Conversational search agents, especially with the advent of large language models, have developed into useful tools to satisfy complex information needs of their users. Former research has shown that personalization (i.e., adaptation of agent responses to the preferences and traits of the user) can increase the relevance and perceived answer quality of these systems even further. However, developing accurate personalization methods typically requires rich datasets, both in terms of user profiles and complex conversations, for which only a few resources are publicly available. Over the past three years, the goal of the TREC Interactive Knowledge Assistance Track (iKAT) has been to bridge this gap. In organizing this shared task, we have developed a collection of complex information needs and associated conversations, made to challenge today's conversational agents and thus highlight aspects in need of further research. In this paper, we present the resources made for iKAT 2025, focusing on multi-session conversations (i.e., multiple dialogues per user), dynamically evolving user models, mixed-initiative dialogues, and large-scale human and automatic assessments. In addition to manually designed user profiles and conversations, the test collection for 2025 also contains dialogues between participating systems and our user simulators. All the resources are publicly available in our repository, including the system evaluation both as a result of the offline (i.e., test collection-based) and interactive tasks (i.e., user simulation-based), as well as their source code and model weights, to foster future research in this direction.

Deep research has emerged as an important task that aims to address hard queries that need extensive open-web exploration. To tackle it, most prior work equips large language model (LLM)-based agents with opaque web search APIs, enabling agents to iteratively issue search queries, retrieve external evidence, and reason over it. Despite search's essential role in deep research, black-box web search APIs leave the behaviour of established text ranking methods in deep research largely unclear. To fill this gap, we reproduce key findings and best practices for text ranking methods in deep research. We examine their effectiveness from three perspectives: (i) retrieval units (documents vs. passages), (ii) pipeline configurations (different retrievers, re-rankers, and re-ranking depths), and (iii) query characteristics (the mismatch between agent-issued queries and the training queries of text rankers). We perform experiments on BrowseComp-Plus, a deep research dataset with a fixed corpus, evaluating 2 open-source agents, 5 retrievers, and 3 re-rankers. We find that agent-issued queries typically follow web-search-style syntax (e.g., quoted exact matches), favouring lexical, learned sparse, and multi-vector retrievers; passage-level units are more efficient under limited context windows, and avoid the difficulties of document length normalisation in lexical retrieval; re-ranking is highly effective. We further propose a query-to-question (Q2Q) method that translates agent-issued queries into natural-language questions, significantly reducing the query mismatch.

The rise of personalized conversational search systems has been driven by advancements in Large Language Models (LLMs), enabling these systems to retrieve and generate answers for complex information needs. However, the automatic evaluation of responses generated by Retrieval Augmented Generation (RAG) systems remains an understudied challenge. In this paper, we introduce a new resource for assessing the retrieval effectiveness and relevance of responses generated by RAG systems, using a nugget-based evaluation framework. Built upon the foundation of TREC iKAT 2023, our dataset extends to the TREC iKAT 2024 collection, which includes 17 conversations and 20,575 relevance passage assessments, together with 2,279 extracted gold nuggets and 62 manually written gold answers from NIST assessors. While maintaining the core structure of its predecessor, this new collection enables a deeper exploration of generation tasks in conversational settings. Key improvements in iKAT 2024 include: (1) ''gold nuggets'' - concise, essential pieces of information extracted from relevant passages of the collection - which serve as a foundation for automatic response evaluation; (2) manually written answers to provide a gold standard for response evaluation; (3) expanded user personas, providing richer contextual grounding; and (4) a transition from Personal Text Knowledge Base (PTKB) ranking to PTKB classification and selection. Built on this resource, we provide a framework for long-form answer generation evaluation, involving nugget extraction and nugget matching, linked to retrieval. This establishes a solid resource for advancing research in personalized conversational search and long-form answer generation. Our resources are publicly available at https://github.com/irlabamsterdam/CONE-RAG.

Neural IR has advanced through two distinct paths: entity-oriented approaches leveraging knowledge graphs and multi-vector models capturing fine-grained semantics. We introduce QDER, a neural re-ranking model that unifies these approaches by integrating knowledge graph semantics into a multi-vector model.

