Rob Nowak
@rdnowak
Followers
2K
Following
3K
Statuses
1K
Director of the Center for the Advancement of Progress
Ho-Chunk land
Joined April 2009
@DimitrisPapail It's not a bad model for the practice of one-dimensional curve fitting to data. And it is not a useful image for thinking about what is happening in high-dimensional problems.
0
0
6
RT @Kangwook_Lee: 🥲 Current image diffusion models struggle to generate complex images such as "a mustachioed squirrel holding an ax-shaped…
0
15
0
elated to announce that I also joined academic twitter 2016
1
0
9
Wow! Awesome. Thanks so much to the authors for pulling this valuable information together in a beautiful and accessible form.
This new textbook by @ml_angelopoulos Rina Barber and @stats_stephen looks really neat!
1
0
17
Laura is an amazing researcher. Excited about this talk!
The UW-Madison ECE is thrilled to host Associate Professor Laura Balzano from the University of Michigan for an insightful talk on finding linear structure in modern machine learning data. For full event details:
0
3
21
Looking forward to this talk!
I'm thrilled to announce that François Charton (@f_charton @AIatMeta) will be kicking off our new AI for Science seminar series next Wednesday. He is at the forefront of using AI for mathematics, cryptography, and theoretical physics. @datascience_uw
0
0
6
RT @byersalex: “And which great efficiency expert will lead this important new effort to streamline?” “Actually there will be two people i…
0
5K
0
Looks like an awesome workshop!
Two more weeks to submit your work on tensors/low-rank factorizations in the workshop
0
0
3
RT @Grigoris_c: Two more weeks to submit your work on tensors/low-rank factorizations in the workshop
0
2
0
@DimitrisPapail ok. this was a pic from WI, but not in the region it suggests, and other places are far away from WI. i think it must be using my profile to guess
0
0
6
@RishiSonthalia @lauralak @joanbruna @GittaKutyniok just saw this. it should be today, Friday Nov 8, AOE. That's 7am on Saturday, Nov 9 US Eastern time.
1
0
1
When it comes to putting AI systems into the field, there are known unknowns and unknown unknowns. Haoyue’s latest work tackles both!
AHA: Human-Assisted Out-of-Distribution Generalization and Detection #NeurIPS 2024 AHA strategically labels examples within a novel maximum disambiguation region to tackle both OOD generalization and detection. With just 2% additional human annotations, AHA achieves new SOTA🚀
1
2
21
Quite possibly my favorite application of active learning!
How do you tailor-make an LLM for your specific use case? It’s all in the data. Introducing SIEVE to customize your training data. Input any natural language filtering prompt, SIEVE retrieves all relevant data from web-scale datasets with high accuracy and at a super affordable price. 🧵on our latest preprint with @rdnowak:
1
0
13
Sounds like an awesome idea! I look forward to reading more.
Distillation is an important process, but why limit ourselves to distilling models into models, instead of into other objects? In new work from my group, we distill model capabilities into programs—a spotlight at #neurips2024.
0
0
14
We have a fantastic and growing AI/ML group!
🥳We have an open position in @UWMadisonECE, focusing on Trustworthy AI. Areas of interest include robustness, security, privacy, interpretability, fairness, and safety in machine learning and AI from systems and/or theoretical perspectives: Apply here!
0
0
24
Maybe you can't plan and reason your way to funny. Check out our latest work.
Sorry for raining on o1’s parade, but the more we train LLMs, the worse they get on generating and understanding captions for the New Yorker cartoons. 🧵on our latest study on humor in AI: •Large gap between AI & top human submissions •Dataset with 250M+ human ratings •New benchmark for AI humor generation •Insights on what it takes to win the New Yorker Cartoon Caption Contest. Collaboration with @stochasticlalit @guoyang328 and many others. Paper:
0
0
14