Adapters, a plug-in neural network module with some tunable parameters, have emerged as a parameter-efficient transfer learning technique for adapting pre-trained models to downstream tasks, especially for natural language processing (NLP) and computer vision (CV) fields. Meanwhile, learning recommendation models directly from raw item modality features --- e.g., texts of NLP and images of CV --- can enable effective and transferable recommender systems (called TransRec). In view of this, a natural question arises:can adapter-based learning techniques achieve parameter-efficient TransRec with good performance? To this end, we perform empirical studies to address several key sub-questions. First, we ask whether the adapter-based TransRec performs comparably to TransRec based on standard full-parameter fine-tuning? does it hold for recommendation with different item modalities, e.g., textual RS and visual RS. If yes, we benchmark these existing adapters, which have been shown to be effective in NLP and CV tasks, in item recommendation tasks. Third, we carefully study several key factors for the adapter-based TransRec in terms of where and how to insert these adapters? Finally, we look at the effects of adapter-based TransRec by either scaling up its source training data or scaling down its target training data. Our paper provides key insights and practical guidance on unified & transferable recommendation --- a less studied recommendation scenario. We release our codes and other materials at https://github.com/westlake-repl/Adapter4Rec/.

Large foundational models, through upstream pre-training and downstream fine-tuning, have achieved immense success in the broad AI community due to improved model performance and significant reductions in repetitive engineering. By contrast, the transferable one-for-all models in the recommender system field, referred to as TransRec, have made limited progress. The development of TransRec has encountered multiple challenges, among which the lack of large-scale, high-quality transfer learning recommendation dataset and benchmark suites is one of the biggest obstacles. To this end, we introduce NineRec, a TransRec dataset suite that comprises a large-scale source domain recommendation dataset and nine diverse target domain recommendation datasets. Each item in NineRec is accompanied by a descriptive text and a high-resolution cover image. Leveraging NineRec, we enable the implementation of TransRec models by learning from raw multimodal features instead of relying solely on pre-extracted off-the-shelf features. Finally, we present robust TransRec benchmark results with several classical network architectures, providing valuable insights into the field.

Information technology has profoundly altered the way humans interact with information. The vast amount of content created, shared, and disseminated online has made it increasingly difficult to access relevant information. Over the past two decades, RSs and search (collectively referred to as information retrieval systems) have evolved significantly to address these challenges. Recent advances in large language models (LLMs) have demonstrated capabilities that surpass human performance in various language-related tasks and exhibit general understanding, reasoning, and decision-making abilities. This survey provides a structured overview of the emerging application of LLM agents for improving recommendation and search systems. We discuss the motivations and roles of LLM agents and propose a taxonomy to organize and summarize existing research. The survey highlights how LLM agents can address key challenges in information retrieval and charts potential directions for future work. To the best of our knowledge, this paper represents one of the earliest efforts to provide a structured and methodologically grounded overview of LLM-based agents for recommendation and search, offering an integrative perspective on this rapidly advancing area. To help understand the existing work, we provide a curated list of papers on LLM agent-based recommendation and search at https://github.com/tsinghua-fib-lab/LLM-Agent-for-Recommendation-and-Search .

The development of Large Language Models (LLMs) is increasingly oriented toward leveraging deliberate reasoning to break down a wider spectrum of complex problems. In Recommender Systems (RS), recent studies suggest that this LLM-based reasoning paradigm empowers reasoning-aware recommender systems to more effectively uncover and interpret user intent. However, the existing literature has yet to clearly define what constitutes a reasoning process in recommender systems and systematically review its progress. To advance research on reasoning-aware recommender systems, this paper provides a comprehensive review of the development of LLM-based recommender systems from the perspective of reasoning techniques. We begin by tracing the evolution of reasoning in recommender systems and, for the first time, offer a clear definition of reasoning within this domain. We then explore diverse reasoning paradigms for reasoning-aware recommender systems, including Text-Based Reasoning, Latent Reasoning, and Retrieval-Enhanced Reasoning, along with their corresponding specific reasoning subtypes. Subsequently, we analyze the underlying mechanisms behind these paradigms and their transition across four dimensions: Reasoning Space, Reasoning Acquisition, Reasoning Flexibility, and Reasoning Data Collection. We also point out current challenges in the evaluation of reasoning-aware recommender systems, including reasoning benchmarks and metrics. By clarifying these issues, we highlight the potential of reasoning-aware recommender systems and outline promising directions for future recommender systems study.