Lee Zamparo

156 Followers
125 Following
137 Posts
Senior scientist @recursionpharma. Ex @ServiceNowRSRCH, @sloan_kettering research scholar, @UofTCompSci PhD. Messages are my own.
So science Twitter seems severely degraded from my experience; have more people stopped using these types of sites altogether, or are they. posting here and I don't see them? #machinelearning #science #computationalbiology #structuralbiology
We have ongoing a special
issue of JMLR on machine learning for climate change. A lot of nice work I’ve seen in my editorial set so far, so please encourage more submissions and research in this important area. 🌍 https://jmlr.org/special_issues/climate_change.html
Journal of Machine Learning Research

Purportedly leaked internal Google document

"We Have No Moat, And Neither Does OpenAI"

https://www.semianalysis.com/p/google-we-have-no-moat-and-neither

Whether this is a real document or not, I believe the larger message that Google and OpenAI are running on borrowed time is legit. I've been saying as much.

Google "We Have No Moat, And Neither Does OpenAI"

Leaked Internal Google Document Claims Open Source AI Will Outcompete Google and OpenAI

SemiAnalysis
How do people raise issues of coding standards at work for packages that aren't your own? What are some strategies to motivate standardization that have worked?

Oh hey, March 17th is Stochastic Parrots Day. Join the @DAIR event!

https://www.eventbrite.com/e/stochastic-parrots-day-tickets-524219965027

I guess I'll be a panelist on cutting through the hype.

Stochastic Parrots Day

A virtual event to commemorate the 2nd anniversary of the paper, On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?

Eventbrite

MSFT lays off its responsible AI team

The thing that strikes me most about this story from @zoeschiffer and @caseynewton is the way in which the MSFT execs describe the urgency to move "AI models into the hands of customers"

https://www.platformer.news/p/microsoft-just-laid-off-one-of-its

>>

Microsoft just laid off one of its responsible AI teams

As the company accelerates its push into AI products, the ethics and society team is gone

Platformer

I actually predicted this last week!

Now that outages have massively increased, Elon Musk pronounces Twitter's old code is a mess that can't be saved -- and there will be a re-write:

https://twitter.com/elonmusk/status/1632810081497513993?cxt=HHwWkoCw1ceH9KgtAAAA

This is Stage 3 of Twitter's demise.

Now pay attention to the next stages.

It will be awhile before Twitter 3.0 will arrive. When it arrives, it won't be "real" Twitter.

See: https://mastodon.social/@atomicpoet/109949723816737406

See screenshots.

Elon Musk on Twitter

“@pmarca A small API change had massive ramifications. The code stack is extremely brittle for no good reason. Will ultimately need a complete rewrite.”

Twitter

If diffusion model sampling tries your patience, check out consistency models: single-step sampling! No adversarial loss!

In addition to being a very cool idea, this paper significantly leans on the formalism from Karras et al. 2022 AKA my favourite diffusion paper😁 Neat!
---
RT @_akhaliq
Consistency Models

achieve the new state-of-the-art FID of 3.55 on CIFAR10 and 6.20 on ImageNet 64 ˆ 64 for one-step
generation

abs: https://arxiv.o…
https://twitter.com/_akhaliq/status/1631470940310609921

AK on Twitter

“Consistency Models achieve the new state-of-the-art FID of 3.55 on CIFAR10 and 6.20 on ImageNet 64 ˆ 64 for one-step generation abs: https://t.co/oMhs20fxGM”

Twitter
Museum membership: totally worth it when you have kids. Almost doesn't matter which museum.

I wrote a public communications piece about #MachineLearning guided adaptive experiments: covering active learning, bandits, Bayesian optimization, and design of experiments. Most of the applications where I've managed to get it to work are in genomics.

https://researchoutreach.org/articles/adaptive-experiments-machine-learning-help-scientific-discovery/

Adaptive experiments: Machine learning helps scientific discovery

Cheng Soon Ong develops statistical machine learning methods and applies machine learning-guided design to biology and adaptive experiments.

Research Outreach