11-10, 11:45–12:30 (America/New_York), Central Park East (6th floor)
Machine language understanding and generation has been undergoing rapid improvements due to recent breakthroughs in machine learning (e.g. large language models like GPT and BERT). And while big tech and NLP engineers were quick to capitalize on these models, the broader developer community lags in adopting these models and realizing their potential in their business domains.
This talk provides a gentle and highly visual overview of some of the main intuitions and real-world applications of large language models. It assumes no prior knowledge of language processing and aims to bring attendees up to date with the fundamental intuitions and applications of large language models.
It's only been a couple of years since humanity reached a language technology breakthrough: software that can write as well as humans do. Large language models (LLMs) like GPT and BERT have taken the machine learning world by storm propelling both the language processing (NLP) and computer vision domains rapidly forward and bringing about a series of surprising application (e.g. DeepMind Gopher, OpenAI GPT3 and Dall-E, Google Pathways Language Model). But while big tech invests heavily in including these models into their products, developers and smaller companies still lag in adopting these models and realizing their potential in their business domains.
This talk provides a gentle and highly visual overview of real-world applications of large language models. It introduces some of the main intuitions to help developers and data scientist who are new to the topic. These include: - how LLMs are trained, - prompt engineering, - embeddings - finetuning
The talk then proceeds to some advanced use cases and practical tips & tricks for use cases such as: - Building LLM-based text classifiers - Semantic search (going beyond keyword search to searching by meaning) - Document clustering - Generating training data with generative language models
This talk assumes no prior knowledge of language processing and aims to bring attendees up to date with the fundamental intuitions and applications of large language models.
No previous knowledge expected
I work on Machine Learning Inference at Cohere AI. Prior to this I spent 3 years at NVIDIA developing Triton Inference Server, an open source solution used to deploy machine learning models into production. I have a Masters in Data Science from the University of Washington.