Artwork

Player FM - Internet Radio Done Right

1,113 subscribers

Checked 11h ago
Added nine years ago
Content provided by Jon Krohn. All podcast content including episodes, graphics, and podcast descriptions are uploaded and provided directly by Jon Krohn or their podcast platform partner. If you believe someone is using your copyrighted work without your permission, you can follow the process outlined here https://ppacc.player.fm/legal.
Player FM - Podcast App
Go offline with the Player FM app!
icon Daily Deals

863: TabPFN: Deep Learning for Tabular Data (That Actually Works!), with Prof. Frank Hutter

1:06:06
 
Share
 

Manage episode 467249047 series 1278026
Content provided by Jon Krohn. All podcast content including episodes, graphics, and podcast descriptions are uploaded and provided directly by Jon Krohn or their podcast platform partner. If you believe someone is using your copyrighted work without your permission, you can follow the process outlined here https://ppacc.player.fm/legal.

Jon Krohn talks tabular data with Frank Hutter, Professor of Artificial Intelligence at Universität Freiburg in Germany. Despite the great steps that deep learning has made in analysing images, audio, and natural language, tabular data has remained its insurmountable obstacle. In this episode, Frank Hutter details the path he has found around this obstacle even with limited data by using a ground-breaking transformer architecture. Named TabPFN, this approach is vastly outperforming other architectures, as testified by a write up of TabPFN’s capabilities in Nature. Frank talks about his work on version 2 of TabPFN, the architecture’s cross-industry applicability, and how TabPFN is able to return accurate results with synthetic data.

This episode is brought to you by ODSC, the Open Data Science Conference. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

In this episode you will learn:

  • (05:57) All about the TabPFN architecture
  • (21:27) Use cases for Bayesian inference
  • (35:07) On getting published in Nature
  • (44:03) How TabPFN handles time series data
  • (51:52) All about Prior Labs

Additional materials: www.superdatascience.com/863

  continue reading

1192 episodes

Artwork
iconShare
 
Manage episode 467249047 series 1278026
Content provided by Jon Krohn. All podcast content including episodes, graphics, and podcast descriptions are uploaded and provided directly by Jon Krohn or their podcast platform partner. If you believe someone is using your copyrighted work without your permission, you can follow the process outlined here https://ppacc.player.fm/legal.

Jon Krohn talks tabular data with Frank Hutter, Professor of Artificial Intelligence at Universität Freiburg in Germany. Despite the great steps that deep learning has made in analysing images, audio, and natural language, tabular data has remained its insurmountable obstacle. In this episode, Frank Hutter details the path he has found around this obstacle even with limited data by using a ground-breaking transformer architecture. Named TabPFN, this approach is vastly outperforming other architectures, as testified by a write up of TabPFN’s capabilities in Nature. Frank talks about his work on version 2 of TabPFN, the architecture’s cross-industry applicability, and how TabPFN is able to return accurate results with synthetic data.

This episode is brought to you by ODSC, the Open Data Science Conference. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

In this episode you will learn:

  • (05:57) All about the TabPFN architecture
  • (21:27) Use cases for Bayesian inference
  • (35:07) On getting published in Nature
  • (44:03) How TabPFN handles time series data
  • (51:52) All about Prior Labs

