Akshay Jacob Thomas

Bayesian statistics and applied mathematics for #composite materials

[Santiago @svpino]

The tokenmaxxing culture is absolutely out of hand.

You don't need an agent for everything.

In fact, I'm 99% sure you don't need an agent for 99% of the problems you need to solve.

Good engineering principles still apply. Simplicity still wins.

A regular, boring ol' script that works every time is 1000 times better than a fancy agent that works sometimes.

We need common sense back.

[Andrew Gordon Wilson @andrewgwils]

Being good at next word prediction is the opposite of what we want for creativity, for scientific breakthroughs.

[Akshay Jacob Thomas @ajacobth]

@Astro_JulioH @TheStrugglingS4 Because that’s when they read it

[Akshay Jacob Thomas @ajacobth]

@betanalpha @danny_smyl @Astro_JulioH

[Akshay Jacob Thomas @ajacobth]

@danny_smyl writing one right now for stochastic PDEs 🥸

[Akshay Jacob Thomas @ajacobth]

@danny_smyl This brings us to the intersection of scientific machine learning and statistical computing. Encoding prior beliefs (physics) in mathematical constructs and quantifying limited knowledge could help in modeling the measurement process, resulting in uncertainty aware ML models.

[Akshay Jacob Thomas @ajacobth]

@LucyInTheSky22 @Uber Recently booked one at 4 30 am from Anaheim to LAX. @Uber showed up

[Akshay Jacob Thomas @ajacobth]

Looking for a probabilistic simulation framework for #composites #additivemanufacturing? Our recent paper deals with the first step to enabling such a #bayesian framework. A huge shout out to @pymc_devs for providing #statistical tools for us engineers.
https://t.co/IwTwFpf8l8

[NASA @NASA]

It's here–the deepest, sharpest infrared view of the universe to date: Webb's First Deep Field.

Previewed by @POTUS on July 11, it shows galaxies once invisible to us. The full set of @NASAWebb's first full-color images & data will be revealed July 12: https://t.co/63zxpNDi4I

[Alexander CS Hendorf @hendorf]

PyMC is the premier tool for Bayesian modelling, the new version 4️⃣ features a JAX backend for faster sampling and includes Aesara - a new theano based tensor library @twiecki @pydatalondon

[Gavin Buckingham @DrGBuckingham]

Who made this GORGEOUS effect?

[Ravin Kumar @canyon289]

Got my copy of Bayes Rules! Sampled 50 pages or so and it's awesome. I really enjoy the quiz yourself sections, shows the authors are not just writing things down but ensuring readers are really learning it as well

Also 💕open access
https://t.co/6VdNNqLA35

[Andrew Gordon Wilson @andrewgwils]

The phrase "inductive biases" can be useful but I agree it's often used loosely. From a probabilistic perspective, inductive biases have a natural definition as the prior distribution over solutions (e.g., what's a priori likely). https://t.co/midasGNPYn

[Demetri (is over at the other place too) @PhDemetri]

Initially, research was applied problems.

But showing novelty of the method is kind of boring to me.

LoOk aT mY FaNcY NeW MoD-- I don't care. Do a useful thing.

[Akshay Jacob Thomas @ajacobth]

@danny_smyl I would add statistics and probability to that ;)

[Grant Sanderson @3blue1brown]

New video!

This is easily the longest one I've ever made. When you find yourself in the mood to settle in with a good puzzle, and a story about two different problem-solving styles, I hope you enjoy it.

https://t.co/TTroxc38r3

[Google DeepMind @GoogleDeepMind]

Introducing the '21 DeepMind x @ai_ucl Reinforcement Learning Lecture Series, a comprehensive introduction to modern RL.

Follow along with our researchers are they explore Markov Decision Processes, sample-based learning algorithms & much more: https://t.co/vGaQ4ryIBD 1/2

[Christoph Molnar 🦋 christophmolnar.bsky.social @ChristophMolnar]

Is machine learning just statistics in disguise?

Technically, a lot of approaches in ML are based on statistical methods, BUT the approach to how to model data differs greatly between the two fields. ML is not statistics in disguise.

[Demetri (is over at the other place too) @PhDemetri]

A typical day for me is

> Be wrong at stats

> Pain

> Run

> Counteract the run with the salty snacks and beer

I love like a king, honestly.