With the growth of misinformation on the web, automated fact checking has garnered immense interest for detecting growing misinformation and disinformation. Current systems have made significant advancements in handling synthetic claims sourced from Wikipedia, and noteworthy progress has been achieved in addressing real-world claims that are verified by fact-checking organizations as well. We compile and release QuanTemp, a diverse, multi-domain dataset focused exclusively on numerical claims, encompassing comparative, statistical, interval, and temporal aspects, with detailed metadata and an accompanying evidence collection. This addresses the challenge of verifying real-world numerical claims, which are complex and often lack precise information, a gap not filled by existing works that mainly focus on synthetic claims. We evaluate and quantify these gaps in existing solutions for the task of verifying numerical claims. We also evaluate claim decomposition based methods, numerical understanding based natural language inference (NLI) models and our best baselines achieves a macro-F1 of 58.32. This demonstrates that QuanTemp serves as a challenging evaluation set for numerical claim verification.

@en

Large language models (LLMs) have recently gained significant attention due to their unparalleled zero-shot performance on various natural language processing tasks. However, the pre-Training data utilized in LLMs is often confined to a specific corpus, resulting in inherent freshness and temporal scope limitations. Consequently, this raises concerns regarding the effectiveness of LLMs for tasks involving temporal intents. In this study, we aim to investigate the underlying limitations of general-purpose LLMs when deployed for tasks that require a temporal understanding. We pay particular attention to handling factual temporal knowledge through three popular temporal QA datasets. Specifically, we observe low performance on detailed questions about the past and, surprisingly, for rather new information. In manual and automatic testing, we find multiple temporal errors and characterize the conditions under which QA performance deteriorates. Our analysis contributes to understanding LLM limitations and offers valuable insights into developing future models that can better cater to the demands of temporally-oriented tasks. The code is available https://github.com/jwallat/temporalblindspots.

@en

An significant challenge in text-ranking systems is handling hard queries that form the tail end of the query distribution. Difficulty may arise due to the presence of uncommon, underspecified, or incomplete queries. In this work, we improve the ranking performance of hard or difficult queries while maintaining the performance of other queries. Firstly, we do LLM-based query enrichment for training queries using relevant documents. Next, a specialized ranker is fine-tuned only on the enriched hard queries instead of the original queries. We combine the relevance scores from the specialized ranker and the base ranker, along with a query performance score estimated for each query. Our approach departs from existing methods that usually employ a single ranker for all queries, which is biased towards easy queries, which form the majority of the query distribution. In our extensive experiments on the DL-Hard dataset, we find that a principled query performance based scoring method using base and specialized ranker offers a significant improvement of up to 48.4% on the document ranking task and up to 25% on the passage ranking task compared to the baseline performance of using original queries, even outperforming SOTA model.@en

With the rapid advancement of neural language models, the deployment of overparameterized models has surged, increasing the need for interpretable explanations comprehensible to human inspectors. Existing post-hoc interpretability methods, which often focus on unigram features of single input textual instances, fail to capture the models’ decision-making process fully. Additionally, many methods do not differentiate between decisions based on spurious correlations and those based on a holistic understanding of the input. Our paper introduces DISCO, a novel method for discovering global, rule-based explanations by identifying causal n-gram associations with model predictions. This method employs a scalable sequence mining technique to extract relevant text spans from training data, associate them with model predictions, and conduct causality checks to distill robust rules that elucidate model behavior. These rules expose potential overfitting and provide insights into misleading feature combinations. We validate DISCO through extensive testing, demonstrating its superiority over existing methods in offering comprehensive insights into complex model behaviors. Our approach successfully identifies all shortcuts manually introduced into the training data (100% detection rate on the MultiRC dataset), resulting in an 18.8% regression in model performance—a capability unmatched by any other method. Furthermore, DISCO supports interactive explanations, enabling human inspectors to distinguish spurious causes in the rule-based output. This alleviates the burden of abundant instance-wise explanations and helps assess the model’s risk when encountering out-of-distribution (OOD) data.

@en

Dual-encoder-based dense retrieval models have become the standard in IR. They employ large Transformer-based language models, which are notoriously inefficient in terms of resources and latency.We propose Fast-Forward indexes - vector forward indexes which exploit the semantic matching capabilities of dual-encoder models for efficient and effective re-ranking. Our framework enables re-ranking at very high retrieval depths and combines the merits of both lexical and semantic matching via score interpolation. Furthermore, in order to mitigate the limitations of dual-encoders, we tackle two main challenges: Firstly, we improve computational efficiency by either pre-computing representations, avoiding unnecessary computations altogether, or reducing the complexity of encoders. This allows us to considerably improve ranking efficiency and latency. Secondly, we optimize the memory footprint and maintenance cost of indexes; we propose two complementary techniques to reduce the index size and show that, by dynamically dropping irrelevant document tokens, the index maintenance efficiency can be improved substantially.We perform an evaluation to show the effectiveness and efficiency of Fast-Forward indexes - our method has low latency and achieves competitive results without the need for hardware acceleration, such as GPUs.

