Simon Razniewski

Tuan-Phong Nguyen, Simon Razniewski, Aparna Varde et al.

Structured knowledge is important for many AI applications. Commonsense knowledge, which is crucial for robust human-centric AI, is covered by a small number of structured knowledge projects. However, they lack knowledge about human traits and behaviors conditioned on socio-cultural contexts, which is crucial for situative AI. This paper presents CANDLE, an end-to-end methodology for extracting high-quality cultural commonsense knowledge (CCSK) at scale. CANDLE extracts CCSK assertions from a huge web corpus and organizes them into coherent clusters, for 3 domains of subjects (geography, religion, occupation) and several cultural facets (food, drinks, clothing, traditions, rituals, behaviors). CANDLE includes judicious techniques for classification-based filtering and scoring of interestingness. Experimental evaluations show the superiority of the CANDLE CCSK collection over prior works, and an extrinsic use case demonstrates the benefits of CCSK for the GPT-3 language model. Code and data can be accessed at https://candle.mpi-inf.mpg.de/.

Jeff Z. Pan, Simon Razniewski, Jan-Christoph Kalo et al.

Large Language Models (LLMs) have taken Knowledge Representation -- and the world -- by storm. This inflection point marks a shift from explicit knowledge representation to a renewed focus on the hybrid representation of both explicit knowledge and parametric knowledge. In this position paper, we will discuss some of the common debate points within the community on LLMs (parametric knowledge) and Knowledge Graphs (explicit knowledge) and speculate on opportunities and visions that the renewed focus brings, as well as related research topics and challenges.

Blerta Veseli, Sneha Singhania, Simon Razniewski et al.

Structured knowledge bases (KBs) are a foundation of many intelligent applications, yet are notoriously incomplete. Language models (LMs) have recently been proposed for unsupervised knowledge base completion (KBC), yet, despite encouraging initial results, questions regarding their suitability remain open. Existing evaluations often fall short because they only evaluate on popular subjects, or sample already existing facts from KBs. In this work, we introduce a novel, more challenging benchmark dataset, and a methodology tailored for a realistic assessment of the KBC potential of LMs. For automated assessment, we curate a dataset called WD-KNOWN, which provides an unbiased random sample of Wikidata, containing over 3.9 million facts. In a second step, we perform a human evaluation on predictions that are not yet in the KB, as only this provides real insights into the added value over existing KBs. Our key finding is that biases in dataset conception of previous benchmarks lead to a systematic overestimate of LM performance for KBC. However, our results also reveal strong areas of LMs. We could, for example, perform a significant completion of Wikidata on the relations nativeLanguage, by a factor of ~21 (from 260k to 5.8M) at 82% precision, usedLanguage, by a factor of ~2.1 (from 2.1M to 6.6M) at 82% precision, and citizenOf by a factor of ~0.3 (from 4.2M to 5.3M) at 90% precision. Moreover, we find that LMs possess surprisingly strong generalization capabilities: even on relations where most facts were not directly observed in LM training, prediction quality can be high.

Hiba Arnaout, Simon Razniewski, Gerhard Weikum et al.

Commonsense knowledge about everyday concepts is an important asset for AI applications, such as question answering and chatbots. Recently, we have seen an increasing interest in the construction of structured commonsense knowledge bases (CSKBs). An important part of human commonsense is about properties that do not apply to concepts, yet existing CSKBs only store positive statements. Moreover, since CSKBs operate under the open-world assumption, absent statements are considered to have unknown truth rather than being invalid. This paper presents the UNCOMMONSENSE framework for materializing informative negative commonsense statements. Given a target concept, comparable concepts are identified in the CSKB, for which a local closed-world assumption is postulated. This way, positive statements about comparable concepts that are absent for the target concept become seeds for negative statement candidates. The large set of candidates is then scrutinized, pruned and ranked by informativeness. Intrinsic and extrinsic evaluations show that our method significantly outperforms the state-of-the-art. A large dataset of informative negations is released as a resource for future research.

