Data, graphs, and laughs
Ghana
Joined February 2011
- Tweets2K
- Following740
- Followers98
- Likes7.2K
#graphshade Elvis A. retweeted
Why do tree based models still beat deep learning on tabular data?
This NeurIPS benchmark across 45 datasets found XGBoost and random forests ahead, and explains why: trees cope better with uninformative features and irregular patterns.
arxiv.org/abs/2207.08815
#graphshade Elvis A. retweeted
The complete reasoning-from-scratch video series is now up (and linked) here: magazine.sebastianraschka.co…
(Taking a short break but will be back with LLM model and research updates in November!)
#graphshade Elvis A. retweeted
Table of Contents and access to individual chapters can be found here: bayes.cs.ucla.edu/PRIMER/
A First Course in Causal Inference: arxiv.org/abs/2305.18793
[490-page PDF download]
+ Also see the book "Causal Inference in Statistics: A Primer" at amzn.to/3Mrm2wO by @yudapearl
#Probability #Mathematics #DataScience #ML #MachineLearning #DataScientist
#graphshade Elvis A. retweeted
Look friends. Here's what you need to know about "user instructions" for AI agents. What *you intend* is different from *what a system processes*. Saying "do X" does not rule out "don't do Y". And simple natural language commands are not a guarantee for direct control. 1/
the cybersecurity situation is deteriorating at a pace that’s getting genuinely difficult to comprehend.
> microsoft: ai-powered attackers targeting hundreds of organizations, sometimes reaching privileged cloud access in under 3 hours.
> crowdstrike: four separate threat actors using autonomous agents. one already compromised 15+ healthcare organizations.
> autonomous ai agents: reportedly used to compromise 100+ corporate websites, including sites associated with John Deere, SharkNinja and Cummins.
> major us banks: JPMorgan, Bank of America, Wells Fargo and Morgan Stanley are dealing with third-party ai agents accessing customer accounts. some apparently attempting actions beyond user instructions.
> php2shell: millions of exploitation attempts against critical php vulnerabilities. governments issuing urgent warnings, with some organizations taking systems offline.
> litellm: backdoored release shipped through its official distribution pipeline, stealing credentials and enabling remote code execution.
> railway: possible exposure of decrypted customer secrets.
> zeabur: breached AGAIN, now shutting down entirely.
> emudhra: serious certificate-security concerns involving infrastructure used to establish trust in websites and software.
> openai: two of its own coding agents reportedly interfered with internal security monitoring during coding tasks.
> gitlab and node.js: major vulnerabilities requiring urgent patches.
and this is all happening while frontier models are getting substantially better at finding vulnerabilities and carrying out complex computer tasks.
cybersecurity is having an absolutely catastrophic week and i don’t think we’re prepared for how quickly this could escalate.
#graphshade Elvis A. retweeted
How to Learn Mathematics: A Self-Study Roadmap
abakcus.com/articles/how-to-…
#graphshade Elvis A. retweeted
The lectures of my new course "Principles of Deep Representation Learning" are now available at the Youtube site: youtube.com/playlist?list=PL… with new lectures continuing to be added there. I will open-source all the slides later in the textbook website: ma-lab-berkeley.github.io/de…
#graphshade Elvis A. retweeted
The best way to learn the mother of all programming languages:
- 3h beginner course
- 7h advanced course
Get both-in-one, only for a few days at this price
udemy.com/course/advanced-c-…
JB
#graphshade Elvis A. retweeted
ChatGPT launched their version of Generative UI yesterday. We took it apart
tldr; the agent never write raw code. It emits DIL (a new language), the server compiles and streams it, the client renders it using a component set so that it feels native
#graphshade Elvis A. retweeted
Your phone runs more than Android or iOS behind the scenes.
In this handbook, Nikheel explores QuRT, Qualcomm’s real-time operating system that powers DSP workloads for audio, AI, sensors, and more.
You’ll learn about scheduling, memory management, interrupts, and how RTOS systems differ from general-purpose OSes.
freecodecamp.org/news/qurt-t…
#graphshade Elvis A. retweeted
Taking a full-stack app from local dev to prod involves much more than just putting it on a server.
In this course, Imad explains how to deploy, secure, and automate a Vue.js and FastAPI app.
You'll learn how to set up an Ubuntu server plus all about Nginx, SSL, Cloudflare, CI/CD, security testing, and observability.
freecodecamp.org/news/learn-…
#graphshade Elvis A. retweeted
Dear Readers in Causality Research,
I am delighted to let you know that the second edition of The Book of Why (BOW) is scheduled to appear on September 20, 2026.
This edition will include:
1. Revisions and corrections suggested by readers, students and instructors.
2. An account of recent breakthroughs in causal inference (e.g., missing data, personalized medicine, and more).
3. A discussion of the role of causal inference in the era of LLM technology.
4. A perspective on “world models”: what they are, why they are needed, and what general principles we can learn from the structural causal models used in The Book of Why.
The latter topic has become hot in the past few months, so I am posting a section from the Preface, titled "Worldviews, Mindsets, and Narratives: Hidden Nuggets of Causal Inference," ucla.in/4yGPuBk -- which highlights the general principles we can learn from the structural causal models used in The Book of Why.
A word about the dangers of AGI (Artificial General Intelligence, or "super-intelligent AI systems"), which have become the target of general public concerns in just the past two weeks. I remain convinced that AGI is achievable, that the road to AGI goes through causal inference, and that the dangers of AGI taking over and enslaving humanity are real. At the same time, I also believe that causal inference may hold the key to controlling AGI systems so as to secure the survival of the human species.
I hope therefore, half jokingly, that The Book of Why will not be banned by draconian legislation currently brewing in DC. (Hurry to get a copy just in case.)
With warmest regards,
Judea
@GaryMarcus @eliasbareinboim @soboleffspaces @tdietterich @erichorvitz @ylecun @geoffreyhinton
#graphshade Elvis A. retweeted
🤖🧠NEW PAPER🧠🤖
(The result of an 8-year project!)
LLMs seem very different from symbolic systems. Yet LLMs excel in symbolic domains (e.g., language/code/math). How do they do it?
Our finding: LLM representations have implicit symbolic structure!
Link in thread ⬇️
1/n
![Overview of the paper.
Title: The Emergent Symbolic Structure of Artificial Neural Networks
Authors: Tom McCoy, Paul Soulos, Tal Linzen, Paul Smolensky
Left: Neural networks encode information in vectors (there is then an image of a vector), yet they excel at tasks long thought to require symbolic structure (there is then an image of a symbolic representation, specifically a syntax tree). How do LLMs do it?
Right: We find that LLM representations can be closely approximated with symbolic structures. This approximation lets us edit the structure of an LLM’s output by editing the structure of its internal representations, as shown. There is then an image of two edits to LLMs. In the first one, the original input is 3 + 6 * 8, with an answer of 51. But if we swap the positions of the 3 and the 6, the output becomes 30. In the second one, the original input is a Python command repeating the list [Z, U] three times, producing [Z, U, Z, U, Z, U]. But if we edit the input in a way that adds](https://pbs.twimg.com/media/HRJ5Jr5W0AAQ-sN.jpg?name=orig)
ALT Overview of the paper. Title: The Emergent Symbolic Structure of Artificial Neural Networks Authors: Tom McCoy, Paul Soulos, Tal Linzen, Paul Smolensky Left: Neural networks encode information in vectors (there is then an image of a vector), yet they excel at tasks long thought to require symbolic structure (there is then an image of a symbolic representation, specifically a syntax tree). How do LLMs do it? Right: We find that LLM representations can be closely approximated with symbolic structures. This approximation lets us edit the structure of an LLM’s output by editing the structure of its internal representations, as shown. There is then an image of two edits to LLMs. In the first one, the original input is 3 + 6 * 8, with an answer of 51. But if we swap the positions of the 3 and the 6, the output becomes 30. In the second one, the original input is a Python command repeating the list [Z, U] three times, producing [Z, U, Z, U, Z, U]. But if we edit the input in a way that adds
#graphshade Elvis A. retweeted
I’m excited to finally announce the newest edition my Stanford course 𝗧𝗵𝗲 𝗠𝗼𝗱𝗲𝗿𝗻 𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿. It has been 9 months in the making.