@en

Dual encoders are highly effective and widely deployed in the retrieval phase for passage and document ranking, question answering, or retrieval-augmented generation (RAG) setups. Most dual-encoder models use transformer models like BERT to map input queries and output targets to a common vector space encoding the semantic similarity. Despite their prevalence and impressive performance, little is known about the inner workings of dense encoders for retrieval. We investigate neural retrievers using the probing paradigm to identify well-understood IR properties that causally result in ranking performance. Unlike existing works that have probed cross-encoders to show query-document interactions, we provide a principled approach to probe dual-encoders. Importantly, we employ causal probing to avoid correlation effects that might be artefacts of vanilla probing. We conduct extensive experiments on one such dual encoder (TCT-ColBERT) to check for the existence and relevance of six properties: term importance, lexical matching (BM25), semantic matching, question classification, and the two linguistic properties of named entity recognition and coreference resolution. Our layer-wise analysis shows important differences between re-rankers and dual encoders, establishing which tasks are not only understood by the model but also used for inference.

@en

As information retrieval systems continue to evolve, accurate evaluation and benchmarking of these systems become pivotal. Web search datasets, such as MS MARCO, primarily provide short keyword queries without accompanying intent or descriptions, posing a challenge in comprehending the underlying information need. This paper proposes an approach to augmenting such datasets to annotate informative query descriptions, with a focus on two prominent benchmark datasets: TREC-DL-21 and TREC-DL-22. Our methodology involves utilizing state-of-the-art LLMs to analyze and comprehend the implicit intent within individual queries from benchmark datasets. By extracting key semantic elements, we construct detailed and contextually rich descriptions for these queries. To validate the generated query descriptions, we employ crowdsourcing as a reliable means of obtaining diverse human perspectives on the accuracy and informativeness of the descriptions. This information can be used as an evaluation set for tasks such as ranking, query rewriting, or others.@en

Contextual ranking models have delivered impressive performance improvements over classical models in the document ranking task. However, these highly over-parameterized models tend to be data-hungry and require large amounts of data even for fine-tuning. In this article, we propose data-augmentation methods for effective and robust ranking performance. One of the key benefits of using data augmentation is in achieving sample efficiency or learning effectively when we have only a small amount of training data. We propose supervised and unsupervised data augmentation schemes by creating training data using parts of the relevant documents in the query-document pairs. We then adapt a family of contrastive losses for the document ranking task that can exploit the augmented data to learn an effective ranking model. Our extensive experiments on subsets of the MS MARCO and TREC-DL test sets show that data augmentation, along with the ranking-adapted contrastive losses, results in performance improvements under most dataset sizes. Apart from sample efficiency, we conclusively show that data augmentation results in robust models when transferred to out-of-domain benchmarks. Our performance improvements in in-domain and more prominently in out-of-domain benchmarks show that augmentation regularizes the ranking model and improves its robustness and generalization capability.

@en

Answering complex questions is a challenging task that requires question decomposition and multistep reasoning for arriving at the solution. While existing supervised and unsupervised approaches are specialized to a certain task and involve training, recently proposed prompt-based approaches offer generalizable solutions to tackle a wide variety of complex question-answering (QA) tasks. However, existing prompt-based approaches that are effective for complex QA tasks involve expensive hand annotations from experts in the form of rationales and are not generalizable to newer complex QA scenarios and tasks. We propose, icat (In-Context Ability Transfer) which induces reasoning capabilities in LLMs without any LLM fine-tuning or manual annotation of in-context samples. We transfer the ability to decompose complex questions to simpler questions or generate step-by-step rationales to LLMs, by careful selection from available data sources of related tasks. We also propose an automated uncertainty-aware exemplar selection approach for selecting examples from transfer data sources. Finally, we conduct large-scale experiments on a variety of complex QA tasks involving numerical reasoning, compositional complex QA, and heterogeneous complex QA which require decomposed reasoning. We show that ICAT convincingly outperforms existing prompt-based solutions without involving any model training, showcasing the benefits of re-using existing abilities.@en

Contextual models like BERT are highly effective in numerous text-ranking tasks. However, it is still unclear as to whether contextual models understand well-established notions of relevance that are central to IR. In this paper, we use probing, a recent approach used to analyze language models, to investigate the ranking abilities of BERT-based rankers. Most of the probing literature has focussed on linguistic and knowledge-aware capabilities of models or axiomatic analysis of ranking models. In this paper, we fill an important gap in the information retrieval literature by conducting a layer-wise probing analysis using four probes based on lexical matching, semantic similarity as well as linguistic properties like coreference resolution and named entity recognition. Our experiments show an interesting trend that BERT-rankers better encode ranking abilities at intermediate layers. Based on our observations, we train a ranking model by augmenting the ranking data with the probe data to show initial yet consistent performance improvements (The code is available at https://github.com/yolomeus/probing-search/ ).

