Novice data scientist: Data is data.
— Christoph Molnar (@ChristophMolnar) August 1, 2022
Expert data scientist: Each data column has a difficult past. A past full of life-altering events (distribution shift), loss (missing data), misunderstandings (wrong encodings), and regrets (no documentation of how data was collected).
Data science cannot be fully automated.
— Christoph Molnar (@ChristophMolnar) August 1, 2022
No amount of parameter tuning, benchmarking, automation, model comparison, automated feature engineering, etc., can automatically figure out what data column location_123_old contains and whether it should be a feature or not.
Stanford CS25 - Transformers United
— Jean de Nyandwi (@Jeande_d) July 29, 2022
A new class of Transformers and their applications in NLP, vision, RL, Biology, audio & speech.
9 lecture videos are available already!!
Youtube: https://t.co/UIpWSYjOox
Website: https://t.co/8DUtInnBaO pic.twitter.com/CE9ZARxcrX
1. On measuring causal contributions via do-interventions https://t.co/SjJVL5hCw0
— Amit Sharma (@amt_shrma) July 29, 2022
Shapley value, but each possible change in input is a do-intervention. Simple, principled idea for attribution. By @yonghanjung @patrickbloebaum @eliasbareinboim et al.
NLP techniques are already having a major impact on social sciences. They could potentially revolutionize many fields in the upcoming years. Thanks @DannyCEbanks for this book recommendation.https://t.co/kxVnBHXqLD pic.twitter.com/t59128tbNQ
— Bojan Tunguz (@tunguz) July 29, 2022
Our biggest launch in years: nbdev2, now boosted with the power of @quarto_pub!
— Jeremy Howard (@jeremyphoward) July 28, 2022
Use @ProjectJupyter to build reliable and delightful software fast. A single notebook creates a python module, tests, @github Actions CI, @pypi/@anacondainc packages, & morehttps://t.co/kpH4xsKbjn
We're excited to announce Quarto, a new open-source scientific and technical publishing system.
— RStudio (@rstudio) July 28, 2022
Quarto is the next generation of R Markdown and takes what we've learned from the last 10 years and weave it into a more complete, cohesive whole.
Read more: https://t.co/ceO07mFT0V
A @Gradio Demo for OCR-free Document Understanding Transformer on @huggingface Spaces
— AK (@_akhaliq) July 28, 2022
demo: https://t.co/o0Gdheg8O4
Get started with Gradio: https://t.co/qh8qpILE1S pic.twitter.com/hN3GWGHnAT
I used to think AI was far closer to engineering than to science. I can see now that what we've made of it is actually much closer to art than to engineering. Its primary output is cultural.
— François Chollet (@fchollet) July 27, 2022
Check out a new hybrid approach that leverages both ML and rule-based semantic engines to build a Transformer-based semantic code completion model, which we show can improve developer efficiency. Read more ↓ https://t.co/bu6dTdPgII
— Google AI (@GoogleAI) July 26, 2022
A Short Chronology Of Deep Learning For Tabular Data:https://t.co/VAXJRBMyzj
— Sebastian Raschka (@rasbt) July 24, 2022
Deep tabular methods are an interesting research direction! So, this morning, I sat down and summarized my thoughts + the recent papers I read.
[5/6] For large(r) datasets (50k), however, the gap between tree-based ML and deep learning is much smaller.
— Sebastian Raschka (@rasbt) July 23, 2022
However, the authors argue that large tabular datasets are rare. Do you agree? (I think it is somewhat true in academic collaboration but in industry?) pic.twitter.com/zNqdPkfzHq
