Mr Timothy Liu1, Dr Aik Beng Ng1, Dr Simon See1
1NVIDIA, Singapore, Singapore
In recent years, large language models (LLM) have achieved state-of-the-art across a wide range of natural language processing (NLP) applications, outperforming previous models which were developed for specific problems. LLMs are also termed as foundation models, as they can be effectively applied to diverse tasks with minimal modification. Increasingly, these models have also been used in diverse scenarios outside of NLP in areas such as life sciences, program synthesis and robotics, with many promising results.
In our talk, we will introduce state-of-the-art LLMs such as GPT-3 and how similar models can be openly accessed by developers and researchers. We will share about the applications of LLMs outside of NLP tasks, focusing on scientific applications in domains such as life sciences (biomedical research, drug discovery, patient care etc.) and physical sciences (chemistry, physics etc.). We will then provide an overview of an open-source software toolkit, NVIDIA NeMo, for researchers and developers to build applications that utilize these LLMs, or even train new LLMs for new domains of data. Lastly, we will also discuss some considerations when using such LLMs.
Biography:
Timothy is a research associate at NVIDIA working on Deep Learning (DL) and Natural Language Processing. His wider interests include understanding generalisation in DL, performant execution of DL models and understanding uncertainty in DL models.
Aik Beng is the regional manager of the NVIDIA AI Technology Center (NVAITC) at NVIDIA. His role is an interesting mix of being a technologist, engineer, researcher, and evangelist in Artificial Intelligence (AI). Together with his team, he actively works on the growth and development of the AI ecosystem, through strategic partnerships and collaborations with institutes of higher learning (universities, polytechnics, etc.), industry and government.
Prof. Simon See is currently the Solution Architecture and Engineering Director and Chief Solution Architect for Nvidia AI Technology Centre. His research interests are in the area of High Performance Computing, Big Data, Artificial Intelligence, machine learning, computational science, Applied Mathematics and simulation methodology.