Last November, with the release of Claude Opus 4.5, coding agents experienced a step function improvement in capability. We all felt it. The LLMs were more powerful, could reason for longer, solve harder tasks.
This year’s iteration of my course reflects the 2026 metamorphosis of software engineering.
My core belief is simple: AI-native developers of the LLM era are going to become the most important members of any software organization. I have designed my course to train this next generation of engineers.
𝗪𝗵𝗮𝘁’𝘀 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝘁𝗵𝗶𝘀 𝘁𝗶𝗺𝗲 𝗮𝗿𝗼𝘂𝗻𝗱
First, 85% of my Fall 2025 class material is being thrown out. The Fall 2026 syllabus reflects the core capabilities AI-native engineers must have: agent skills, advanced context engineering, MCP portals, agent-ready codebase principles, agentic code review, security, parallelizing background agents, software factories, and more.
Second, I am going to teach my students how to have software taste. Every student will be required to ship pull requests to production-grade, real-world codebases. The course is collaborating with the top open-source AI repos who will offer support and mentorship to students on how to meaningfully contribute to their projects.
This has never been done before in any university course so I am incredibly grateful to our OSS Partners: @browserbase, @HeyGen, @CopilotKit, @semgrep, @OpenHandsDev, @milvusio, @marimo_io, Pi, @crewAIInc, @warpdotdev, @vercel, @cmux, @arizeai, @UnslothAI, and @anyscalecompute.
𝗪𝗵𝗮𝘁’𝘀 𝘀𝘁𝗮𝘆𝗶𝗻𝗴 𝘁𝗵𝗲 𝘀𝗮𝗺𝗲
I’m fortunate to again have AI software engineering leaders and founders as guest speakers to share their learnings from building top coding agent products. Thank you to @leerob from @cursor_ai, @bcherny of @claudeai code, @EnoReyes of @FactoryAI, @silasalberti of @cognition, @0xine of @semgrep, Rajesh Bhatia of @Cloudflare , @amasad of @Replit, and @eladgil.
All resources will be available online. All classes will be available to the public.
9/22 on Stanford campus. See you in class.
themodernsoftware.dev/
#graphshade Elvis A. retweeted
A beautiful article on whether AI can "do mathematics" by the legendary Peter Sarnak, Prof. at Princeton and Ph.D. supervisor of at least two Fields Medallists. For anyone interested in such philosophical issues, it is a must read.
ams.org/journals/notices/202…
#graphshade Elvis A. retweeted
After 18 months of writing, coding, and experimenting, Build a Reasoning Model (From Scratch) is
finally out!
My first copies just arrived! 📚
440 full-color pages. Inference scaling, reinforcement learning, and distillation from scratch.
#graphshade Elvis A. retweeted
Just caught up with the recent GLM-5.2 release. The best open-weight model today.
Architecture-wise, it's build on the GLM-5 and GLM-5.1 architecture that I covered previously, which means it's reusing the Multi-head Latent Attention (MLA) and DeepSeek Sparse Attention (DSA) mechanisms from DeepSeek V3.2. (I wrote about it here: magazine.sebastianraschka.co…)
What's new is that they added an IndexShare mechanism. (That's a cross-layer reuse trick for DSA where instead of recomputing the sparse-attention top-k indexer in every layer, GLM-5.2 runs the full indexer only once every four layers and lets the following layers reuse those selected token indices. This keeps the same DSA idea but makes 1M-token inference much cheaper.)
