Context Lab

Since you typically know more about your *own* past than your future, this affects what you say and do. And your "biased" knowledge of your own autobiographical timeline is transmitted to other people when you communicate!

From examining millions of real and fictional human conversations, we found that people refer to past events ~1.5 more often than future events, on average. This affects what other people *learn* (or infer) from our conversations.

Now out in @NatureComms (lead: @Xu_Xinming)! How do conversations shape how (and what) people guess about the unknown past and future (e.g., in other people's lives)? Paper/code/data: nature.com/articles/s4146…

Fun writeup of one of our group's recent papers (pnas.org/doi/10.1073/pn…)-- thanks @PsyPost!

We also did some digging to get at the specific networks that seem to exhibit this mode switching process most strongly (DMN!) or weakly (primary sensory cortex!) and also used some cool @neurosynth analyses to help us interpret the results.

When we are less engaged, our brain switches to a mode that prioritizes flexibility in favor of reliability.

New study out in @PNASNews (lead: @LucyLongOwen)! Our brains comprise billions of tiny components interact via trillions of connections, changing at sub-ms timescales. How is it so robust to injury and noise? Paper: doi.org/10.1073/pnas.2… Code/data: github.com/ContextLab/pca…

We've been doing some @LeetCode together to help hone our coding skills. We've set up a project for going through the daily problems: github.com/ContextLab/lee… We welcome participation from "outsiders" too; instructions are at the link above if you'd like to participate! 🧑‍💻

I'm feeling very happy, lucky, and grateful to have reached this next career milestone (tenure!) with the support of my family, lab, and colleagues! Special thanks to my current grad students @paxt0n4 and @Xu_Xinming for their incredible work, creativity, and general awesomeness!

We've (lead: @paxt0n4) added an awesome new system for automatically managing dependencies, that's isolated from the runtime environment. Davos now turns your notebook into a "fully enclosed" + shareable doc (w/ code/figures/data/text) that installs its own dependencies!

Excited to announce the latest release of our "Davos" Python package (v0.2.3)! 🎉🥳 Add Davos to a Jupyter notebook by inserting a new cell with this code at the top: %pip install davos import davos Then replace all uses of "import" with "smuggle" github.com/ContextLab/dav…

MIND is in the news today! 📰🗞️ home.dartmouth.edu/news/2023/08/v…

We're also grateful to @OpenAI, @huggingface, and @Meta, whose awesome LLM wizardry gives Chatify its brain powers 🧠💕🪄, and to the developers of @ProjectJupyter, LangChain, and GPTCache for providing a solid foundation for creating such neat tools!

Special shoutout to our core developers 🧑‍💻🧑‍💻🧑‍💻🧑‍💻: @HemuManju, @jeremyRmanning, @j6m8, and @KordingLab 🥳, and to all of our (many, many) pre-alpha testers 🧪🏅!

We're excited for the alpha release of our Chatify 🤖package in collaboration with @neuromatch! Chatify adds an LLM-powered "AI tutor" widget to Jupyter notebooks. Try it by adding this code to the top cell: !pip install chatify %load_ext chatify Link: github.com/ContextLab/cha…