Show newer

Together with FDA and @HochreiterSepp's IARAI we've been working on an -oriented challenge that is focused on tracking safety of the drug after clinical trials.
Submit before 16th may to join us at @CAMDA_conf 2022 in Wisconsin!
bipress.boku.ac.at/camda2022/

The second - that the model's github implementation _actually_ works out of the box! It should be a standard, but for some reason it rarely works for me

Show thread

The first one is the dual - comparing all models both default and tuned .
Sure, it doesn't make much difference for production deployment of the model, but good defaults are very convenient during and early experiments

Show thread

Popov et al.'s NODE might not be the newest , but I love two things about it. 🧵

arxiv.org/abs/1909.06312

Neural Oblivious Decision Ensembles for Deep Learning on Tabular Data

Nowadays, deep neural networks (DNNs) have become the main instrument for machine learning tasks within a wide range of domains, including vision, NLP, and speech. Meanwhile, in an important case of heterogenous tabular data, the advantage of DNNs over shallow counterparts remains questionable. In particular, there is no sufficient evidence that deep learning machinery allows constructing methods that outperform gradient boosting decision trees (GBDT), which are often the top choice for tabular problems. In this paper, we introduce Neural Oblivious Decision Ensembles (NODE), a new deep learning architecture, designed to work with any tabular data. In a nutshell, the proposed NODE architecture generalizes ensembles of oblivious decision trees, but benefits from both end-to-end gradient-based optimization and the power of multi-layer hierarchical representation learning. With an extensive experimental comparison to the leading GBDT packages on a large number of tabular datasets, we demonstrate the advantage of the proposed NODE architecture, which outperforms the competitors on most of the tasks. We open-source the PyTorch implementation of NODE and believe that it will become a universal framework for machine learning on tabular data.

arxiv.org

is cool, but if you're fine with a considerably cheaper bootleg version, then AI21's Jumbo is for you

studio.ai21.com/playground

CI/CD configuration is currently exceeding the capacity of my brain.

Looks like I need a better runner

RT @fchollet
I'm hearing disturbing rumors that a number of fabulously rich men are turning out to be very right-wing and are seeking to buy up media outlets in order to increase their political influence and shore up their views. This is unprecedented and no one could have predicted it.

Ahh, the bliss of finding a bug in the function that you've used to optimise hyperparams for the past week.
It's nearly as pleasant as stubbing your toe on a screwdriver's pointy end.

RT @AcademicsSay
if you’re worried about billionaires owning publically sourced content on social media wait until you hear about academic publishing

It turns out that my earphones monitor what I am doing while listening to music.

So yeah, I should probably go out for a run

RT @CAMDA_conf
Join leading researchers at '22! Top keynotes on literature by Zhiyong Lu of @NCBI, on by @lschriml of @genomestandards and @diseaseontology. Submit late breaking methods or take the until 16 May! camda.info @iscb @IARAInews

Mildly savage abstracts are one of my favourite types of abstracts

Prezesie Glapiński, trzymamy za słowo

It is what it is
---
RT @docmilanfar
“We achieve state of the art results on standard benchmarks”


twitter.com/docmilanfar/status

RT @Pooyahat
I love the idea of 4-day workweeks. Think about the amount of research that can be done on 3 day weekends.

Congrats @metasub! It's a huge accomplishment, I'm glad that I could help at least by collecting a few samples :)

So if anybody's still wondering about licking that handrail, read the publication and go wild.
---
RT @pawel_labaj
Congrats to whole @metasub! Results in @CellPressNews doi.org/10.1016/j.cell.2021.05 comes from our main project global City Sampling Day but we do more cool stuff metasub.org/projects/ In @…
twitter.com/pawel_labaj/status

RT @hanna_cichy
Na pytanie o ile zdrożeje koszyk wielkanocny nie umiem odpowiedzieć, ale jak wynotowałam w ramach przygotowania wzrost cen składników do ciasta to robi wrażenie. Pozostaje piec szarlotkę - nietradycyjnie, ale jabłka to jedyne, co idzie w dół

Show older
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.