
Machine Learning scientist @amazon. Passed through @ververicaData, @pandoramusic, @spotify. Opinions and retweets my own.
Seattle, WA
Joined August 2014
- Tweets1.6K
- Following473
- Followers516
- Likes3.4K
Pinned Tweet
After a long and winding road, our work on uncertainty estimation for online regression forests is now published in @JmlrOrg ! Big thanks to my co-authors @gdfm7 and Henrik for sticking through and getting this over the finish line! jmlr.org/papers/v20/19-006.h…
Theodore Vasiloudis retweeted
Excited to be speaking at the Stanford Graph Learning workshop about GNNs for Spotify RecSys applications! The event is great with lots of interesting talks! Livestream here: youtu.be/GYW286H3SKw
Theodore Vasiloudis retweeted
A summer endeavor, developing MLC: the first open lecture series on ML compilation. Machine learning compilation is an emerging field for systematic optimization and deployment of AI workloads. Hope to share adventures and fun with the community mlc.ai/summer22/ 🚀
Theodore Vasiloudis retweeted
In our latest episode of “How to Fix the Internet,” comedian and host of @WTFpod @marcmaron and @ProducerMcD tell never-before-shared details of the battle to save podcasting from a patent troll, in a case that EFF took all the way to the Supreme Court. eff.org/deeplinks/2022/01/po…
Theodore Vasiloudis retweeted
ICYMI: The best science career advice we heard in 2021 from @alliekmiller, @DalianaLiu, @sneha_rajana, @crislopeslara, @thvasilo, and many others. Read the full article here: amazon.science/working-at-am…
Theodore Vasiloudis retweeted
I'm starting a series of blog posts on "optimization nuggets": short and beautiful proofs in optimization (let me know what you think!). First one shows that SGD converges exponentially fast to a neighborhood of the solution:
fa.bianp.net/blog/2021/expon…
This is a very well-written intro to the topic of automated reasoning, and glad to see the techniques making their way onto real products. Logic geeks will be happy to see this :)
To launch our new research area, @byroncook, head of the AWS Automated Reasoning group, wrote a "gentle introduction" to the field, with some simple code examples (including a two-line illustration of the halting problem). #reInvent #AutomatedReasoning amazon.science/blog/a-gentle…
Theodore Vasiloudis retweeted
Join Amazon Scholars Michael I. Jordan, @mkearnsupenn, and Amazon VP and distinguished scientist Bernhard Schölkopf, as they discuss the history of machine learning in the past decade, its social impacts, the role of causal reasoning, and more. Register: amzn.to/315vPkl
Theodore Vasiloudis retweeted
Excited to announce that we've finally released @XGBoostProject v1.0.0! This cannot happen without contributions from over 390 contributors and the support from the community. Thank you! #MachineLearning
Release notes: bit.ly/2P7MWsY
GitHub repo: bit.ly/2wzcRDl
Theodore Vasiloudis retweeted
Graph neural networks are driving lots of progress in machine learning by extending deep learning approaches to complex graph data and applications.
Let’s take a look at a few methods ↓
Theodore Vasiloudis retweeted
What if I told you that fine-tuning T5-Large (0.8B params) on a couple hundred examples could outperform GPT-3 (175B params) on a bunch of tasks?
I’m thrilled to share our new work “A Few More Examples May Be Worth *Billions* of Parameters”.
Joint w/ @PSH_Lewis @riedelcastro @omerlevy_
Paper: arxiv.org/abs/2110.04374
1/4
Theodore Vasiloudis retweeted
I am excited to share my latest work: 8-bit optimizers – a replacement for regular optimizers. Faster 🚀, 75% less memory 🪶, same performance📈, no hyperparam tuning needed 🔢. 🧵/n
Paper: arxiv.org/abs/2110.02861
Library: github.com/facebookresearch/…
Video: youtube.com/watch?v=IxrlHAJt…
Theodore Vasiloudis retweeted
Happy to see @ApacheArrow performance and memory use benefits coming to @XGBoostProject github.com/dmlc/xgboost/pull…
Theodore Vasiloudis retweeted
I'm delighted to share that our paper "Fairness in Ranking under Uncertainty" has been accepted to #NeurIPS2021. This is joint work with my advisor Thorsten Joachims and David Kempe from USC. Looking forward to sharing the camera ready version soon! 📄

