We tackle the challenge of building real-world multimodal assistants for complex real-world tasks. We describe the practicalities and challenges of developing and deploying GRILLBot, a leading (first and second prize winning in 2022 and 2023) system deployed in the Alexa Prize TaskBot Challenge. Building on our Open Assistant Toolkit (OAT) framework, we propose a hybrid architecture that leverages Large Language Models (LLMs) and specialised models tuned for specific subtasks requiring very low latency. OAT allows us to define when, how and which LLMs should be used in a structured and deployable manner. For knowledge-grounded question answering and live task adaptations, we show that LLM reasoning abilities over task context and world knowledge outweigh latency concerns. For dialogue state management, we implement a code generation approach and show that specialised smaller models have 84% effectiveness with 100x lower latency. Overall, we provide insights and discuss tradeoffs for deploying both traditional models and LLMs to users in complex real-world multimodal environments in the Alexa TaskBot challenge. These experiences will continue to evolve as LLMs become more capable and efficient -- fundamentally reshaping OAT and future assistant architectures.

Conversational information seeking has evolved rapidly in the last few years with the development of Large Language Models (LLMs), providing the basis for interpreting and responding in a naturalistic manner to user requests. The extended TREC Interactive Knowledge Assistance Track (iKAT) collection aims to enable researchers to test and evaluate their Conversational Search Agent (CSA). The collection contains a set of 36 personalized dialogues over 20 different topics each coupled with a Personal Text Knowledge Base (PTKB) that defines the bespoke user personas. A total of 344 turns with approximately 26,000 passages are provided as assessments on relevance, as well as additional assessments on generated responses over four key dimensions: relevance, completeness, groundedness, and naturalness. The collection challenges CSAs to efficiently navigate diverse personal contexts, elicit pertinent persona information, and employ context for relevant conversations. The integration of a PTKB and the emphasis on decisional search tasks contribute to the uniqueness of this test collection, making it an essential benchmark for advancing research in conversational and interactive knowledge assistants.

Contains fulltext : 308793.pdf (Publisher’s version ) (Open Access)

We present two empirical studies to investigate users’ expectations and behaviours when using digital assistants, such as Alexa and Google Home, in a kitchen context: First, a survey (N = 200) queries participants on their expectations for the kinds of information that such systems should be able to provide. While consensus exists on expecting information about cooking steps and processes, younger participants who enjoy cooking express a higher likelihood of expecting details on food history or the science of cooking. In a follow-up Wizard-of-Oz study (N = 48), users were guided through the steps of a recipe either by an active wizard that alerted participants to information it could provide or a passive wizard who only answered questions that were provided by the user. The active policy led to almost double the number of conversational utterances and 1.5 times more knowledge-related user questions compared to the passive policy. Also, it resulted in 1.7 times more knowledge communicated than the passive policy. We discuss the findings in the context of related work and reveal implications for the design and use of such assistants for cooking and other purposes such as DIY and craft tasks, as well as the lessons we learned for evaluating such systems.

This tutorial will provide an overview of recent advances on neuro-symbolic approaches for information retrieval. A decade ago, knowledge graphs and semantic annotations technology led to active research on how to best leverage symbolic knowledge. At the same time, neural methods have demonstrated to be versatile and highly effective.

Current query expansion models use pseudo-relevance feedback to improve first-pass retrieval effectiveness; however, this fails when the initial results are not relevant. Instead of building a language model from retrieved results, we propose Generative Relevance Feedback (GRF) that builds probabilistic feedback models from long-form text generated from Large Language Models. We study the effective methods for generating text by varying the zero-shot generation subtasks: queries, entities, facts, news articles, documents, and essays. We evaluate GRF on document retrieval benchmarks covering a diverse set of queries and document collections, and the results show that GRF methods significantly outperform previous PRF methods. Specifically, we improve MAP between 5-19% and NDCG@10 17-24% compared to RM3 expansion, and achieve state-of-the-art recall across all datasets.

