The rapid advancement of Large Language Models (LLMs) has introduced a paradigm shift in Information Retrieval (IR), moving beyond conventional keyword queries and ranked result lists. LLMs now play a critical role in the evolution of IR technologies and introduce new interaction forms like Retrieval-Augmented Generation, which is a more dynamic and interactive retrieval process that integrates various aspects of Information Access, like Question Answering, into the dialog between a searcher and the search engine. We explore the multi-faceted impact of LLMs on IR, particularly in three distinct layers where they have become an integral part of the retrieval process, namely the retrieval system and processing pipeline that can make use of a richer semantic representation using advanced language models, the interaction layer, and the broader IR ecosystem. For the latter, we focus on evaluation issues as well as bias, fairness, and ethical concerns. We also highlight some recent cases of using LLMs in the medical domain to demonstrate the impact on one specific domain.

Evaluating the interactions between users and systems presents many challenges. Simulation offers a reliable, re-usable, and repeatable methodology to explore how different users, user behaviours and/or retrieval systems impact performance. With Large Language Models and Generative AI now widely available and accessible, new affordances are possible. These allow researchers to create more ''realistic'' simulated users that can generate queries and judge items like humans, and to develop new retrieval systems where responses and interactions are conversational and based on retrieval augmented generation. This resource paper presents a community-led initiative to update the Simulation of Interactive Information Retrieval (SimIIR) Framework to enable the simulation of conversational search using LLMs. The largest update provides a conversational search workflow which involves a number of new possible interactions with a search system or agent -- enabling a host of new development and evaluation opportunities. Other developments include the Markovian Users, Cognitive States, LLM-based components for assessing snippets/documents/responses, generating queries, deciding on when to stop/continue, and PyTerrier integration. This paper aims to mark the release of SimIIR 3.0 and invites the community to build, extend, and use the resource.

Information Retrieval (IR) research is deeply rooted in experimentation and evaluation, and the Text REtrieval Conference (TREC) has been playing a central role in making that possible since its inauguration in 1992. TREC's mission centers around providing the infrastructure and resources to make IR evaluations possible at scale. Over the years, a plethora of different retrieval problems were addressed, culminating in data artifacts that remained as valuable and useful tools for the IR community. Even though the data are largely available from TREC's website, there is currently no resource that facilitates a cohesive way to obtain metadata information about the run file - the IR community's de-facto standard data format for storing rankings of system-oriented IR experiments.To this end, the work at hand introduces a software suite that facilitates access to metadata of experimental resources, resulting from over 30 years of IR experiments and evaluations at TREC. With a particular focus on the run files, the paper motivates the requirements for better access to TREC metadata and details the concepts, the resources, the corresponding implementations, and possible use cases. More specifically, we contribute a web interface to browse former TREC submissions. Besides, we provide the underlying metadatabase and a corresponding RESTful interface for more principled and structured queries about the TREC metadata.

Considering query variance in information retrieval (IR) experiments is beneficial for retrieval effectiveness. Especially ranking ensembles based on different topically related queries retrieve better results than rankings based on a single query alone. Recently, generative instruction-tuned Large Language Models (LLMs) improved on a variety of different tasks in capturing human language. To this end, this work explores the feasibility of using synthetic query variants generated by instruction-tuned LLMs in data fusion experiments. More specifically, we introduce a lightweight, unsupervised, and cost-efficient approach that exploits principled prompting and data fusion techniques. In our experiments, LLMs produce more effective queries when provided with additional context information on the topic. Furthermore, our analysis based on four TREC newswire benchmarks shows that data fusion based on synthetic query variants is significantly better than baselines with single queries and also outperforms pseudo-relevance feedback methods. We publicly share the code and query datasets with the community as resources for follow-up studies.

Reproducibility is a fundamental part of scientific progress. Compared to other scientific fields, computational sciences are privileged as experimental setups can be preserved with ease, and regression experiments allow the validation of computational results by bitwise similarity. When evaluating information access systems, the system users are often considered in the experiments, be it explicit as part of user studies or implicit as part of evaluation measures. Usually, system-oriented Information Retrieval (IR) experiments are evaluated with effectiveness measurements and batches of multiple queries. Successful reproduction of an Information Retrieval (IR) system is often determined by how well it approximates the averaged effectiveness of the original (reproduced) system. Earlier work suggests that the nave comparison of average effectiveness hides differences that exist between the original and reproduced systems. Most importantly, such differences can affect the recipients of the retrieval results, i.e., the system users. To this end, this work sheds light on what implications for users may be neglected when a system-oriented Information Retrieval (IR) experiment is prematurely considered reproduced. Based on simulated reimplementations with comparable effectiveness as the reference system, we show what differences are hidden behind averaged effectiveness scores. We discuss possible future directions and consider how these implications could be addressed with user simulations.

Information retrieval systems have been evaluated using the Cranfield paradigm for many years. This paradigm allows a systematic, fair, and reproducible evaluation of different retrieval methods in fixed experimental environments. However, real-world retrieval systems must cope with dynamic environments and temporal changes that affect the document collection, topical trends, and the individual user's perception of what is considered relevant. Yet, the temporal dimension in IR evaluations is still understudied.To this end, this work investigates how the temporal generalizability of effectiveness evaluations can be assessed. As a conceptual model, we generalize Cranfield-type experiments to the temporal context by classifying the change in the essential components according to the create, update, and delete operations of persistent storage known from CRUD. From the different types of change different evaluation scenarios are derived and it is outlined what they imply. Based on these scenarios, renowned state-of-the-art retrieval systems are tested and it is investigated how the retrieval effectiveness changes on different levels of granularity.We show that the proposed measures can be well adapted to describe the changes in the retrieval results. The experiments conducted confirm that the retrieval effectiveness strongly depends on the evaluation scenario investigated. We find that not only the average retrieval performance of single systems but also the relative system performance are strongly affected by the components that change and to what extent these components changed.

Simulations in various forms have been used to evaluate information access systems, like search engines, recommender systems, or conversational agents. In the form of the Cranfield paradigm, a simulation setup is well-known in the IR community, but user simulations have recently gained interest. While user simulations help to reduce the complexity of evaluation experiments and help with reproducibility, they can also contribute to a better understanding of users. Building on recent developments in methods and toolkits, the Sim4IA workshop aims to bring together researchers and practitioners to form an interactive and engaging forum for discussions on the future perspectives of the field. An additional aim is to plan an upcoming TREC/CLEF campaign.

No existing evaluation infrastructure for shared tasks currently supports both reproducible on- and offline experiments. In this work, we present an architecture that ties together both types of experiments with a focus on reproducibility. The readers are provided with a technical description of the infrastructure and details of how to contribute their own experiments to upcoming evaluation tasks.

Academic search systems aid users in finding information covering specific topics of scientific interest and have evolved from early catalog-based library systems to modern web-scale systems. However, evaluating the performance of the underlying retrieval approaches remains a challenge. An increasing amount of requirements for producing accurate retrieval results have to be considered, e.g., close integration of the system's users. Due to these requirements, small to mid-size academic search systems cannot evaluate their retrieval system in-house. Evaluation infrastructures for shared tasks alleviate this situation. They allow researchers to experiment with retrieval approaches in specific search and recommendation scenarios without building their own infrastructure. In this paper, we elaborate on the benefits and shortcomings of four state-of-the-art evaluation infrastructures on search and recommendation tasks concerning the following requirements: support for online and offline evaluations, domain specificity of shared tasks, and reproducibility of experiments and results. In addition, we introduce an evaluation infrastructure concept design aiming at reducing the shortcomings in shared tasks for search and recommender systems.