Mikhail Genkin

Exited to share our study where we find that encoding principles of cognitive variables are similar to those of sensory variables. biorxiv.org/content/10.110…

“Like a group of skiers descending a mountain, each [neuron] prefers a slightly different path, but all are shaped by the same slope,” says @Princeton’s @EngelTatiana on her lab’s new @Nature study of how the brain makes decisions. 📰: pni.princeton.edu/news/2025/all-…

Out today in @Nature: we show that individual neurons have diverse tuning to a decision variable computed by the entire population, revealing a unifying geometric principle for the encoding of sensory and dynamic cognitive variables. nature.com/articles/s4158…

@max_katz @grok, what do you think about it?

Хорошо ли Дональд Трамп руководит Америкой

@AndrewYNg Nuclear Engineering. If the same principles were applied to AI, every single AI developer would have to prove that the chance of terrible accident is less than X (e.g. 10^-7) per year before deploying their model.

In AI, the ratio of attention on hypothetical, future, forms of harm to actual, current, realized forms of harm seems out of whack. Many of the hypothetical forms of harm, like AI "taking over", are based on highly questionable hypotheses about what technology that does not currently exist might do. Every field should examine both future and current problems. But is there any other engineering discipline where this much attention is on hypothetical problems rather than actual problems?

A unifying perspective on neural manifolds and circuits for cognition — a Perspective by Christopher Langdon, Mikhail Genkin & Tatiana A. Engel @chrismlangdon @MGENK @EngelTatiana go.nature.com/3GHKOCV

What approach can explain the link between the brain and behavior? Shall we focus on neural circuits or manifolds? With @chrismlangdon & @MGENK, we provide a unifying perspective on manifolds and circuits, just out in @NatRevNeurosci: nature.com/articles/s4158…. #tweeprint 👇

ChatGPT is sometimes amazing, and sometimes hilariously wrong. Here, it confidently explains to me why an abacus is faster than a GPU. 😃

In decision-making tasks excitatory and inhibitory neurons show selectively for an upcoming choice. @EngelTatiana, @anne_churchland, and I explore the effects of inhibitory choice selectivity in decision making circuits using mean-field and RNN models. biorxiv.org/content/10.110…

Excited to share our bioRxiv preprint with @EngelTatiana. Cortical neurons show mixed selectivity for multiple task variables. We develop a latent circuit model for inferring interpretable circuit mechanisms from heterogeneous neural responses. biorxiv.org/content/10.110…

What advantages can neuro-inspired ANN architectures have for motor control / robotics? We translate C. elegans locomotion circuits into an ANN model controlling a simulated Swimmer agent. arxiv.org/abs/2201.05242 github.com/nikhilxb/ncap w/ @TonyZador, @EngelTatiana

Out today in @NatureComms the work led by @YanliangShi. We show that cortical state dynamics and selective attention define the spatial pattern of noise correlations. Many thanks to @SteinmetzNeuro, Tirin Moore, and @boahen_k for the great collaboration. nature.com/articles/s4146…

@SfNtweets Thanks, after very long waiting abd trying web/app for a video chat, I finally connected to find out that my discussion session is full…

🚧 #SfN21 Attendees: The virtual environment is currently experiencing an issue. We are working to resolve this ASAP. Apologies for the inconvenience.

@mikkolehtimaki @EngelTatiana @owenkhughes @NatureComms Currently, we work with a 1D latent space that only has 2 decision boundaries. To go beyond binary decisions, we will likely need to extend the framework for 2D space, which is one of our main future directions. In 2D one can model an arbitrary number of choices.

@MGENK @EngelTatiana @owenkhughes @NatureComms Alright, very interesting! Is it possible to extend this framework beyond binary decisions? You'd need multidimensional dynamics?

Excited to see this work with @MGENK and @owenkhughes out in @NatureComms today. The paper is open access: nature.com/articles/s4146… The code + tutorials are available on GitHub: github.com/engellab/neura…

@mikkolehtimaki @EngelTatiana @owenkhughes @NatureComms @mikkolehtimaki Thanks! We derive the expressions for the likelihood and its variational derivatives in terms of continuous functions. We then compute these expressions numerically in a Lagrange polynomials basis. Thus, we do not assume any parametric form for phi(x) a priori.

@EngelTatiana @MGENK @owenkhughes @NatureComms Beautiful paper! Seems like there are infinite possibilities for choosing the analytical form of phi(x). How did you narrow down the choices?

Our study on learning Langevin dynamics from noisy observations of non-stationary systems is published in Nat Commun. Our method infers the dynamics underlying decision-making from spiking activity. Paper: nature.com/articles/s4146…. Python package: github.com/engellab/neura…

Tomorrow, Sep. 24, 11am - 2 pm at @CogCompNeuro, @EngelTatiana will talk about flexible computational framework for modeling neural population dynamics during decision-making, and I will give a tutorial on our python package neuralflow: github.com/engellab/neura….

Join us tomorrow Sep 24 11am-2pm ET for the virtual @CogCompNeuro Keynote & Tutorial. I will talk about flexible models of neural population dynamics and decision making, and @MGENK will give a tutorial on our software package github.com/engellab/neura…. 2021.ccneuro.org/event.php?e=5

After the rough year of separation and isolation, everyone was looking forward to traveling to see our families. All Russians working in the US on a visa cannot do it. If they leave the US, they won’t be able to come back. Our parents cannot come either. No visas = more isolation