Sardelich

Quantitative PM. PhD in AI/ML. Retired from #Physics, worked as #Quant for +17 years.

[Conrad Hackett @conradhackett]

Be sure to catch the end of this clever @NASA "climate spiral" video showing monthly global temperature changes between 1880 and 2021
https://t.co/pj6Hucz2Xg

[Grey Court School @GreyCourtTweets]

Our very own Physics teacher & breakdance guru, Faraz Khan, is featured in a new Dept of Education/Channel 4 collaboration. The charismatic Mr Khan is an amazing role model. We're delighted his talent & commitment have been showcased. Watch it here 👏👏👏
https://t.co/loMQub1zVS

[Sardelich @SardelichM]

Yes, it was 40 years ago!

[Prash Chandramohan @MDMGeek]

So amazing this video is from 1981.

[Sardelich @SardelichM]

@jackclarkSF @indexingai @OECD Wish You All the Best :)

[Zachary Lipton @zacharylipton]

One interesting (but incomplete) measure of the state of the deep learning bubble. While PyTorch has caught up with TensorFlow, aggregate interest in DL frameworks seems to be maybe half what it was at the peak. I can imagine many interpretations, share yours.

[Oriol Vinyals @OriolVinyalsML]

Biology's ImageNet moment has arrived: #AlphaFold2 solves Protein Folding at the #CASP14 competition!

Blog: https://t.co/KH9g8FFXvS

[Sardelich @SardelichM]

@zacharylipton LOL hahahahha

[Zachary Lipton @zacharylipton]

Thrilled to release our greatest paper yet in Nature Machine Learning: ***Multimodal Multitask Multidomain Multi-Attention for Multilabel Classification with Multiple Adversaries*** https://t.co/q8YD8lr6MG

[Sardelich @SardelichM]

To Build Truly Intelligent Machines, Teach Them Cause and Effect https://t.co/II7wSoyuPW via @QuantaMagazine

[Sardelich @SardelichM]

SocGen Makes A Striking Discovery: For More Than Half The Market, It Already Feels Like A Recession | Zero Hedge https://t.co/zcSwsThW6x

[Sardelich @SardelichM]

Uber lays off 435 people across engineering and product teams https://t.co/P3JmTlt2oM via @techcrunch

[Sardelich @SardelichM]

Mike Bloomberg's philanthropic arm is shelling out $160 million to keep kids from vaping https://t.co/soi32yABsh via @techcrunch

[Sardelich @SardelichM]

🏎 Smaller, faster, cheaper, lighter: Introducing DilBERT, a distilled version of BERT by @victorsanh https://t.co/LwxWlgkJke

[Sardelich @SardelichM]

Goldman Sachs is spending $100 million to shave milliseconds off stock trades https://t.co/Rxl0nJipuW

[Sardelich @SardelichM]

https://t.co/pDBP7Ac1XM

[Sardelich @SardelichM]

Your Data Were ‘Anonymized’? These Scientists Can Still Identify You https://t.co/SGv8XupFRN

[Sardelich @SardelichM]

@mikeharrisNY @lopezdeprado @spyrosmakrid @camharvey The problem is not ML. In some sense, nothing can be more hacking than the 3-4 well-known risk premium. The answer would be cross-validation and dealing with the concept drift, which permeates ML or any other approach.

[Sardelich @SardelichM]

https://t.co/X49VPAV7bb

[Sardelich @SardelichM]

Study: Immigrants and their kids founded 45% of U.S. Fortune 500 companies https://t.co/PtMQdPpqcm

[Sardelich @SardelichM]

Natural Adversarial Examples Slash Image Classification Accuracy by 90% by @Synced_Global https://t.co/1NlX7GX8qu