Federico Adolfi reposted
By-the-book analysis updated! • Better theorems • More experiments • Fewer bugs • Discussion section So very proud of this work
Who's Who of AI
Federico Adolfi
Why follow
Researcher with public evidence across AI research, Safety & security.
- AI signals
- 2 past 30d
- Sources
- 2 distinct domains
- Discussions
- 13 past 30d
- Latest signal
- 7d ago
Computation & Complexity | AI Interpretability | Meta-theory | Computational Cognitive Science
Max Planck Institute ESI
https://fedeadolfi.github.io
On the job market!
What they're sharing
Beyond Smoothed Analysis: Analyzing the Simplex Method by the Book arxiv.org
View on Bluesky →
Neuromodulation enhances the capability and efficiency of spiking neural networks | bioRxiv biorxiv.org
View on Bluesky →
From Empirical Problem-Solving to Theoretical Problem-Finding Perspectives on the Cognitive Sciences - Computational Brain & Behavior link.springer.com
Start - Georg-August-Universität Göttingen uni-goettingen.de
Articles & links
Federico Adolfi reposted
New preprint (well, very updated). 🤖🧠🧪 We find that an abstract model of neuromodulation lets spiking neural networks perform much better, particularly in challenging noisy environments, using less energy. Relevant to #neuroscience and #neuromorphic computing. 🧵👇 www.biorxiv.o…
This one👇🏻 because: 1) the cognitive and brain sciences and AI/ML have typically focused on problem solving and there’s not nearly as much discussion of how we find important problems or what makes a problem worthy of attention; 2) the paper could use more citations. link.spri…
Next week I'll be at University of Göttingen, Campus Institute Data Science. Come say hi ! On Thursday I'll give a talk in the CIDAS lecture series: on the integration of interpretability theory and experiment across AI and the cognitive sciences. www.uni-goettingen.de/en/6901…
Their own posts
Recent commentary
Doing formal/computational theory in a multidisciplinary intersection is wonderful because neuroscientists say it’s not biological enough, it’s too theoretical for AI folks, too applied for theoretical computer scientists, not empirical enough for cognitive scientists, & too formal for philosophers💫
Machine Learning books: and that’s why NETFLIX uses these POWERFUL TOOLS for RECOMMENDATION SYSTEMS Netflix: our personalized selection for you is 🥁🥁🥁 the show you’ve been watching all week we think you will like it
Reviewing the slide deck for my talk today: Right, so one slide to introduce them to interpretability, one slide to give them some background on complexity, one slide to bring them all and in the darkness bi — *delete delete* never mind got a bit carried away there
Gen Xer: *publishes single paper as tenured professor with chatbot as coauthor* hi fellow kids Gen Zer: *publishes dozens of AI generated papers before grad school without disclosure* sure ok
Not to be all ‘I told you so’ but arguing unconditionally that science is fixed and improved as we automate more and more of it is kind of backfiring with the whole “neuroscientists have 2 years to make a contribution and then they are all obsolete because AI has just graduated Summa Cum Claude”
Next week I’ll be at University of Geneva. I’ll give a talk on Wednesday on interpretability and complexity at the intersection of AI and the Cognitive/Brain Sciences. Happy to chat if you’re in the area!
Their network
In Federico Adolfi's orbit
Center = Federico Adolfi. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.