Additional materials: www.superdatascience.com/863

  continue reading

1192 episodes

All episodes

×
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
Returning after the “Super Bowl of AI”, NVIDIA GTC, Sama Bali and Logan Lawler talk to Jon Krohn about their respective work at tech giants NVIDIA and Dell. Sama and Logan discuss the next-gen Blackwell GPUs to their collaboration with Dell in launching Pro-Max PCs specially designed to take on heavy computational workloads as well as the incredible performance of GB 10 and GB 300 workstations, and the widening accessibility of AI developer tools and models. Additional materials: www.superdatascience.com/883 This episode is brought to you by ODSC, the Open Data Science Conference , by Adverity, the conversational analytics platform and by the Dell AI Factory with NVIDIA . Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (07:29) About Dell’s Pro Max PCs (14:01) Why having a Blackwell GPU from Nvidia is a great option for those new to training and deploying AI models (36:47) When it makes sense for a data scientist to switch from a Unix to a Windows based system (46:33) Logan’s and Sama’s predictions for AI…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
This week’s five-minute Friday heads to the Netherlands to find out more about Dutch company ASML, the brains behind the lithography machines that build AI chips. Jon Krohn walks through how ASML came to dominate the market, where they’re headed next, and how ASML’s complex machines shape AI chips as well as the very future of AI. Additional materials: www.superdatascience.com/882 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
Emily Webber speaks to Jon Krohn about her work at Amazon Web Services, from its Annapurna Labs-developed Nitro System, a foundational technology that can enhance securities and performance in the cloud and how Trainium2 became AWS’ most powerful AI chip with four times the compute of Trainium. Hear the specs of AWS’s chips and when to use them. Additional materials: www.superdatascience.com/881 This episode is brought to you by ODSC , the Open Data Science Conference . Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (08:36) Emily’s work on AWS’ SageMaker and Trainium (23:54) How AWS Neuron lets builders tailor their approach to using frameworks (29:07) Why using an accelerator is better than using a GPU (35:29) The key differences between AWS Trainium and AWS Trainium2 (52:45) How to select between AWS Trainium and AWS Trainium2…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
First developed in China, Manus AI and DeepSeek have made great waves on an international scale. Sought-after for their cost-effectiveness compared to US-made tech, Manus AI and DeepSeek are quickly becoming dominant technologies inside the country. In this five-minute Friday, Jon Krohn asks: Do these technologies warrant the huge amount of resources spent on them by multiple industries in China, and what makes hype become a mainstay? Additional materials: www.superdatascience.com/880 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.…
 