This research aims to explore various methods for assessing user feedback in mixed-initiative conversational search (CS ) systems. While CS systems enjoy profuse advancements across multiple aspects, recent research fails to successfully incorporate feedback from the users. One of the main reasons for that is the lack of system-user conversational interaction data. To this end, we propose a user simulator-based framework for multi-turn interactions with a variety of mixed-initiative CS systems. Specifically, we develop a user simulator, dubbed ConvSim, that, once initialized with an information need description, is capable of providing feedback to system's responses, as well as answering potential clarifying questions. Our experiments on a wide variety of state-of-the-art passage retrieval and neural re-ranking models show that effective utilization of user feedback can lead to 16% retrieval performance increase in terms of nDCG@3. Moreover, we observe consistent improvements as the number of feedback rounds increases (35% relative improvement in terms of nDCG@3 after three rounds). This points to a research gap in the development of specific feedback processing modules and opens a potential for significant advancements in CS. To support further research in the topic, we release over 30 000 transcripts of system-simulator interactions based on well-established CS datasets.

Conversational information seeking (CIS) involves interaction sequences between one or more users and an information system. Interactions in CIS are primarily based on natural language dialogue, while they may include other types of interactions, such as click, touch, and body gestures. CIS recently attracted significant attention and advancements continue to be made. This tutorial follows the content of the recent Conversational Information Seeking book authored by several of the tutorial presenters. The tutorial aims to be an introduction to CIS for newcomers to CIS in addition to the recent advanced topics and state-of-the-art approaches for students and researchers with moderate knowledge of the topic. A significant part of the tutorial is dedicated to hands-on experiences based on toolkits developed by the presenters for conversational passage retrieval and multi-modal task-oriented dialogues. The outcomes of this tutorial include theoretical and practical knowledge, including a forum to meet researchers interested in CIS.

CODEC is a document and entity ranking benchmark that focuses on complex research topics. We target essay-style information needs of social science researchers, i.e. "How has the UK's Open Banking Regulation benefited Challenger Banks". CODEC includes 42 topics developed by researchers and a new focused web corpus with semantic annotations including entity links. This resource includes expert judgments on 17,509 documents and entities (416.9 per topic) from diverse automatic and interactive manual runs. The manual runs include 387 query reformulations, providing data for query performance prediction and automatic rewriting evaluation.

The role of conversational assistants continues to evolve, beyond simple voice commands to ones that support rich and complex tasks in the home, car, and even virtual reality. Going beyond simple voice command and control requires agents and datasets blending structured dialogue, information seeking, grounded reasoning, and contextual question-answering in a multimodal environment with rich image and video content. In this demo, we introduce Task-oriented Multimodal Agent Dialogue (TaskMAD), a new platform that supports the creation of interactive multimodal and task-centric datasets in a Wizard-of-Oz experimental setup. TaskMAD includes support for text and voice, federated retrieval from text and knowledge bases, and structured logging of interactions for offline labeling. Its architecture supports a spectrum of tasks that span open-domain exploratory search to traditional frame-based dialogue tasks. It's open-source and offers rich capability as a platform used to collect data for the Amazon Alexa Prize Taskbot challenge, TREC Conversational Assistance track, undergraduate student research, and others. TaskMAD is distributed under the MIT license.

Task-based Virtual Personal Assistants (VPAs) rely on multi-domain Dialogue State Tracking (DST) models to monitor goals throughout a conversation. Previously proposed models show promising results on established benchmarks, but they have difficulty adapting to unseen domains due to domain-specific parameters in their model architectures. We propose a new Similarity-based Multi-domain Dialogue State Tracking model (SM-DST) that uses retrieval-inspired and fine-grained contextual token-level similarity approaches to efficiently and effectively track dialogue state. The key difference with state-of-the-art DST models is that SM-DST has a single model with shared parameters across domains and slots. Because we base SM-DST on similarity it allows the transfer of tracking information between semantically related domains as well as to unseen domains without retraining. Furthermore, we leverage copy mechanisms that consider the system’s response and the dialogue state from previous turn predictions, allowing it to more effectively track dialogue state for complex conversations. We evaluate SM-DST on three variants of the MultiWOZ DST benchmark datasets. The results demonstrate that SM-DST significantly and consistently outperforms state-of-the-art models across all datasets by absolute 5-18% and 3-25% in the few- and zero-shot settings, respectively.

