Gideon (@gidsy) profile picture

Co-founder @cerenaut_ai (AI+Neuroscience) & @atidia_health (improving health outcomes), Affogato lover, Builder of stuff

Melbourne
Joined March 2008
Photo by Gideon (@gidsy)Photo by Gideon (@gidsy)Photo by Gideon (@gidsy)Photo by Gideon (@gidsy)Photo by Gideon (@gidsy)Photo by Gideon (@gidsy)Photo by Gideon (@gidsy)Photo by Gideon (@gidsy)Photo by Gideon (@gidsy)Photo by Gideon (@gidsy)
All bilaterally symmetric animals on earth have brains with 2 hemispheres. Why, and could it be an important principle for AI/ML? We explored these questions in: arxiv.org/abs/2209.06862 1/n
1
7
1
21
Hope to see you on Bluesky @gidsy.bsky.social
29
Replying to @fchollet
I have just posted this in another thread, but I did exactly that - added a 2d positional embedding to NanoGPT - and it works OK on ARC tasks (see medium.com/@causalwizard/arc… )
1
3
18
1,577
Introducing PaperQA2, the first AI agent that conducts entire scientific literature reviews on its own. PaperQA2 is also the first agent to beat PhD and Postdoc-level biology researchers on multiple literature research tasks, as measured both by accuracy on objective benchmarks and assessments by human experts. We are publishing a paper and open-sourcing the code. This is the first example of AI agents exceeding human performance on a major portion of scientific research, and will be a game-changer for the way humans interact with the scientific literature. Paper and code are below, and congratulations in particular to @m_skarlinski, @SamCox822, @jonmlaurent, James Braza, @MichaelaThinks, @mjhammerling, @493Raghava, @andrewwhite01, and others who pulled this off. 1/
77
724
159
3,284
556,193
Super excited re our publication on ‘AI surgical risk forecasting and prevention’ in the prestigious journal BMC. tinyurl.com/36s2xcat
1
2
6
81
But current risk calculators are: - inferior to ML solutions, - cumbersome and under-utilised, and - do not leverage the available data.
1
39
Our methodology takes the patient record and trains models to predict the following outcomes, with interpretable and actionable insights: - length-of-stay, - complications, - readmission, and - mortality
36
An interesting debate is happening about how you can prove causation:
People dunking on this too quickly, "correlation is not causation", "we don't vote count studies", etc. *Not every important question can be answered with a perfect RCT* For decades, the cigarette-cancer link was similarly correlational, based on case studies, etc. 🧵 1/N
2
1
50
PhD position: I’m looking for a PhD candidate to lead a project on how modulating variability can improve motor learning in my lab at Tel Aviv University. The position will be part of the TReND doctoral network with 11 other PhD students from around Europe movementsciences.sites.tau.a…
1
12
2
20
2,505
Join us as a PhD candidate! Explore how brain patterns follow intrinsic biological cycles using brain stimulation, cognitive paradigms & neuroimaging. Join our multidisciplinary team with @pjkaroly @NaimFeil @cashmachine15 @AndrewZalesky More details at: findanexpert.unimelb.edu.au/…
7
3
15
4,211
By popular request, Two-Way Fixed-Effects models, Difference-in-Differences, and Panel Data support will be enabled in causalwizard.app when this tweet is 5 days old. #buildinpublic #epitwitter #causality #causaltwitter
2
1
1
67
"Columbia Untisemity": Israel's No. 1 Satire program mocks the support for Hamas on college campuses @mulisegev @Eretz_Nehederet
2,834
10,159
6,743
31,840
17,799,495
This is really important. Instead of grappling with #causality people pretend it's not possible with historical data (it is!) and then happily (incorrectly!) draw causal conclusions from associative methods! It's dead easy to grapple with causal methods, stop making excuses :)
You tell folks that correlation isn’t causation but this study shows people just keep assuming that it is, even when researchers use careful wording. “We show that people do in fact infer causality from statements of association, under minimal conditions” gershmanlab.com/pubs/Gershma…
1
3
73
New paper! We meta-train plastic networks to do transitive inference (learn A>B & B>C, infer A>C) Training discovers a new neural learning algorithm, that reinstates past items in memory to quickly reconfigure existing knowledge upon learning a new item biorxiv.org/content/10.1101/…
1
12
2
58
19,536
62% of Israelis are either totally against the Netanyahu government’s anti-Supreme Court legislation (33) or only prepared for it to pass with the opposition’s agreement (29). Only 22% are in favor of it passing unilaterally.
2
6
313
This short thread clearly explains some profound features of the do-Calculus. Worth a quick read. #Causalinference
This Post is from an account that no longer exists. Learn more
1
2
59
ball of a pen rolling over paper
5
54
3
422
35,533
חייבים לעצור את אלימות המתנחלים שמשתוללת כבר כל יום בגדה המערבית. חייבים להילחם בהפיכה במקור שלה: ההתנחלויות והכיבוש. בואו איתנו לצעוד עם תושבי הכפר בורקה אל חומש, המאחז שגנב להם אדמות, ששרף את רכושם ושמפורר את המדינה שלנו. לפרטים והרשמה להסעות מכל הארץ👇
11
15
2
85
3,139