Georgia DeaconuinAI AdvancesHow to use InstantID to generate identity preserving imagesOr how to generate better pictures of yourselfFeb 13Feb 13
Georgia DeaconuinAI AdvancesLLM Safety Training and Jail-BreakingSince ChatGPT brought LLMs into the spotlight, we witnessed the cat and mouse game between LLM developers trying to implement guardrails…Jan 20Jan 20
Georgia DeaconuinTowards Data ScienceTowards LLM explainability: why did my model produce this output ?The recent release of larger, better LLMs that showcase new capabilities has been paired with growing concerns over AI safetyDec 15, 20235Dec 15, 20235
Georgia DeaconuDisplay Spark logs from AWS Glue jobs in a local Spark History ServerWith AWS Glue it is easy to create and run Spark ETL jobs without having to think about the underlying infrastructure. Unfortunately it…Aug 19, 2023Aug 19, 2023
Georgia Deaconu5 things you never wanted to know about the EurovisionAn analysis of the Eurovision voting data from 1975 to 2019 in order to check some personal biases and deal with some frustration.May 17, 2023May 17, 2023
Georgia DeaconuinTowards Data ScienceHow Data Leakage affects model performance claimsThis year has seen several important scientific advancements enabled by machine learning driven research. Along with the enthusiasm came…Jan 2, 20231Jan 2, 20231
Georgia DeaconuinTowards Data ScienceMonitoring Databricks jobs through calls to the REST APIMonitoring jobs that run in a Databricks production environment requires not only setting up alerts in case of failure but also being able…Oct 9, 20222Oct 9, 20222
Georgia DeaconuUnderstanding core concepts in Apache SparkApache Spark is one of the most popular frameworks for big data processing nowadays. Its success is widely attributed to the 10x reduction…Mar 21, 2022Mar 21, 2022
Georgia DeaconuinTowards Data ScienceWhy we need to deal with imbalanced classesAn illustration of the effects of imbalanced features on model evaluationFeb 16, 2022Feb 16, 2022
Georgia DeaconuinTowards Data ScienceCommon Mistakes in Hyper-Parameters TuningHyper-parameter tuning for machine learning models is a trial and error game. To succeed, it is best to avoid the following mistakes.Jan 28, 20221Jan 28, 20221