The field of online advertising is witnessing significant developments with the integration of Large Language Models (LLMs) and reinforcement learning techniques. Researchers are exploring innovative methods to optimize ad text generation, focusing on improving click-through rates (CTR) and ad visibility. A key direction in this field is the use of online feedback and preference optimization to fine-tune LLMs and generate high-CTR ad texts. Additionally, studies are investigating the effectiveness of deep neural networks (DNNs) in Learning-to-Rank (LTR) tasks, comparing their performance to traditional tree-based models. Noteworthy papers in this area include:

• CTR-Driven Ad Text Generation via Online Feedback Preference Optimization, which proposes a novel two-stage framework for ad text generation.
• Improving Generative Ad Text on Facebook using Reinforcement Learning, which introduces a post-training method using historical ad performance data as a reward signal and achieves significant CTR improvements in a large-scale A/B test.