with wet lab experimental data: github.com/jsunn-y/SGPO (w/ @jsunn_y, @WendaChu32619, @danielpkhalil, @francesarnold) Many talented students graduating soon and exploring next opportunities! Come chat at our posters 👇

My group at Caltech is recruiting: two postdocs (mechanical design of endovascular neural interfaces; cell-type-specific neural interfaces) + multiple Ph.D. positions. If you build at the interface of electronics, chemistry & neuroscience, let’s talk. 🔗 zlab.caltech.edu/openings

Excited to announce the Foundation Models for the Brain and Body workshop at #NeurIPS2025!🧠 We invite short papers or interactive demos on AI for neural, physiological or behavioral data. Submit by Aug 22 👉 brainbodyfm-workshop.github.io

We're excited to announce the Brain-to-Text '25 competition, with a new intracortical speech neuroscience dataset and $9,000 in prizes generously provided by @BlackrockNeuro_!. Can you do better than us at decoding speech-related neural activity into text? kaggle.com/competitions/b…

With over 500 delegates in attendance, the BCI Meeting was a tremendous success. A heartfelt thank you to all of you for the engaging discussions, meaningful exchanges, and the great energy throughout the event.

Our brain-to-voice synthesis brain-computer interface paper was published in @Nature today! This neuroprosthesis synthesized the voice of a man with ALS instantaneously, enabling him to ‘speak’ flexibly and modulate the prosody of his BCI-voice. 1/7 Paper: rdcu.be/eqH3C

📢 Heads-up to the neuro-ML/BCI community that we'll launch the Brain-to-Text '25 competition in ~a month, and it'll run through Dec. 31. More intracortical electrodes! More varied utterances! More... speechy? Rumor even has it there'll be cash prizes this year Stay tuned...

In #Banff for the @BCISociety #BCI2025 meeting! How can Machine Learning benefit BCIs? 🤝Would love to meet! (DM me) 📍Poster: Jun 4 (Wed) 4-6pm. P2-G-68 🗣Talk: Jun 5 (Thurs) Morning. W21: Large neural data models for BCIs. Organized by @_JoelYe🙌

Please share/amplify! Apply now: forestneurotech.org/jobs 📍 Location: Bay Area, CA

I'm excited to share our latest work — "Data Heterogeneity Limits the Scaling Effect of Pretraining in Neural Data Transformers", where we examined the effect of scaling up pretraining data in neural foundation models carefully.🧐 (1/9) Preprint: biorxiv.org/content/10.110…

Thoughts are now an input device. Today, @Apple announced its new BCI Human Interface Device (#HID) protocol—and Synchron is proud to be the first brain-computer interface company to achieve native integration with iPhone, iPad, and Apple Vision Pro. That means: 🧠 No touch. 🧠 No voice. 🧠 No movement. Just thought. Using Synchron’s brain-computer interface, people with paralysis can now control Apple devices directly with their minds. Apple is formally recognizing brain signals as a native input category—joining the ranks of touch, keyboard, and voice. That’s a paradigm shift not just in accessibility, but in how we think about interacting with technology. We're grateful to Apple’s accessibility team, our clinical collaborators, and—most of all—the patients leading the way. Read the full press release here: businesswire.com/news/home/2025… #Accessibility #BrainComputerInterface #Appl

We have released VARS-fUSI: Variable sampling for fast and efficient functional ultrasound imaging (fUSI) using neural operators. The first deep learning fUSI method to allow for different sampling durations and rates during training and inference. biorxiv.org/content/10.110… 1/

If you missed us, we'll be speaking at 4:06pm (Friday) in Garnet 216-218! Hope to see you there! :)

Come find us in Hall 3! Can’t wait to meet you and talk about modeling brain data across heterogeneous datasets!