Search

The use of large language models (LLMs) for relevance assessment in information retrieval has gained significant attention, with recent studies suggesting that LLM-based judgments provide comparable evaluations to human judgments. Notably, based on TREC 2024 data, Upadhyay et al. make a bold claim that LLM-based relevance assessments, such as those generated by the UMBRELA system, can fully replace traditional human relevance assessments in TREC-style evaluations. This paper critically examines this claim, highlighting practical and theoretical limitations that undermine the validity of this conclusion. First, we question whether the evidence provided by Upadhyay et al. really supports their claim, particularly if a test collection is used asa benchmark for future improvements. Second, through a submission deliberately intended to do so, we demonstrate the ease with which automatic evaluation metrics can be subverted, showing that systems designed to exploit these evaluations can achieve artificially high scores. Theoretical challenges -- such as the inherent narcissism of LLMs, the risk of overfitting to LLM-based metrics, and the potential degradation of future LLM performance -- must be addressed before LLM-based relevance assessments can be considered a viable replacement for human judgments.

Traditional evaluation of information retrieval (IR) systems relies on human-annotated relevance labels, which can be both biased and costly at scale. In this context, large language models (LLMs) offer an alternative by allowing us to directly prompt them to assign relevance labels for passages associated with each query. In this study, we explore alternative methods to directly prompt LLMs for assigned relevance labels, by exploring two hypotheses: Hypothesis 1 assumes that it is helpful to break down "relevance" into specific criteria - exactness, coverage, topicality, and contextual fit. We explore different approaches that prompt large language models (LLMs) to obtain criteria-level grades for all passages, and we consider various ways to aggregate criteria-level grades into a relevance label. Hypothesis 2 assumes that differences in linguistic style between queries and passages may negatively impact the automatic relevance label prediction. We explore whether improvements can be achieved by first synthesizing a summary of the passage in the linguistic style of a query, and then using this summary in place of the passage to assess its relevance. We include an empirical evaluation of our approaches based on data from the LLMJudge challenge run in Summer 2024, where our "Four Prompts" approach obtained the highest scores in Kendall's tau.

This resource paper addresses the challenge of evaluating Information Retrieval (IR) systems in the era of autoregressive Large Language Models (LLMs). Traditional methods relying on passage-level judgments are no longer effective due to the diversity of responses generated by LLM-based systems. We provide a workbench to explore several alternative evaluation approaches to judge the relevance of a system's response that incorporate LLMs: 1. Asking an LLM whether the response is relevant; 2. Asking the LLM which set of nuggets (i.e., relevant key facts) is covered in the response; 3. Asking the LLM to answer a set of exam questions with the response. This workbench aims to facilitate the development of new, reusable test collections. Researchers can manually refine sets of nuggets and exam questions, observing their impact on system evaluation and leaderboard rankings. Resource available at https://github.com/TREMA-UNH/autograding-workbench, Comment: 10 pages. To appear in the Resource & Reproducibility Track of SIGIR 2024

Current IR evaluation is based on relevance judgments, created either manually or automatically, with decisions outsourced to Large Language Models (LLMs). We offer an alternative paradigm, that never relies on relevance judgments in any form. Instead, a text is defined as relevant if it contains information that enables the answering of key questions. We use this idea to design the EXAM Answerability Metric to evaluate information retrieval/generation systems for their ability to provide topically relevant information. We envision the role of a human judge to edit and define an exam question bank that will test for the presence of relevant information in text. We support this step by generating an initial set of exam questions. In the next phase, an LLM-based question answering system will automatically grade system responses by tracking which exam questions are answerable with which system responses. We propose two evaluation measures, the recall-oriented EXAM Cover metric, and the precision-oriented EXAM Qrels metric, the latter which can be implemented with trec_eval. This paradigm not only allows for the expansion of the exam question set post-hoc but also facilitates the ongoing evaluation of future information systems, whether they focus on retrieval, generation, or both., Comment: 10 pages

Time series forecasting is a challenging task due to the existence of complex and dynamic temporal dependencies. This can lead to incorrect predictions by even the best forecasting models. Using more training data is one way to improve the accuracy, but this source is often limited. In contrast, we are building on successful denoising approaches for image generation by advocating for an end-to-end forecasting and denoising paradigm. We propose an end-to-end forecast-blur-denoise forecasting framework by encouraging a division of labors between the forecasting and the denoising models. The initial forecasting model is directed to focus on accurately predicting the coarse-grained behavior, while the denoiser model focuses on capturing the fine-grained behavior that is locally blurred by integrating a Gaussian Process model. All three parts are interacting for the best end-to-end performance. Our extensive experiments demonstrate that our proposed approach is able to improve the forecasting accuracy of several state-of-the-art forecasting models as well as several other denoising approaches., Comment: 10 pages