@en

Configurable software systems have become increasingly popular as they enable customized software variants. The main challenge in dealing with configuration problems is that the number of possible configurations grows exponentially as the number of features increases. Therefore, algorithms for testing customized software have to deal with the challenge of tractably finding potentially faulty configurations given exponentially large configurations. To overcome this problem, prior works focused on sampling strategies to significantly reduce the number of generated configurations, guaranteeing a high t-wise coverage. In this work, we address the configuration sampling problem by proposing a deep reinforcement learning (DRL) based sampler that efficiently finds the trade-off between exploration and exploitation, allowing for the efficient identification of a minimal subset of configurations that covers all t-wise feature interactions while minimizing redundancy. We also present the CS-Gym, an environment for the configuration sampling. We benchmark our results against heuristic-based sampling methods on eight different feature models of software product lines and show that our method outperforms all sampling methods in terms of sample size. Our findings indicate that the achieved improvement has major implications for cost reduction, as the reduction in sample size results in fewer configurations that need to be tested.@en

This tutorial presents explainable information retrieval (ExIR), an emerging area focused on fostering responsible and trustworthy deployment of machine learning systems in the context of information retrieval. As the field has rapidly evolved in the past 4-5 years, numerous approaches have been proposed that focus on different access modes, stakeholders, and model development stages. This tutorial aims to introduce IR-centric notions, classification, and evaluation styles in ExIR, while focusing on IR-specific tasks such as ranking, text classification, and learning-to-rank systems. We will delve into method families and their adaptations to IR, extensively covering post-hoc methods, axiomatic and probing approaches, and recent advances in interpretability-by-design approaches. We will also discuss ExIR applications for different stakeholders, such as researchers, practitioners, and end-users, in contexts like web search, patent and legal search, and high-stakes decision-making tasks. To facilitate practical understanding, we will provide a hands-on session on applying ExIR methods, reducing the entry barrier for students, researchers, and practitioners alike.

@en

Neural document ranking models perform impressively well due to superior language understanding gained from pre-Training tasks. However, due to their complexity and large number of parameters these (typically transformer-based) models are often non-interpretable in that ranking decisions can not be clearly attributed to specific parts of the input documents.In this article, we propose ranking models that are inherently interpretable by generating explanations as a by-product of the prediction decision. We introduce the Select-And-Rank paradigm for document ranking, where we first output an explanation as a selected subset of sentences in a document. Thereafter, we solely use the explanation or selection to make the prediction, making explanations first-class citizens in the ranking process. Technically, we treat sentence selection as a latent variable trained jointly with the ranker from the final output. To that end, we propose an end-To-end training technique for Select-And-Rank models utilizing reparameterizable subset sampling using the Gumbel-max trick.We conduct extensive experiments to demonstrate that our approach is competitive to state-of-The-Art methods. Our approach is broadly applicable to numerous ranking tasks and furthers the goal of building models that are interpretable by design. Finally, we present real-world applications that benefit from our sentence selection method.

@en

This paper proposes a novel approach towards better interpretability of a trained text-based ranking model in a post-hoc manner. A popular approach for post-hoc interpretability text ranking models are based on locally approximating the model behavior using a simple ranker. Since rankings have multiple relevance factors and are aggregations of predictions, existing approaches that use a single ranker might not be sufficient to approximate a complex model, resulting in low fidelity. In this paper, we overcome this problem by considering multiple simple rankers to better approximate the entire ranking list from a black-box ranking model. We pose the problem of local approximation as a Generalized Preference Coverage (GPC) problem that incorporates multiple simple rankers towards the listwise explanation of ranking models. Our method Multiplex uses a linear programming approach to judiciously extract the explanation terms, so that to explain the entire ranking list. We conduct extensive experiments on a variety of ranking models and report fidelity improvements of 37%–54% over existing competitors. We finally compare explanations in terms of multiple relevance factors and topic aspects to better understand the logic of ranking decisions, showcasing our explainers’ practical utility.