Blerta Veseli, Simon Razniewski, Jan-Christoph Kalo et al.

Structured knowledge bases (KBs) are an asset for search engines and other applications, but are inevitably incomplete. Language models (LMs) have been proposed for unsupervised knowledge base completion (KBC), yet, their ability to do this at scale and with high accuracy remains an open question. Prior experimental studies mostly fall short because they only evaluate on popular subjects, or sample already existing facts from KBs. In this work, we perform a careful evaluation of GPT's potential to complete the largest public KB: Wikidata. We find that, despite their size and capabilities, models like GPT-3, ChatGPT and GPT-4 do not achieve fully convincing results on this task. Nonetheless, they provide solid improvements over earlier approaches with smaller LMs. In particular, we show that, with proper thresholding, GPT-3 enables to extend Wikidata by 27M facts at 90% precision.

Lihu Chen, Simon Razniewski, Gerhard Weikum

Despite their impressive scale, knowledge bases (KBs), such as Wikidata, still contain significant gaps. Language models (LMs) have been proposed as a source for filling these gaps. However, prior works have focused on prominent entities with rich coverage by LMs, neglecting the crucial case of long-tail entities. In this paper, we present a novel method for LM-based-KB completion that is specifically geared for facts about long-tail entities. The method leverages two different LMs in two stages: for candidate retrieval and for candidate verification and disambiguation. To evaluate our method and various baselines, we introduce a novel dataset, called MALT, rooted in Wikidata. Our method outperforms all baselines in F1, with major gains especially in recall.

Sneha Singhania, Simon Razniewski, Gerhard Weikum

The widespread usage of latent language representations via pre-trained language models (LMs) suggests that they are a promising source of structured knowledge. However, existing methods focus only on a single object per subject-relation pair, even though often multiple objects are correct. To overcome this limitation, we analyze these representations for their potential to yield materialized multi-object relational knowledge. We formulate the problem as a rank-then-select task. For ranking candidate objects, we evaluate existing prompting techniques and propose new ones incorporating domain knowledge. Among the selection methods, we find that choosing objects with a likelihood above a learned relation-specific threshold gives a 49.5% F1 score. Our results highlight the difficulty of employing LMs for the multi-valued slot-filling task and pave the way for further research on extracting relational knowledge from latent language representations.

Shrestha Ghosh, Simon Razniewski, Gerhard Weikum

Questions on class cardinality comparisons are quite tricky to answer and come with its own challenges. They require some kind of reasoning since web documents and knowledge bases, indispensable sources of information, rarely store direct answers to questions, such as, ``Are there more astronauts or Physics Nobel Laureates?'' We tackle questions on class cardinality comparison by tapping into three sources for absolute cardinalities as well as the cardinalities of orthogonal subgroups of the classes. We propose novel techniques for aggregating signals with partial coverage for more reliable estimates and evaluate them on a dataset of 4005 class pairs, achieving an accuracy of 83.7%.

Luca Giordano, Simon Razniewski

Parametric knowledge in large language models (LLMs) is a cornerstone of their success, yet remains poorly understood. Existing knowledge benchmarks typically rely on predefined questions (e.g., "What is the birth date of M.L. King?"), evaluating only knowledge that benchmark designers explicitly choose to query, a problematic availability bias. In this paper, we introduce open knowledge evaluation, a new paradigm for LLM knowledge benchmarking. Instead of asking narrow questions, it evaluates models on the knowledge they choose to surface in response to open-ended elicitation prompts (e.g., "Tell me everything you know about M.L. King"). This shifts the focus from predefined answer retrieval toward characterizing the knowledge models naturally express. We instantiate this paradigm with BeQu (Beyond Questions), a benchmark of 10,000 entities paired with reference corpora for statement verification. Using BeQu, we evaluate a broad range of language models and analyze the effects of reasoning effort, model scale, prompt format, and knowledge domain. Data and leaderboard are available on this work's GitHub repository and at the benchmark's website.