Query-specific article generation is the task of, given a search query, generate a single article that gives an overview of the topic. We envision such articles as an alternative to presenting a ranking of search results. While generative Large Language Models (LLMs) like chatGPT also address this task, they are known to hallucinate new information, their models are secret, hard to analyze and control. Some generative LLMs provide supporting references, yet these are often unrelated to the generated content. As an alternative, we propose to study article generation systems that integrate document retrieval, query-specific clustering, and summarization. By design, such models can provide actual citations as provenance for their generated text. In particular, we contribute an evaluation framework that allows to separately trains and evaluate each of these three components before combining them into one system. We experimentally demonstrate that a system comprised of the best-performing individual components also obtains the best F-1 overall system quality., Comment: 5 pages, 1 figures

When asked, large language models (LLMs) like ChatGPT claim that they can assist with relevance judgments but it is not clear whether automated judgments can reliably be used in evaluations of retrieval systems. In this perspectives paper, we discuss possible ways for LLMs to support relevance judgments along with concerns and issues that arise. We devise a human--machine collaboration spectrum that allows to categorize different relevance judgment strategies, based on how much humans rely on machines. For the extreme point of "fully automated judgments", we further include a pilot experiment on whether LLM-based relevance judgments correlate with judgments from trained human assessors. We conclude the paper by providing opposing perspectives for and against the use of~LLMs for automatic relevance judgments, and a compromise perspective, informed by our analyses of the literature, our preliminary experimental evidence, and our experience as IR researchers.

The vision of HIPstIR is that early stage information retrieval (IR) researchers get together to develop a future for non-mainstream ideas and research agendas in IR. The first iteration of this vision materialized in the form of a three day workshop in Portsmouth, New Hampshire attended by 24 researchers across academia and industry. Attendees pre-submitted one or more topics that they want to pitch at the meeting. Then over the three days during the workshop, we self-organized into groups and worked on six specific proposals of common interest. In this report, we present an overview of the workshop and brief summaries of the six proposals that resulted from the workshop.

We investigate the problem of understanding the message (gist) conveyed by images and their captions as found, for instance, on websites or news articles. To this end, we propose a methodology to capture the meaning of image-caption pairs on the basis of large amounts of machine-readable knowledge that has previously been shown to be highly effective for text understanding. Our method identifies the connotation of objects beyond their denotation: where most approaches to image understanding focus on the denotation of objects, i.e., their literal meaning, our work addresses the identification of connotations, i.e., iconic meanings of objects, to understand the message of images. We view image understanding as the task of representing an image-caption pair on the basis of a wide-coverage vocabulary of concepts such as the one provided by Wikipedia, and cast gist detection as a concept-ranking problem with image-caption pairs as queries. To enable a thorough investigation of the problem of gist understanding, we produce a gold standard of over 300 image-caption pairs and over 8,000 gist annotations covering a wide variety of topics at different levels of abstraction. We use this dataset to experimentally benchmark the contribution of signals from heterogeneous sources, namely image and text. The best result with a Mean Average Precision (MAP) of 0.69 indicate that by combining both dimensions we are able to better understand the meaning of our image-caption pairs than when using language or vision information alone. We test the robustness of our gist detection approach when receiving automatically generated input, i.e., using automatically generated image tags or generated captions, and prove the feasibility of an end-to-end automated process.

Nowadays, where multimedia data is continuously generated, stored, and distributed, multimedia indexing, with its purpose of group- ing similar data, becomes more important than ever. Understanding the gist (=message) of multimedia instances is framed in related work as a ranking of concepts from a knowledge base, i.e., Wikipedia. We cast the task of multimedia indexing as a gist understanding problem. Our pipeline benefits from external knowledge and two subsequent learning- to-rank (l2r) settings. The first l2r produces a ranking of concepts rep- resenting the respective multimedia instance. The second l2r produces a mapping between the concept representation of an instance and the targeted class topic(s) for the multimedia indexing task. The evaluation on an established big size corpus (MIRFlickr25k, with 25,000 images), shows that multimedia indexing benefits from understanding the gist. Finally, with a MAP of 61.42, it can be shown that the multimedia in- dexing task benefits from understanding the gist. Thus, the presented end-to-end setting outperforms DBM and competes with Hashing-based methods.

The sequential dependence model (SDM) is a popular retrieval model which is based on the theory of probabilistic graphical models. While it was originally introduced by Metzler and Croft as a Markov Random Field (aka discriminative probabilistic model), in this paper we demonstrate that it is equivalent to a generative probabilistic model. To build an foundation for future retrieval models, this paper details the axiomatic underpinning of the SDM model as discriminative and generative probabilistic model. The only difference arises whether model parameters are estimated in log-space or Multinomial-space. We demonstrate that parameter-estimation with grid-tuning is negatively impacting the generative formulation, an effect that vanishes when parameters are estimated with coordinate-gradient descent. This is concerning, since empirical differences may be falsely attributed to improved models., Comment: SIGIR'17 Workshop on Axiomatic Thinking for Information Retrieval and Related Tasks (ATIR)

Retrieving paragraphs to populate a Wikipedia article is a challenging task. The new TREC Complex Answer Retrieval (TREC CAR) track introduces a comprehensive dataset that targets this retrieval scenario. We present early results from a variety of approaches -- from standard information retrieval methods (e.g., tf-idf) to complex systems that using query expansion using knowledge bases and deep neural networks. The goal is to offer future participants of this track an overview of some promising approaches to tackle this problem.