Really the concept of end-user programming! As LLMs get better at coding, I think this becomes increasingly possible.
CODE
-
AI Scientist: Automating Full Scientific Paper Generation Under $15
By
–
Introducing the AI Scientist: Revolutionizing Research with AI Imagine an AI that can independently develop and write a full scientific paper for under $15! The AI Scientist leverages advanced LLMs to automate the entire research process, from discovery to summarization.
-
Grading Notes: LLM Evaluation Method for Specialized Domains
By
–
Grading Notes were developed for high-quality LLM-as-a-judge evaluation in specialized domains. It is a simple and effective method to enable the evaluation of domain-specific AI – and we even used it in our development of Databricks Assistant!
-
Training Steps Impact on Model Realism Bias
By
–
Para que veáis a qué me refiero con cómo el modelo a más steps de entrenamiento se va sesgando al realismo, este es el mismo prompt+seed para un modelo entrenado con 250 steps y con 1250 steps.
-
Fine-tuning Flux LoRA: Overtraining Effects on Model Rigidity
By
–
Si nos movemos a estilos menos realistas ahí a Flux le cuesta más. Creo que he sobrentrenado al modelo (4000 steps) y por eso cuesta mucho sacarle del realismo Mientras escribo, estoy entrenando un nuevo Lora con menos steps, que creo será menos rígido. Aún así, ni tan mal!
-
Matrix Powers and Graph Walks in AI Algorithms
By
–
According to Danka, the directed graph representation is beneficial because the powers of the matrix correspond to walks in the graph. W/i the elements of the square matrix, all possible 2-step walks are accounted for in the sum defining the elements of A².
-
Encoding Matrices as Graphs for Complex Behavior Analysis
By
–
Breaking down how encoding matrices as graphs makes their complex behaviors easier to study A thread by @TivadarDanka
-
Developing AI for Automation Production and Artifact Understanding
By
–
Strong agree. AI is best used to produce, modify, and understand an artifact. We're developing an AI for producing automations.
-
AI Scientist’s Amusing Bugs: Self-Modifying Code Behavior
By
–
The Bloopers in our blog is quite interesting: https://
sakana.ai/ai-scientist/ Sometimes it tries to modify its own script. For example, when its experiments took too long to run, instead of making its code run faster, it simply tried to modify its own code to extend its timeout period! -
Building a Movie Recommendation System Using Vector Search
By
–
Recommendation System is a set of algorithms that can analyze our likes and dislikes and recommend us with cool new items. And most of the time, all the recommendations are perfect!https://t.co/CP1hRqhjiP
— Satya Mallick (@LearnOpenCV) 13 août 2024
In our latest article, we are building a movie recommendation system… pic.twitter.com/iaOXoGrfcyRecommendation System is a set of algorithms that can analyze our likes and dislikes and recommend us with cool new items. And most of the time, all the recommendations are perfect! https://
learnopencv.com/recommendation
-system-using-vector-search/
… In our latest article, we are building a movie recommendation system