Language Modelling for Optimization
Date: Wednesday, Sep 13, 2023, 14:30 - 16:00
- Xingyou Song, Google Deepmind
- Yutian Chen, Google Deepmind
Undeniably, Large Language Models (LLMs) have stirred an extraordinary wave of innovation in the machine learning research domain, leaving substantial impact across diverse sectors such as reinforcement learning, robotics, and computer vision. Their incorporation within these areas has been rapid and transformative, marking a significant paradigm shift.
Despite this fact, the field of AutoML has been much less affected by such a paradigm shift, even though AutoML presents a unique landscape ripe for exploration when integrating with LLMs. In this tutorial, we will survey the literature and discuss the most promising ways LLMs can revolutionize AutoML. This includes harnessing the vast wealth of information encapsulated in free-form text to enrich task comprehension, and utilizing highly flexible sequence models such as Transformers to engineer superior optimization strategies and enhance performance prediction.