Why Manual Machine Learning Beats LLMs in Some Cases"
Nothing is loaded from YouTube until you press play, and then without tracking cookies.
About this video
Large language models excel with unstructured data, but manually tuned algorithms still outperform them in specific tasks. Dive into how evolving codebases highlight this shift in AI. #LanguageModels #DataScience #MachineLearning #AIChallenges #TechEvolution #Llama3 #NeuralNetworks #DataEmbedding #ArtificialIntelligence #DeepLearning
Transcript
189 wordsCaptions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.
- 0:00
a lot of the machine learning algorithms that are manually tuned will generally do better than like the autogenerated ones from from large language models I think the key where large langage models are really good is like I have some unstructured data I don't really know how to compress it into something that's a good embedding that I can given that embedding I can do some numerical math on and I just think like we haven't figured out how to do those things without all of these like poison attacks
- 0:25
and and trust issues and also the fact that if you know if you you look at like llama 3 code base evolved over time you will see that like it's started as like here's a Transformer and here's some light stuff that converts your input query to format for the Transformer and here's a an output thing and what you've seen is the code bases have gone from like 80% neural net to like 10% neural
- 0:49
net and the rest of the stuff on the outside is growing that stuff has a lot of subjectivity in it