Generative AI Terminology - An evolving taxonomy to get you started with Generative Artificial Intelligence

In the realm of Generative AI, newcomers may find themselves daunted by the technical terminology. To alleviate this mental hurdle, various resources compile lists of these terms. 'Generative AI Terminology - An evolving taxonomy to get you started with Generative Artificial Intelligence' categorises these terms into 12 groups, aiming to present the information in a non-technical manner suitable for individuals with a basic grasp of machine learning.

The compiled list encompasses the following categories:

🔵 Categories of models - Foundation models, LLM, SLM, VLMs, etc.

🔵 Common LLM terms - Prompts, Temperature, Hallucinations, Tokens, etc.

🔵 Stages in the LLM lifecycle - Pre-training, Supervised Fine Tuning, RLHF, etc.

🔵 Evaluations in LLM - ROUGE, BLEU, BIG-bench, GLUE, etc.

🔵 LLM architecture - Encoder, Decoder, Transformer, Attention, etc.

🔵 Retrieval augmented generation - Vector DBs, Chunking, Evaluations, etc.

🔵 LLM agents - Memory, Planning, ReAct, CoT, ToT, etc.

🔵 LMM architecture - GAN, VAE, CLIP, etc.

🔵 Cost & efficiency - GPU, PEFT, LoRA, Quantization, etc.

🔵 LLM security - Prompt Injection, Data poisoning, etc.

🔵 Deployment & inference - Pruning, Distillation, Flash Attention, etc.

🔵 A list of providers supporting LLMOps

Like the generative AI space, this taxonomy is also evolving.

Who can benefit from this taxonomy -

• Product folks looking to get a hold on the Generative AI space
• Leaders following Generative AI
• ML Engineers, Data Scientists looking to update their knowledge on LLMs and Generative AI
• AI enthusiasts

I'd also like to get your feedback on this taxonomy. Please feel free to get in touch

Connect on LinkedIn for feedback and discussion