Leonardo Cotta
@cottascience
Followers
1K
Following
5K
Media
32
Statuses
1K
Postdoc Fellow @VectorInst. Machine Learning and Causal Inference. Schooled at: @purduecs @dcc_ufmg. He/Ele. From BH 🔺 🇧🇷
Toronto, Ontario
Joined July 2018
🗞️ Exciting news! This is now published in the proceedings of @royalsociety A (w/ open access). Make sure to check the original thread and drop me a line if you have any questions or feedback 😊.
⚠️ Most modern link prediction tasks are actually causal!.❓ What are the needed causal assumptions and estimators for this task?.💡 Our new work ( w/ @beabevi_, Nesreem, @brunofmr) shows that INVARIANCES ARE SUFFICIENT answers to both!.🧵1/10
0
12
40
this is empirical science at its best. identify failure modes and update your belief/strategy.
Hinton ~1983 thought Boltzmann Machines > backprop, but debugged himself out of the infatuation.Boltzmann Machines failed to learn, so he printed out weights, 8 cm thick, and inspected them for weeks.it's the dreaded local minimum.so 1 yr later, in desperation, he tried backprop
0
0
4
RT @phil_fradkin: The news is out! We're starting Blank Bio to build a computational toolkit assisted with RNA foundation models. If you wa….
0
25
0
RT @iclr_conf: ICLR 2026 will take place in. 📍Rio de Janeiro, Brazil.📅 April 23–27, 2026. Save the date - see you in Rio! #ICLR2026 https:….
0
99
0
RT @cgeorgiaw: Lots of hype around “AI scientists” but they have all been focused on the fun parts, not the hard parts. The focus has bee….
0
3
0
RT @HaonanDuan: When designing AI scientist benchmarks, the challenge lies in how to simulate realistic experimental data. We find that the….
0
3
0
RT @cjmaddison: What makes a great scientist? Most AI scientist benchmarks miss the key skill: designing and analyzing experiments. 🧪 We'r….
0
22
0
RT @CellStemCell: Online Now! Bacterial ADP-heptose triggers stem cell regeneration in the intestinal epithelium following injury https://t….
cell.com
Here, Goyal et al. identify that the ALPK1-TIFA innate immune signaling pathway regulates regeneration in the intestinal epithelium. Importantly, detection of ADP-heptose by ALPK1 triggers intestinal...
0
5
0
RT @bose_joey: 🎉Personal update: I'm thrilled to announce that I'm joining Imperial College London @imperialcollege as an Assistant Profess….
0
34
0
RT @ashVaswani: Check out our latest research on data. We're releasing 24T tokens of richly labelled web data. We found it very useful for….
0
84
0
RT @martoskreto: 🧵(1/6) Delighted to share our @icmlconf 2025 spotlight paper: the Feynman-Kac Correctors (FKCs) in Diffusion. Picture this….
0
41
0
great tutorial! these definitions can be directly extended to the different positional encodings used in transformer models for sequences too (changing the group action).
🚨Tutorial on Positional and Structural Graph Representations. Given at the National Institute for Theory and Mathematics in Biology. Thanks to @ClaireDonnat and Olga Klopp for the invitation!. @HaggaiMaron, the tutorial is finally out!.
0
0
4
RT @bwhite5290: Announcing Axiom: Eliminating drug toxicity, without using animals!. Alex Beatson and I founded Axiom a little over over a….
0
45
0
RT @Pavel_Asparagus: The problem is all virtue signaling — the city is built around getting a seed round from Sequoia so people focus on si….
0
33
0
0
52
0
this is pretty cool.
I’m so excited to share something we’ve been working on for awhile: North is @cohere’s AI workspace for enterprises. Today we’re releasing the platform for early access!
0
0
4
RT @brekelmaniac: I wrote a thing about "RL or control as Bayesian inference", which encompasses.- RLHF and controlled generation in LLMs.-….
0
105
0
RT @cjmaddison: Very glad to read this. Scientists already know how to represent biological, chemical, and physical concepts in a flexible,….
0
1
0
The principled way to combine diffusion models with a stellar team.
🧵(1/5) Have you ever wanted to combine different pre-trained diffusion models but don't have time or data to retrain a new, bigger model?. 🚀 Introducing SuperDiff 🦹♀️ – a principled method for efficiently combining multiple pre-trained diffusion models solely during inference!
1
1
12