Conversational assistants are becoming prevalent among the wider population due to their simplicity and increasing utility. However, the shortcomings of these tools are as renowned as their benefits. In this work, we present a "first look" at an extensive collection of conversational queries, aiming to identify limitations and improvement opportunities specifically related to information access (i.e., search interactions). We explore over 600,000 Google Assistant interactions from 173 unique users, examining usage trends and the resulting deficiencies and strengths of these assistants. We aim to provide a balanced assessment, highlighting the assistant's shortcomings in supporting users and delivering relevant information to user needs and areas where it demonstrates a reasonable response to user inputs. Our analysis shows that, although most users conduct information-seeking tasks, there is little evidence of complex information-seeking behaviour, with most interactions consisting of simple, imperative instructions. Finally, we find that conversational devices allow users to benefit from increased naturalistic interactions and the ability to apply acquired information in situ, a novel observation for conversational information seeking.

Advancements in large language models (LLMs) have changed information retrieval, offering users a more personalised and natural search experience with technologies like OpenAI ChatGPT, Google Bard (Gemini), or Microsoft Copilot. Despite these advancements, research into user tasks and information needs remains scarce. This preliminary work analyses a Google Bard prompt log with 15,023 interactions called the Bard Intelligence and Dialogue Dataset (BIDD), providing an understanding akin to query log analyses. We show that Google Bard prompts are often verbose and structured, encapsulating a broader range of information needs and imperative (e.g., directive) tasks distinct from traditional search queries. We show that LLMs can support users in tasks beyond the three main types based on user intent: informational, navigational, and transactional. Our findings emphasise the versatile application of LLMs across content creation, LLM writing style preferences, and information extraction. We document diverse user interaction styles, showcasing the adaptability of users to LLM capabilities.

This paper explores the utility of a Large Language Model (LLM) to automatically generate queries and query variants from a description of an information need. Given a set of information needs described as backstories, we explore how similar the queries generated by the LLM are to those generated by humans. We quantify the similarity using different metrics and examine how the use of each set would contribute to document pooling when building test collections. Our results show potential in using LLMs to generate query variants. While they may not fully capture the wide variety of human-generated variants, they generate similar sets of relevant documents, reaching up to 71.1% overlap at a pool depth of 100.

Where do queries -- the words searchers type into a search box -- come from? The Information Retrieval community understands the performance of queries and search engines extensively, and has recently begun to examine the impact of query variation, showing that different queries for the same information need produce different results. In an information environment where bad actors try to nudge searchers toward misinformation, this is worrisome. The source of query variation -- searcher characteristics, contextual or linguistic prompts, cognitive biases, or even the influence of external parties -- while studied in a piecemeal fashion by other research communities has not been studied by ours. In this paper we draw on a variety of literatures (including information seeking, psychology, and misinformation), and report some small experiments to describe what is known about where queries come from, and demonstrate a clear literature gap around the source of query variations in IR. We chart a way forward for IR to research, document and understand this important question, with a view to creating search engines that provide more consistent, accurate and relevant search results regardless of the searcher's framing of the query.

Feature engineering is a fundamental but poorly documented component in Learning-to-Rank (LTR) search engines. Such features are commonly used to construct learning models for web and product search engines, recommender systems, and question-answering tasks. In each of these domains, there is a growing interest in the creation of open-access test collections that promote reproducible research. However, there are still few open-source software packages capable of extracting high-quality machine learning features from large text collections. Instead, most feature-based LTR research relies on "canned" test collections, which often do not expose critical details about the underlying collection or implementation details of the extracted features. Both of these are crucial to collection creation and deployment of a search engine into production. So in this regard, the experiments are rarely reproducible with new features or collections, or helpful for companies wishing to deploy LTR systems.

Language model pre-training has spurred a great deal of attention for tasks involving natural language understanding, and has been successfully applied to many downstream tasks with impressive results. Within information retrieval, many of these solutions are too costly to stand on their own, requiring multi-stage ranking architectures. Recent work has begun to consider how to "backport" salient aspects of these computationally expensive models to previous stages of the retrieval pipeline. One such instance is DeepCT, which uses BERT to re-weight term importance in a given context at the passage level. This process, which is computed offline, results in an augmented inverted index with re-weighted term frequency values. In this work, we conduct an investigation of query processing efficiency over DeepCT indexes. Using a number of candidate generation algorithms, we reveal how term re-weighting can impact query processing latency, and explore how DeepCT can be used as a static index pruning technique to accelerate query processing without harming search effectiveness.

Reducing excessive costs in feature acquisition and model evaluation has been a long-standing challenge in learning-to-rank systems. A cascaded ranking architecture turns ranking into a pipeline of multiple stages, and has been shown to be a powerful approach to balancing efficiency and effectiveness trade-offs in large-scale search systems. However, learning a cascade model is often complex, and usually performed stagewise independently across the entire ranking pipeline. In this work we show that learning a cascade ranking model in this manner is often suboptimal in terms of both effectiveness and efficiency. We present a new general framework for learning an end-to-end cascade of rankers using backpropagation. We show that stagewise objectives can be chained together and optimized jointly to achieve significantly better trade-offs globally. This novel approach is generalizable to not only differentiable models but also state-of-the-art tree-based algorithms such as LambdaMART and cost-efficient gradient boosted trees, and it opens up new opportunities for exploring additional efficiency-effectiveness trade-offs in large-scale search systems.

No abstract available.

Complex machine learning models are now an integral part of modern, large-scale retrieval systems. However, collection size growth continues to outpace advances in efficiency improvements in the learning models which achieve the highest effectiveness. In this paper, we re-examine the importance of tightly integrating feature costs into multi-stage learning-to-rank (LTR) IR systems. We present a novel approach to optimizing cascaded ranking models which can directly leverage a variety of different state-of-the-art LTR rankers such as LambdaMART and Gradient Boosted Decision Trees. Using our cascade model, we conclusively show that feature costs and the number of documents being re-ranked in each stage of the cascade can be balanced to maximize both efficiency and effectiveness. Finally, we also demonstrate that our cascade model can easily be deployed on commonly used collections to achieve state-of-the-art effectiveness results while only using a subset of the features required by the full model.