@en

Pre-trained contextual language models such as BERT, GPT, and XLnet work quite well for document retrieval tasks. Such models are fine-tuned based on the query-document/query-passage level relevance labels to capture the ranking signals. However, the documents are longer than the passages and such document ranking models suffer from the token limitation (512) of BERT. Researchers proposed ranking strategies that either truncate the documents beyond the token limit or chunk the documents into units that can fit into the BERT. In the later case, the relevance labels are either directly transferred from the original query-document pair or learned through some external model. In this paper, we conduct a detailed study of the design decisions about splitting and label transfer on retrieval effectiveness and efficiency. We find that direct transfer of relevance labels from documents to passages introduces label noise that strongly affects retrieval effectiveness for large training datasets. We also find that query processing times are adversely affected by fine-grained splitting schemes. As a remedy, we propose a careful passage level labelling scheme using weak supervision that delivers improved performance (3–14% in terms of nDCG score) over most of the recently proposed models for ad-hoc retrieval while maintaining manageable computational complexity on four diverse document retrieval datasets.

@en

With the ever-increasing popularity and applications of graph neural networks, several proposals have been made to explain and understand the decisions of a graph neural network. Explanations for graph neural networks differ in principle from other input settings. It is important to attribute the decision to input features and other related instances connected by the graph structure. We find that the previous explanation generation approaches that maximize the mutual information between the label distribution produced by the model and the explanation to be restrictive. Specifically, existing approaches do not enforce explanations to be valid, sparse, or robust to input perturbations. In this paper, we lay down some of the fundamental principles that an explanation method for graph neural networks should follow and introduce a metric RDT-Fidelity as a measure of the explanation's effectiveness. We propose a novel approach Zorro based on the principles from rate-distortion theory that uses a simple combinatorial procedure to optimize for RDT-Fidelity. Extensive experiments on real and synthetic datasets reveal that Zorro produces sparser, stable, and more faithful explanations than existing graph neural network explanation approaches.

@en

As algorithms are increasingly augmenting and substituting human decision-making, understanding how the introduction of computational agents changes the fundamentals of human behavior becomes vital. This pertains to not only users, but also those parties who face the consequences of an algorithmic decision. In a controlled experiment with 480 participants, we exploit an extended version of two-player ultimatum bargaining where responders choose to bargain with either another human, another human with an AI decision aid or an autonomous AI-system acting on behalf of a passive human proposer. Our results show strong responder preferences against the algorithm, as most responders opt for a human opponent and demand higher compensation to reach a contract with autonomous agents. To map these preferences to economic expectations, we elicit incentivized subject beliefs about their opponent's behavior. The majority of responders maximize their expected value when this is line with approaching the human proposer. In contrast, responders predicting income maximization for the autonomous AI-system overwhelmingly override economic self-interest to avoid the algorithm.

@en

Quality control is essential for creating extractive question answering (EQA) datasets via crowdsourcing. Aggregation across answers, i.e. word spans within passages annotated, by different crowd workers is one major focus for ensuring its quality. However, crowd workers cannot reach a consensus on a considerable portion of questions. We introduce a simple yet effective answer aggregation method that takes into account the relations among the answer, question, and context passage. We evaluate answer quality from both the view of question answering model to determine how confident the QA model is about each answer and the view of the answer verification model to determine whether the answer is correct. Then we compute aggregation scores with each answer’s quality and its contextual embedding produced by pre-trained language models. The experiments on a large real crowdsourced EQA dataset show that our framework outperforms baselines by around 16% on precision and effectively conduct answer aggregation for extractive QA task.@en

Contextual ranking models have delivered impressive performance improvements over classical models in the document ranking task. However, these highly over-parameterized models tend to be data-hungry and require large amounts of data even for fine tuning. This paper proposes a simple yet effective method to improve ranking performance on smaller datasets using supervised contrastive learning for the document ranking problem. We perform data augmentation by creating training data using parts of the relevant documents in the query-document pairs. We then use a supervised contrastive learning objective to learn an effective ranking model from the augmented dataset. Our experiments on subsets of the TREC-DL dataset show that, although data augmentation leads to an increasing the training data sizes, it does not necessarily improve the performance using existing pointwise or pairwise training objectives. However, our proposed supervised contrastive loss objective leads to performance improvements over the standard non-augmented setting showcasing the utility of data augmentation using contrastive losses. Finally, we show the real benefit of using supervised contrastive learning objectives by showing marked improvements in smaller ranking datasets relating to news (Robust04), finance (FiQA), and scientific fact checking (SciFact).

@en

We introduce SparCAssist, a general-purpose risk assessment tool for the machine learning models trained for language tasks. It evaluates models' risk by inspecting their behavior on counterfactuals, namely out-of-distribution instances generated based on the given data instance. The counterfactuals are generated by replacing tokens in rational subsequences identified by ExPred, while the replacements are retrieved using HotFlip or the Masked-Language-Model-based algorithms. The main purpose of our system is to help the human annotators to assess the model's risk on deployment. The counterfactual instances generated during the assessment are the by-product and can be used to train more robust NLP models in the future.

@en