Understanding and comparing the behavior of retrieval models is a fundamental challenge that requires going beyond examining average effectiveness and per-query metrics, because these do not reveal key differences in how ranking models' behavior impacts individual results. DiffIR is a new open-source web tool to assist with qualitative ranking analysis by visually 'diffing' system rankings at the individual result level for queries where behavior significantly diverges. Using one of several configurable similarity measures, it identifies queries for which the rankings of models compared have important differences in individual rankings and provides a visual web interface to compare the rankings side-by-side. DiffIR additionally supports a model-specific visualization approach based on custom term importance weight files. These support studying the behavior of interpretable models, such as neural retrieval methods that produce document scores based on a similarity matrix or based on a single document passage. Observations from this tool can complement neural probing approaches like ABNIRML to generate quantitative tests. We provide an illustrative use case of DiffIR by studying the qualitative differences between recently developed neural ranking models on a standard TREC benchmark dataset.

Deep Learning Hard (DL-HARD) is a new annotated dataset designed to more effectively evaluate neural ranking models on complex topics. It builds on TREC Deep Learning (DL) topics by extensively annotating them with question intent categories, answer types, wikified entities, topic categories, and result type metadata from a commercial web search engine. Based on this data, we introduce a framework for identifying challenging queries. DL-HARD contains fifty topics from the official DL 2019/2020 evaluation benchmark, half of which are newly and independently assessed. We perform experiments using the official submitted runs to DL on DL-HARD and find substantial differences in metrics and the ranking of participating systems. Overall, DL-HARD is a new resource that promotes research on neural ranking methods by focusing on challenging and complex topics.

CAsT-19 is a new dataset that supports research on conversational information seeking. The corpus is 38,426,252 passages from the TREC Complex Answer Retrieval (CAR) and Microsoft MAchine Reading COmprehension (MARCO) datasets. Eighty information seeking dialogues (30 train, 50 test) are an average of 9 to 10 questions long. A dialogue may explore a topic broadly or drill down into subtopics. Questions contain ellipsis, implied context, mild topic shifts, and other characteristics of human conversation that may prevent them from being understood in isolation. Relevance assessments are provided for 30 training topics and 20 test topics.

Conversational Information Seeking (CIS) is an emerging area of Information Retrieval focused on interactive search systems. As a result there is a need for new benchmark datasets and tools to enable their creation. In this demo we present the Agent Dialogue (AD) platform, an open-source system developed for researchers to perform Wizard-of-Oz CIS experiments. AD is a scalable cloud-native platform developed with Docker and Kubernetes with a flexible and modular micro-service architecture built on production-grade state-of-the-art open-source tools (Kubernetes, gRPC streaming, React, and Firebase). It supports varied front-ends and has the ability to interface with multiple existing agent systems, including Google Assistant and open-source search libraries. It includes support for centralized structure logging as well as offline relevance annotation.

Tools capable of automatic code generation have the potential to augment programmer's capabilities. While straightforward code retrieval is incorporated into many IDEs, an emerging area is explicit code generation. Code generation is currently approached as a Machine Translation task, with Recurrent Neural Network (RNN) based encoder-decoder architectures trained on code-description pairs. In this work we introduce and study modern Transformer architectures for this task. We further propose a new model called the Relevance Transformer that incorporates external knowledge using pseudo-relevance feedback. The Relevance Transformer biases the decoding process to be similar to existing retrieved code while enforcing diversity. We perform experiments on multiple standard benchmark datasets for code generation including Django, Hearthstone, and CoNaLa. The results show improvements over state-of-the-art methods based on BLEU evaluation. The Relevance Transformer model shows the potential of Transformer-based architectures for code generation and introduces a method of incorporating pseudo-relevance feedback during inference.

Task-based Virtual Personal Assistants (VPAs) such as the Google Assistant, Alexa, and Siri are increasingly being adopted for a wide variety of tasks. These tasks are grounded in real-world entities and actions (e.g., book a hotel, organise a conference, or requesting funds). In this work we tackle the task of automatically constructing actionable knowledge graphs in response to a user query in order to support a wider variety of increasingly complex assistant tasks. We frame this as an entity property ranking task given a user query with annotated properties. We propose a new method for property ranking, CrossBERT. CrossBERT builds on the Bidirectional Encoder Representations from Transformers (BERT) and creates a new triplet network structure on cross query-property pairs that is used to rank properties. We also study the impact of using external evidence for query entities from textual entity descriptions. We perform experiments on two standard benchmark collections, the NTCIR-13 Actionable Knowledge Graph Generation (AKGG) task and Entity Property Identification (EPI) task. The results demonstrate that CrossBERT significantly outperforms the best performing runs from AKGG and EPI, as well as previous state-of-the-art BERT-based models. In particular, CrossBERT significantly improves Recall and NDCG by approximately 2-12% over the BERT models across the two used datasets.

Conversational search and recommendation systems that use natural language interfaces are an increasingly important area raising a number of research and interface design questions. Despite the increasing popularity of digital personal assistants, the number of conversational recommendation systems is limited and their functionality basic. In this demonstration we introduce Vote Goat, a conversational recommendation agent built using Google's DialogFlow framework. The demonstration provides an interactive movie recommendation system using a speech-based natural language interface. The main intents span search and recommendation tasks including: rating movies, receiving recommendations, retrieval over movie metadata, and viewing crowdsourced statistics. Vote Goat uses gamification to incentivize movie voting interactions with the 'Greatest Of All Time' (GOAT) movies derived from user ratings. The demo includes important functionality for research applications with logging of interactions for building test collections as well as A/B testing to allow researchers to experiment with system parameters.

Semantic technologies such as controlled vocabularies, thesauri, and knowledge graphs have been used throughout the history of information retrieval for a variety of tasks. Recent advances in knowledge acquisition, alignment, and utilization have given rise to a body of new approaches for utilizing knowledge graphs in text retrieval tasks and it is therefore time to consolidate the community efforts and study how such technologies can be employed in information retrieval systems in the most effective way. It is also time to start and deepen the dialogue between researchers and practitioners in order to ensure that breakthroughs, technologies, and algorithms in this space are widely disseminated. The goal of this workshop is to bring together and grow a community of researchers and practitioners who are interested in using, aligning, and constructing knowledge graphs and similar semantic resources for information retrieval applications.

The amount of structured content published on the Web has been growing rapidly, making it possible to address increasingly complex information access tasks. Recent years have witnessed the emergence of large scale human-curated knowledge bases as well as a growing array of techniques that identify or extract information automatically from unstructured and semi-structured sources. The ESAIR workshop series aims to advance the general research agenda on the problem of creating and exploiting semantic annotations. The eighth edition of ESAIR sets its focus on applications. We dedicate a special "annotations in action" track to demonstrations that showcase innovative prototype systems, in addition to the regular research and position paper contributions. The workshop also features invited talks from leaders in the field. The desired outcome of ESAIR'15 is a roadmap and research agenda that guides academic efforts and aligns them with industrial directions and developments.

Recent advances in automatic entity linking and knowledge base construction have resulted in entity annotations for document and query collections. For example, annotations of entities from large general purpose knowledge bases, such as Freebase and the Google Knowledge Graph. Understanding how to leverage these entity annotations of text to improve ad hoc document retrieval is an open research area. Query expansion is a commonly used technique to improve retrieval effectiveness. Most previous query expansion approaches focus on text, mainly using unigram concepts. In this paper, we propose a new technique, called entity query feature expansion (EQFE) which enriches the query with features from entities and their links to knowledge bases, including structured attributes and text. We experiment using both explicit query entity annotations and latent entities. We evaluate our technique on TREC text collections automatically annotated with knowledge base entity links, including the Google Freebase Annotations (FACC1) data. We find that entity-based feature expansion results in significant improvements in retrieval effectiveness over state-of-the-art text expansion approaches.

