Maria Eckstein
@eckstein_maria
Followers
1K
Following
238
Statuses
91
follow me on @mariaeckstein.bsky.social Research Scientist @DeepMind, learning & decision making; PhD @UCBerkeley with @Anne_On_TW; https://t.co/uJV6g4qQR7
London, England
Joined March 2020
You'll get hands-on implementation experience (programming experience background is helpful, but enthusiasm is key!). If you're an undergrad, grad student, or postdoc interested in AI, cogsci, or neuroscience, this course is for you! 🧠🤖
0
0
2
A fantastic event, consider submitting an abstract!
📢📢Announcing this year's conference on the Mathematics of Neuroscience & AI (Split, Croatia, 27-30th May): Abstract submissions and early bird registration are open. The first 100 tickets can get an extra 20% off using discount code SUPEREARLYNM25. Keynote Speakers: Prof Jay McClelland (@StanfordHAI & @GoogleDeepMind) Prof Wolfgang Maass (@UniGraz) Prof Christine Constantinople (@nyuniversity ) Prof Jakob Foerster (@UniofOxford & @AIatMeta) Dr Ida Momennejad (@Microsoft) and more to come! Session Chairs: Dr @VFascianelli (@Columbia) Dr @framigna (@nyuniversity & @Princeton) @YasmineAyman20 (@Harvard) Dr @jcrwhittington (@UniofOxford & @Stanford) Prof @billdthompson (@UCBerkeley) Dr @eckstein_maria (@GoogleDeepMind) @Michael_Lepori (@BrownUniversity) Prof @sucholutsky (@nyuniversity) Location: The conference will be held at Croatia’s National Theatre in Split, 10 minutes walk from the ancient harbour. Reserve your spot now!
0
2
9
RT @HamidrezaJamal9: 🚀 Excited for the Physics of Cognition workshop at @BernsteinNeuro in Frankfurt this Sunday and Monday! Join us to exp…
0
4
0
RT @iancballard: The Computational and Cognitive Neural Sciences lab at UC Riverside is recruiting psychology PhD students to join our team…
0
19
0
RT @scychan_brains: @AndrewLampinen is one of my favorite people in the world to collaborate with, and anyone would be lucky to work with h…
0
7
0
@cogscikid @neuro_kim @nathanieldaw @summerfieldlab @kevinjmiller10 .. that incremental updating is not required. A model that calculates choice variables anew on each trial does just as well as one that keeps track of them and performs a small update on each trial.
0
0
0
@cogscikid @neuro_kim @nathanieldaw @summerfieldlab @kevinjmiller10 Great question, Wilka! The class of algorithms we are questioning are those that are based exclusively on incremental updating of choice variables (e.g., Q-values). Our models suggest that humans separate memory variables from decision variables (aka, logits). And they suggest..
0
0
0
@summerfieldlab @nathanieldaw @kevinjmiller10 If any of this sounds interesting, I am at #CCN2024 and would love to chat! Catch me during posters (B10) or at a coffee break :)
0
0
5
We can add or remove model functions and see which architecture fits best; we can add or remove inputs, outputs, or connections between them; and the networks will automatically find the functions that predict the observed behavior best.
0
0
8
RT @CogCompNeuro: For folks who aren't able to attend in-person, we are excited to be able to stream all talks this year on our youtube cha…
0
140
0
RT @BernsteinNeuro: Register for the #BernsteinConference by July 31 and secure the early bird fees! 👉 https://t.co…
0
9
0