1/ What could you build if your app had the entire knowledge of Wikipedia? This post by @JayAlammar & @Nils_Reimers shows you how to use 100 million embedding vectors that cover Wikipedia in 10 languages. Download them and start building today!
MACHINE LEARNING
-
LLMs Immortalize Documented Knowledge and Diminish Traditional Expertise
By
–
i remember how fun the oreilly perl book was as a kid, to know the whole thing. our shared, slow-moving technical knowledge is becoming less important maybe. but LLMs make the parts we document well immortal. if the model knows, someone always will
-
Key Components of End-to-End MLOps Platform Integration
By
–
Here is a cheat sheet listing the key components of an end-to-end MLOps platform. The main advantage of a platform like @abacusai
: You can get most of these components out of the box in an integrated solution. -
Why AI Writes Grammatically Perfect Yet Meaningless Text
By
–
The Difference Between Speaking and Thinking
The human brain could explain why AI programs are so good at writing grammatically superb nonsense. https://
bit.ly/3LyAaBp -
In-Context Learning Discussion and Collaboration
By
–
Anyone focusing on in-context learning? Would love to chat. DMs are open!
-
SAS Innovate 2023: Data Science, Machine Learning and AI Conference
By
–
#Analytics conferences have a solid agenda of #DataScience #MachineLearning and #AI. Get all that and more at #SASInnovate 2023 in Orlando on May 8-10. Register now and start building your agenda here: https://
sas.com/gms/redirect.j
sp?detail=PLN2755_1799220624
… by @SASsoftware ———
#SASVisionary #DataLeadership #ML -
A Cookbook of Self-Supervised Learning Techniques and Methods
By
–
A Cookbook of Self-Supervised Learning: https://
arxiv.org/abs/2304.12210
=====
#BigData #DataScience #AI #MachineLearning #Algorithms #DataScientists -
LLM Customization Ecosystem: From Prompting to Finetuning
By
–
LLM customization ecosystem is heating up – Remarkable that prompt engineering works at all, but stagnates
– Retrieval can help few-shot prompts, but still…
– Finetuning (BC/RL) is the cannon. But is much more involved
Congrats @realSharonZhou & @GregoryDiamos on the launch! -
Meta’s Efficient AI Development Ecosystems for Production
By
–
AI development ecosystems are increasingly complex and challenging to maintain. Companies like Meta need to develop highly efficient systems to build, serve and improve AI models for production uses. Here's how we make this work effective + efficient
-
T5 Span Corruption: Data Masking and Autoregressive Decoder Loss
By
–
Span corruption in t5 is just a data operation over regular text. Masking stuff in inputs and moving them to targets. It's still fundamentally autoregressive loss on the decoder end.