Recent research in video retrieval has been successful at finding videos when the query consists of tens or hundreds of sample relevant videos for training supervised models. Instead, we investigate unsupervised zero-shot retrieval where no training videos are provided: a query consists only of a text statement. For retrieval, we use text extracted from images in the videos, text recognized in the speech of its audio track, as well as automatically detected semantically meaningful visual video concepts identified with widely varying confidence in the videos. In this work we introduce a new method for automatically identifying relevant concepts given a text query using the Markov Random Field (MRF) retrieval framework. We use source expansion to build rich textual representations of semantic video concepts from large external sources such as the web. We find that concept-based retrieval significantly outperforms text based approaches in recall. Using an evaluation derived from the TRECVID MED'11 track, we present early results that an approach using multi-modal fusion can compensate for inadequacies in each modality, resulting in substantial effectiveness gains. With relevance feedback, our approach provides additional improvements of over 50%.

We present the architecture behind Twitter's real-time related query suggestion and spelling correction service. Although these tasks have received much attention in the web search literature, the Twitter context introduces a real-time "twist": after significant breaking news events, we aim to provide relevant results within minutes. This paper provides a case study illustrating the challenges of real-time data processing in the era of "big data". We tell the story of how our system was built twice: our first implementation was built on a typical Hadoop-based analytics stack, but was later replaced because it did not meet the latency requirements necessary to generate meaningful real-time results. The second implementation, which is the system deployed in production today, is a custom in-memory processing engine specifically designed for the task. This experience taught us that the current typical usage of Hadoop as a "big data" platform, while great for experimentation, is not well suited to low-latency processing, and points the way to future work on data analytics platforms that can handle "big" as well as "fast" data.

Conventional retrieval systems view documents as a unit and look at different retrieval types within a document. We introduce Proteus, a frame-work for seamlessly navigating books as dynamic collections which are defined on the fly. Proteus allows us to search various retrieval types. Navigable types include pages, books, named persons, locations, and pictures in a collection of books taken from the Internet Archive. The demonstration shows the value of multi-type browsing in dynamic collections to peruse new data.

Many forms of linguistic analysis, such as part of speech tagging, named entity recognition, and other sequence labeling tasks are performed on short spans of text and assume statistical dependence within a window of only a few tokens. We propose using passage retrieval to induce non-local dependencies in structured classification that generalizes earlier work in context aggregation for named-entity recognition. We introduce a new method for feature expansion inspired by psuedo-relevance feedback (PRF). Our results on the CoNLL 2003 task show that features from cross-document feature expansion improves NER effectiveness over previous aggregation models. Utilizing all the tokens in a sentence for query context consistently perform best on both intrinsic and extrinsic evaluations. Tagging models incorporating feature expansion outperform the leading NER system when evaluated on out of domain data, a collection of publicly available scanned books on the topic of historic Deerfield, MA. Finally, the results show that retrieval based feature expansion using an external collection of unlabeled text can result in further effectiveness improvements.

As user demands become increasingly sophisticated, search engines today are competing in more than just returning document results from the Web. One area of competition is providing web object results from structured data extracted from a multitude of information sources. We address the problem of performing keyword retrieval over a collection of objects containing a large degree of duplication as different Web-based information sources provide descriptions of the same object. We develop a method for coreference aware retrieval that performs topic-specific coreference resolution on retrieved objects in order to improve object search results. Our results demonstrate that coreference has a significant impact on the effectiveness of retrieval in the domain of local search. Our results show that a coreference aware system outperforms naive object retrieval by more than 20% in P5 and P10.

A document or web page in isolation may appear completely reasonable, but may represent a biased perspective on the topic being discussed. Given the topic of a document, we propose new metrics provocativeness and balance that suggest when the topic could be controversial. We explore the use of these metrics to characterize the subjectivity of the topics in the TREC Blog Track.