Victor Codina @victorcodina

Data Scientist at @Adevinta #RecSys #MachineLearning #BigData es.linkedin.com/in/victorcodina Barcelona Joined February 2010

Tweets

439
Followers

153
Following

89
Likes

77

Radek Osmulski 🇺🇦 @radekosmulski

2 years ago

How do you train a RecSys model and put it in production? ✅ overcoming extreme cold start problem ✅ serving real-time session-based recommendations using Merlin and Triton ✅ recommendation click rate up to 170% vs control Please find the talk here: register.nvidia.com/flow/nvidia/gt…

2 8 51 7K 17

Download Image

Jaime Teevan @jteevan

3 years ago

I got my PhD in AI to build amazing information retrieval (IR) systems. Today's Bing announce shows how transformative the combo is when taken to its limit. I'm so giddy about it I'm giggling. But how do you make an AI laugh? Tell it an IR-relevant joke. blogs.microsoft.com/blog/2023/02/0…

3 20 172 32K 17

Sundar Pichai @sundarpichai

3 years ago

1/ In 2021, we shared next-gen language + conversation capabilities powered by our Language Model for Dialogue Applications (LaMDA). Coming soon: Bard, a new experimental conversational #GoogleAI service powered by LaMDA. blog.google/technology/ai/…

690 3K 15K 7.4M 1K

Alexandros @alexk_z

3 years ago

Our latest work on User Intent Modelling for Generative Sequential Recommendation Models #recsys #TheWebConf arxiv.org/abs/2211.09832

0 4 22 2K 3

Andrej Karpathy @karpathy

3 years ago

Didn't tweet nanoGPT yet (quietly getting it to good shape) but it's trending on HN so here it is :) : github.com/karpathy/nanoG… Aspires to be simplest, fastest repo for training/finetuning medium-sized GPTs. So far confirmed it reproduced GPT-2 (124M). 2 simple files of ~300 lines

33 272 2K 381K 497

Download Image

Radek Osmulski 🇺🇦 @radekosmulski

3 years ago

Some very interesting things went into our recent release so we decided to write about them 🙂 • Tensorflow Transformers for session-based recs • multi-GPU training with Horovod • hyperparameter optimization with Optuna Hope you find them useful 🙏 medium.com/nvidia-merlin/…

0 4 44 0 21

Dan Go @FitFounder

3 years ago

Cheat codes I know at 42 I wish I knew at 22:

2K 27K 130K 0 67K

Alex Xu @alexxubyte

3 years ago

Twitter Architecture 2022 vs. 2012. What’s changed over the past 10 years? Thank you, @elonmusk for the transparency. {1/2}

410 5K 26K 0 6K

Download Image

Papers with Code @paperswithcode

3 years ago

🪐 Introducing Galactica. A large language model for science. Can summarize academic literature, solve math problems, generate Wiki articles, write scientific code, annotate molecules and proteins, and more. Explore and get weights: galactica.org

211 2K 8K 0 3K

Download Video

Radek Osmulski 🇺🇦 @radekosmulski

3 years ago

The OTTO – Multi-Objective Recommender System competition just launched on @kaggle!🥳 For anyone who might be interested, I wrote a recap of the activity on the forums so far. Should be a good way to dive in! 🙂 Link: kaggle.com/competitions/o…

5 28 133 0 43

Download Image

Jo Kristian Bergum @jobergum

3 years ago

This is a cool paper from Google Research, builds on ColBERT with multi-vector representations instead of a single vector representation. This is the future and generalizes much better to new domains than single representations. See also blog.vespa.ai/pretrained-tra…

arXiv CS-CL @arxiv_cscl

3 years ago

0 2 25 0 8

1 13 70 0 26

Aman Jha @amanjha__

3 years ago

It's here! Upload *any paper* to Explainpaper and start instantly getting explanations! Ask follow up questions if you need a more in-depth answer. Go to explainpaper.com and go read all the papers you've been saving! 📝📝📝

160 1K 5K 0 3K

Download Video

Frank Hutter @FrankRHutter

3 years ago

This may revolutionize data science: we introduce TabPFN, a new tabular data classification method that takes 1 second & yields SOTA performance (better than hyperparameter-optimized gradient boosting in 1h). Current limits: up to 1k data points, 100 features, 10 classes. 🧵1/6

104 745 4K 0 2K

Download Image

Google AI @GoogleAI

3 years ago

Learn how to use the TPU embedding API and Cloud TPU v4 to accelerate the training of recommendation systems. Sign up for TPU v4 access at goo.gle/3VkTu7Y and read more ↓ goo.gle/3yCsjMd

4 24 93 0 7

ACM RecSys @ACMRecSys

4 years ago

Are you ready? The #RecSys2022 Challenge is set to start on March 7! We are proud to reveal that this year's challenge is in collaboration with Fashion-AI experts @dressipi: recsyschallenge.com/2022/. You can find more about Dressipi on their website: dressipi.com.

0 15 45 0 2

Maciej Kula @Maciej_Kula

5 years ago

Deep learning based recommenders are a huge leap over traditional MF in expressiveness and flexibility. I'm excited to be working on a library that makes experimenting with them faster and easier!

1 3 12 0 1

Adevinta @Adevinta

5 years ago

We’ve agreed to acquire @eBay’s Classifieds Group. We’ll be the world’s largest online classifieds company, with leading digital marketplaces in 20 countries, reaching 1 billion people and with 3 billion average monthly visits. More: bit.ly/2WFqY4l #acquisition #ebay

2 40 77 0 0

Download Image

Softcatalà @softcatala

5 years ago

Fil 🧵 molt interessant 👇 de @jordimash!!!

Jordi Mas @jordimash

5 years ago

Fil 🧵 molt interessant 👇 de @jordimash!!!

22 465 821 0 20

0 11 11 0 0

Shishir Kumar @Shishir_iitkgp

5 years ago

Ever wondered how Youtube recommends videos? Explore Google's SlateQ algorithm - a state of the art Reinforcement Learning(RL) technique developed for slate recommender problems. Link: medium.com/@kumar.shishir… @usfca_msds #ReinforcementLearning #DataScience