BONUS TSNE: here's a graph of the embeddings of the different-lengthed purple strings, colored by the number of purples what a strange structure!
MACHINE LEARNING
-
Text Embedding Collision Discovery: Purple Repetition Experiment
By
–
I think I found the first evidence of a collision in text embeddings! i.e. for an embedding function f, i found a text pair x, y where x ≠ y but emb(x) = emb(y) ok granted it isn't thatttt interesting but what I did was repeat the word purple from 1 to 8191 times. 'purple ' is
-
Finding Hidden Optimization Opportunities in Complex Systems
By
–
so much fun to pore over the details of a working system to find ideas for optimizations no one has noticed
-
MusicLM Reconstructs Music from Brain Activity Signals
By
–
How can we reconstruct music from brain activity? Learn how our music generation model, MusicLM, was used to reconstruct music in this Research@ NYC lightning talk from Timo Denk, Software Engineer at Google → https://t.co/U8ph36azjv pic.twitter.com/IOgM0iH8BN
— Google AI (@GoogleAI) 9 février 2024How can we reconstruct music from brain activity? Learn how our music generation model, MusicLM, was used to reconstruct music in this Research@ NYC lightning talk from Timo Denk, Software Engineer at Google → https://
goo.gle/3SNJll2 -
Data Quality and Human Factor in AI Systems
By
–
I've been thinking about data quality & human factor in the process a lot lately, so write a short post on the topic: https://
lilianweng.github.io/posts/2024-02-
05-human-data-quality/
… More: If you are into the topic, my team is hiring Research Engineer for a new sub-team Human-AI Interaction: https://
openai.com/careers/resear
ch-engineer-human-ai-interaction
… -
Revisiting Four-Year-Old Satire on Silly Models
By
–
Was just revisiting this blog post from four years ago … https://
inverseprobability.com/2019/12/14/min
istry-of-silly-models
… It was written as a spoof … but has become more prescient than I expected. The Ministry of Silly Models -
Fine-tuning Code Llama 70B for Specific Coding Tasks
By
–
#CodeLlama-70B is the latest and greatest in open-source #codegen models. But, what if you want to #finetune it for a specific coding task? To help you get started, here's a step-by-step #tutorial w/ sample data and #code. Happy fine-tuning!
-
AI Decodes Full Passages from Herculaneum Scrolls
By
–
Researchers reveal first full passages decoded using #AI from famously inscrutable Herculaneum scrolls
-
MLflow Ambassador Program Launch for Community Engagement
By
–
Introducing the @MLflow Ambassador Program! Ambassadors will get to engage the community and showcase their domain expertise through technical content, Meetups and more — apply here: https://
bit.ly/3Ovg2RH
