New LLM framework SAGE optimizes ad auto-bidding with parameter efficiency

Researchers have developed SAGE, a new framework for LLM-based auto-bidding in advertising auctions. SAGE utilizes a parameter-efficient multi-modal alignment approach to adapt LLMs for constrained bidding environments. Key components include temporal-semantic positional embeddings, gated cross-attention for modality fusion, and a constraint-gated LoRA module that activates specific LLM experts. Experiments show SAGE achieves superior performance while requiring less than 10% of the parameters needed for full fine-tuning. AI

IMPACT This framework could enable more efficient and effective use of LLMs in advertising technology by reducing computational costs.

RANK_REASON The item is a research paper detailing a new framework for LLM-based auto-bidding. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.IR (Information Retrieval) →

• alphaXiv
• arXiv
• CatalyzeX
• Connected Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• Litmaps
• LLM
• Lora
• SAGE
• ScienceCast
• scite Smart Citations

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →