AutoGluon

Fast and Accurate ML in 3 Lines of Code

#AutoML #opensource

[Nick Erickson @innixma]

🚀 AutoGluon 1.5 is out - pushing the frontier of tabular and time series learning!

🏆New SOTA on TabArena: +60 Elo, 2x faster training, 50% faster inference, 3% lower rel error vs 1.4
🧠New Models: RealTabPFN-2.5, TabDPT, Chronos-2
⏱️AutoGluon-TimeSeries: 80% win-rate vs 1.4

[Nick Erickson @innixma]

💥 Tabular ML just got a massive upgrade.

🚀 AutoGluon 1.4.0 is here — our biggest release yet.
🔥 Extreme Preset → biggest performance leap ever
🧠 5 new SOTA models
🤖 MLZero 1.0 (zero‑code multimodal AutoML)
📊 TabArena benchmark
⚡ Time series speedups

[Nick Erickson @innixma]

🚨 Join us tomorrow at the 1st ICML Workshop on Foundation Models for Structured Data!

📅 July 18 | ⏰ 9 AM - 5 PM
📍 West Ballroom D
🎉 70 accepted papers
🎤 8 spotlight orals
🌟 6 invited talks from top experts in tabular & timeseries data!
#ICML2025 #FMSD

[David Holzmüller @DHolzmueller]

I'm excited to see RealMLP on top of the full TabArena leaderboard!
The RealMLP version in TabArena uses an improved hyperparameter search space, which I designed on other benchmarks and (mostly) before I was invited as a co-author of TabArena. Here are some thoughts: 🧵

[Lennart Purucker @LennartPurucker]

🚨What is SOTA on tabular data, really? We are excited to announce 𝗧𝗮𝗯𝗔𝗿𝗲𝗻𝗮, a living benchmark for machine learning on IID tabular data with:

📊 an online leaderboard (submit!)
📑 carefully curated datasets
📈 strong tree-based, deep learning, and foundation models

🧵

[Nick Erickson @innixma]

With TabArena, we are able to ensemble the results of all methods together, strongly outperforming all previous approaches, including the state-of-the-art, AutoGluon. The next release of AutoGluon, we plan to incorporate methods such as RealMLP to achieve a new SOTA.

2/

[Nick Erickson @innixma]

We put all the best practices for tabular data benchmarking in one place to make TabArena, a living benchmark for tabular data. Compare with the very best implementations of state-of-the-art methods out of the box! Here are my thoughts: 🧵

[Nick Erickson @innixma]

Less than 2 weeks left to submit to the AutoML 2025 Non-archival track! We are looking for (1) journal/conference papers, (2) original 4-page papers, and (3) non-traditional content on the topic of AutoML. Details: https://t.co/v7YhfBpGUp

[Nick Erickson @innixma]

Deadline Extended to May 23rd AoE for the Foundation Models for Structured Data #ICML2025 workshop to better accomodate those with NeurIPS paper submissions! If you have a paper on structured FMs at NeurIPS, consider submitting to the FMSD workshop! https://t.co/GwnimFbpPb

[Mononito Goswami @MononitoGoswami]

🚨 Final countdown: Submissions for our @icmlconf '25 Workshop on Foundation Models for Structured Data close today (May 19) at 11:59 PM AoE ⏰.

We're excited to see your best work!

Submit now: https://t.co/tELBRPcO5E

[Nick Erickson @innixma]

Today is the last day to submit papers to the Foundation Models for Structured Data (FMSD) Workshop at ICML 2025! Deadline is May 19th, 11:59 PM AoE! Submit your short-form 4 page papers that delve into the emerging structured data foundation model field. https://t.co/GwnimFbpPb

[Nick Erickson @innixma]

🚀 AutoGluon 1.3 is out!

✅ Stability & usability improvements
🔥 AutoML Benchmark 2025: AutoGluon is SOTA in tabular ML
🧠 Huge boost to Multimodal with "Bag of Tricks"
📦 Major refactors, bug fixes, and 140+ commits from 20 contributors!

https://t.co/gKPmRDpO2n

[Nick Erickson @innixma]

📢 We are excited to announce "#FMSD: 1st Workshop on Foundation Models for Structured Data" has been accepted to #ICML 2025! Call for Papers: https://t.co/Zj2e9pXpQJ

[AutoML Seminar @AutomlSeminar]

Really great talk by Oleksandr from the @autogluon team on #Chronos, an new foundation model for #timeseries: https://t.co/1LoTD43PDq

[AutoML Conference @automl_conf]

The #AutoMLConf deadline was extended! We’re looking for groundbreaking contributions in #AutoML, NAS, LLMs, agent-based systems, hyperparameter optimization, benchmarking, algorithm selection, and more. Don’t miss your chance to showcase your work in #NYC!
📅 Deadline: March 31

[Nick Erickson @innixma]

📢 Exciting news! #AutoML2025 introduces a Non-Archival Track to spotlight engaging, impactful, and creative content!

Submit short papers, journal papers, interactive demos, videos, blog posts, surveys & more!
Deadline: June 16th AoE

Details 👉 https://t.co/v7YhfBpGUp #AutoML

[Nick Erickson @innixma]

‼️Now things are getting interesting :).

This is exceptional work that is going to push the tabular data field forward massively. We integrated TabPFN into AutoGluon starting in our v1.2 release with TabPFNMix, and I suspect this trend will continue with TabPFNv2!

Time to code.

[Nick Erickson @innixma]

🚀 AutoGluon 1.2 is here! Key highlights:
1. Tabular & TimeSeries: 70% win-rate over v1.1.
2. New Features: TabPFNMix foundation model, parallel fit, Chronos-Bolt (250x faster + accurate).
3. AutoGluon-Assistant: Zero-code ML with LLMs.
#AutoML https://t.co/AgMOSPGwlA

[Anirudh Dagar @gollum_here]

🚀 Become a Kaggle Grandmaster with Zero Code!

Excited to announce the release of AutoGluon Assistant (AG-A): An ML Assistant, powered by AutoGluon and LLMs, turn your natural language task descriptions into winning ML models.
~ Team @autogluon

https://t.co/UGjOv1FBzz

#AutoML

[Nick Erickson @innixma]

🥇🥈🥉AutoGluon was used in the 1st, 2nd, 3rd, and 4th place solutions to @kaggle's latest ML competition out of 2718 competitors! https://t.co/eQj5G3gG65