What is a privacy budget?
Machine Learning / AI
Why was DBSCAN chosen for the anomaly detection task instead of classification?
What offline learning techniques for LTR from click logs (counterfactual evaluation) are used?
What is FastText and how does it solve the OOV problem?
Have you had experience working with NLP tasks?
What happens if you delete the first or last tree in boosting?
What is DeepSpeed ZeRO Stage 1/2/3?
What are the main components of the basic Retrieval-Augmented Generation (RAG)?
What operating system and configuration do you use on your local computer?
Name an ensemble learning method other than bagging and boosting.
What is diffusion distillation?
What is diversity and MMR (Maximal Marginal Relevance)?
What is MLM (masked language modeling)? What proportion of tokens is masked?
What is a reranker and why is it needed?
What is sliding-window attention in Mistral?
What is DDPG and what tasks is it suitable for?
What is policy and how does a stochastic one differ from a deterministic one?
What are Thompson sampling and UCB?
Have you encountered data drift?
Why is LayerNorm applied over the last dimension rather than over the batch?