May 2019 - Bag of Tricks in Machine Learning

More data usually beats better algorithms

May 21, 2019 by admin

https://anand.typepad.com/datawocky/2008/03/more-data-usual.html

ShujianFollow

Retweet on Twitter Shujian Retweeted

Zexuan Zhong @ZexuanZhong·

6 Oct

Dense retrieval models (e.g. DPR) achieve SOTA on various datasets. Does this really mean dense models are better than sparse models (e.g. BM25)?
No! Our #EMNLP2021 paper shows dense retrievers even fail to answer simple entity-centric questions.

https://arxiv.org/abs/2109.08535 (1/6)

Reply on Twitter 1445857517276438532 Retweet on Twitter 144585751727643853226 Like on Twitter 1445857517276438532153 Twitter 1445857517276438532

Retweet on Twitter Shujian Retweeted

Florian Mai 🇪🇺@_florianmai·

2 Oct

Did Google just casually train a 137B parameter LM to basically solve text style transfer (very close to human performance) for any style (the model accepts arbitrary target styles) through an awfully simple method and described all of that in 4 pages? WTF
http://arxiv.org/abs/2109.03910

Reply on Twitter 1444343658011107337 Retweet on Twitter 144434365801110733773 Like on Twitter 1444343658011107337519 Twitter 1444343658011107337

Retweet on Twitter Shujian Retweeted

Dmitry Kan @DmitryKan·

2 Oct

I just published Milvus, Pinecone, Vespa, Weaviate, Vald, GSI: what unites these buzz words and what makes each stand out https://link.medium.com/a2ouMMXx1jb

Reply on Twitter 1444242568720572418 Retweet on Twitter 14442425687205724188 Like on Twitter 144424256872057241825 Twitter 1444242568720572418

Shujian @Shujian_Liu·

30 Sep

Search On '21 in under 7 minutes https://youtu.be/LqAzpJthtIY via @YouTube

Reply on Twitter 1443438443435073536 Retweet on Twitter 1443438443435073536 Like on Twitter 14434384434350735361 Twitter 1443438443435073536

Retweet on Twitter Shujian Retweeted

Even Oldridge @Even_Oldridge·

27 Sep

Incredibly excited and proud to share work by the @nvidiaai Merlin team at #RecSys2021. We introduce an open source library which wraps @huggingface transformers library and makes it accessible for session-based RecSys. Our blog has all the details: https://medium.com/nvidia-merlin/transformers4rec-4523cc7d8fa8

Reply on Twitter 1442541594343661579 Retweet on Twitter 144254159434366157937 Like on Twitter 1442541594343661579125 Twitter 1442541594343661579