Greg Michaelson speaks to Jon Krohn about the latest developments at Zerve, an operating system for developing and delivering data and AI products, including a revolutionary feature allowing users to run multiple parts of a program’s code at once and without extra costs. You’ll also hear why LLMs might spell trouble for SaaS companies, Greg’s ‘good-cop, bad-cop’ routine that improves LLM responses, and how RAG (retrieval-augmented generation) can be deployed to create even more powerful AI applications. Additional materials: www.superdatascience.com/879 This episode is brought to you by Trainium2, the latest AI chip from AWS and by the Dell AI Factory with NVIDIA . Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (04:00) Zerve’s latest features (35:26) How Zerve’s built-in API builder and GPU manager lowers barriers to entry (40:54) How to get started with Zerve (41:49) Will LLMs make SaaS companies redundant? (52:29) How to create fairer and more transparent AI systems (56:07) The future of software developer workflows…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
AI stacks, AGI, training neural networks, and AI authenticity: Jon Krohn rounds up his interviews from March with this episode of “In Case You Missed It”. In his favorite clips from the month, he speaks to Andriy Burkov (Episode 867), Natalie Monbiot (Episode 873), Richmond Alake (Episode 871) and Varun Godbole (Episode 869). Additional materials: www.superdatascience.com/878 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
NPUs, AIPC, and Dell’s growing suite of AI products: Shirish Gupta speaks to Jon Krohn about neural processing units and what makes them a go-to tool for AI inference workloads, reasons to move your workloads from the cloud and to your local devices, what the mnemonic AIPC stands for and why it will soon be on everyone’s lips, and he offers a special intro to Dell’s new Pro-AI Studio Toolkit. Hear about several real-world AIPC applications run by Dell’s clients, from detecting manufacturing defects to improving efficiencies for first responders, massively supporting actual life-or-death situations. Additional materials: www.superdatascience.com/877 This episode is brought to you by ODSC , the Open Data Science Conference . Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (03:28) What neural processing units (NPUs) are (23:53) About Dell Pro AI Studio (35:03) Use cases for Dell Pro AI Studio (45:16) How AI development workflows and applications will change (49:01) About Dell’s AI factory ecosystem…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
Small, simple, accessible: Hugging Face makes a huge contribution to the agentic AI wave with its smolagents. Jon Krohn explores how this small-but-mighty new Python library can act as the best personal assistant you never had. Hear about its features and use cases in this five-minute Friday. Additional materials: www.superdatascience.com/876 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
Why are semiconductors so essential in this digital age, and how are they made? Jon Krohn speaks to electronics CEO Kai Beckmann about Merck KGaA, Darmstadt, Germany’s intricate manufacturing process, how we can use AI to develop materials that power next-gen AI technologies, and how a chip with the processing power of the human brain might one day be able to run on the power of a low-watt light bulb. Additional materials: www.superdatascience.com/875 This episode is brought to you by the Dell AI Factory with NVIDIA . Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (06:26) How Merck KGaA, Darmstadt, Germany supports groundbreaking developments in AI (13:42) Material science’s biggest challenges for AI (29:55) What heterogeneous integration is (34:37) How optical tech influences the electronics industry (49:04) Navigating upturns and downturns in the semiconductor industry (53:08) How AI regulations benefit humanity…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
In this Five-Minute Friday, Jon Krohn talks baseball. For decades, coaches have relied on player performance stats to make in-game decisions and refine their season strategies. Now, AI led by Statcast is taking baseball strategy even further, massively broadening analytics data to include pitch, swing and catch trajectories, spin rates, biomechanical information, player matchups, and how to enhance player performances. Listen to the episode to find out what other industries can learn from the “data-friendly” sport of baseball. Additional materials: www.superdatascience.com/874 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
Natalie Monbiot is an independent advisor and collaborator for projects that concern the “virtual human”, and she is “going all in on the virtual human economy”. Jon Krohn speaks to Natalie about these new ventures, how to mitigate the divide between AI users and nonusers, and how anyone can collaborate with AI without compromising their own creativity. Additional materials: www.superdatascience.com/873 This episode is brought to you by the Dell AI Factory with NVIDIA , by Trainium2, the latest AI chip from AWS and by ODSC, the Open Data Science Conference . Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (07:21) Natalie’s influences for her work (18:30) Will machines surpass human intelligence? (29:08) Using LLMs as collaborators and partners (40:15) How platforms demand user engagement and time (56:54) Natalie Monbiot at Wizly…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
In this five-minute Friday, Jon Krohn looks into Microsoft’s recent release of Majorana 1, a new quantum processing unit that uses topological qubits, a step away from the fragile qubits currently in use. Get Jon’s thoughts about this “transistor for the quantum age”, potential applications for quantum computing, and why this marks an exciting future for data science and machine learning practitioners. Additional materials: www.superdatascience.com/8 72 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
Agentic AI, AI success strategies, and why flexibility will be so important to keep up with the AI market: Jon Krohn talks to Richmond Alake about the NoSQL database MongoDB, including why it’s a great addition to your toolkit for developing (agentic) AI applications, with a look under the hood at its native vector database. Richmond also talks about why he expects multi-agent AI architectures to go mainstream in 2025. Additional materials: www.superdatascience.com/871 This episode is brought to you by the Dell AI Factory with NVIDIA and by ODSC , the Open Data Science Conference . Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (04:10) How Richmond became a Staff Developer Advocate (07:40) How NoSQL database differs from a relational database (16:50) The advantages of working with the cloud-based MongoDB Atlas (32:26) Richmond’s predictions for agentic AI (40:38) How to create an effective AI strategy…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
In this Five-Minute Friday, Jon Krohn looks into what he considers the world’s most powerful research tool to date, OpenAI’s Deep Research. Find out how OpenAI trained Deep Research to compile literature reviews of limitless topics, what similar tools are on the market, and where Jon sees the tool as having real-world value including how he uses it daily. Additional materials: www.superdatascience.com/870 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.…
 
S
Super Data Science: ML & AI Podcast with Jon Krohn
Super Data Science: ML & AI Podcast with Jon Krohn podcast artworkSuper Data Science: ML & AI Podcast with Jon Krohn podcast artwork
 
Jon Krohn talks to Varun Godbole about AI prompt engineering, generative wisdom, and AI generalists in this episode all about the interrelationships between humans and AI. Additional materials: www.superdatascience.com/869 This episode is brought to you by the Dell AI Factory with NVIDIA and by ODSC, the Open Data Science Conference . Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (10:44) Using deep learning to predict breast cancer (15:55) All about Varun’s Tuning Playbook (29:56) On the explosion of interest and news about AI and data science (46:35) About Varun’s Wise AI…
 
Loading …

Welcome to Player FM!

Player FM is scanning the web for high-quality podcasts for you to enjoy right now. It's the best podcast app and works on Android, iPhone, and the web. Signup to sync subscriptions across devices.

 

icon Daily Deals
icon Daily Deals
icon Daily Deals

Quick Reference Guide

Listen to this show while you explore
Play