Julien Romero, Simon Razniewski

Structured knowledge bases (KBs) are the backbone of many know\-ledge-intensive applications, and their automated construction has received considerable attention. In particular, open information extraction (OpenIE) is often used to induce structure from a text. However, although it allows high recall, the extracted knowledge tends to inherit noise from the sources and the OpenIE algorithm. Besides, OpenIE tuples contain an open-ended, non-canonicalized set of relations, making the extracted knowledge's downstream exploitation harder. In this paper, we study the problem of mapping an open KB into the fixed schema of an existing KB, specifically for the case of commonsense knowledge. We propose approaching the problem by generative translation, i.e., by training a language model to generate fixed-schema assertions from open ones. Experiments show that this approach occupies a sweet spot between traditional manual, rule-based, or classification-based canonicalization and purely generative KB construction like COMET. Moreover, it produces higher mapping accuracy than the former while avoiding the association-based noise of the latter.

Suhas Shrinivasan, Simon Razniewski

Knowledge Bases (KBs) provide structured representation of the real-world in the form of extensive collections of facts about real-world entities, their properties and relationships. They are ubiquitous in large-scale intelligent systems that exploit structured information such as in tasks like structured search, question answering and reasoning, and hence their data quality becomes paramount. The inevitability of change in the real-world, brings us to a central property of KBs -- they are highly dynamic in that the information they contain are constantly subject to change. In other words, KBs are unstable. In this paper, we investigate the notion of KB stability, specifically, the problem of KBs changing due to real-world change. Some entity-property-pairs do not undergo change in reality anymore (e.g., Einstein-children or Tesla-founders), while others might well change in the future (e.g., Tesla-board member or Ronaldo-occupation as of 2022). This notion of real-world grounded change is different from other changes that affect the data only, notably correction and delayed insertion, which have received attention in data cleaning, vandalism detection, and completeness estimation already. To analyze KB stability, we proceed in three steps. (1) We present heuristics to delineate changes due to world evolution from delayed completions and corrections, and use these to study the real-world evolution behaviour of diverse Wikidata domains, finding a high skew in terms of properties. (2) We evaluate heuristics to identify entities and properties likely to not change due to real-world change, and filter inherently stable entities and properties. (3) We evaluate the possibility of predicting stability post-hoc, specifically predicting change in a property of an entity, finding that this is possible with up to 83% F1 score, on a balanced binary stability prediction task.

Julien Romero, Simon Razniewski

Compiling comprehensive repositories of commonsense knowledge is a long-standing problem in AI. Many concerns revolve around the issue of reporting bias, i.e., that frequency in text sources is not a good proxy for relevance or truth. This paper explores whether children's texts hold the key to commonsense knowledge compilation, based on the hypothesis that such content makes fewer assumptions on the reader's knowledge, and therefore spells out commonsense more explicitly. An analysis with several corpora shows that children's texts indeed contain much more, and more typical commonsense assertions. Moreover, experiments show that this advantage can be leveraged in popular language-model-based commonsense knowledge extraction settings, where task-unspecific fine-tuning on small amounts of children texts (childBERT) already yields significant improvements. This provides a refreshing perspective different from the common trend of deriving progress from ever larger models and corpora.

Muhammed Saeed, Simon Razniewski

Benchmarks such as MMLU suggest flagship language models approach factuality saturation, with scores above 90\%. We show this picture is incomplete. \emph{LLMpedia} generates encyclopedic articles entirely from parametric memory, producing ${\sim}$1M articles across three model families without retrieval. For gpt-5-mini, the verifiable true rate on Wikipedia-covered subjects is only 74.7\% -- more than 15 percentage points below the benchmark-based picture, consistent with the availability bias of fixed-question evaluation. Beyond Wikipedia, frontier subjects verifiable only through curated web evidence fall further to 63.2\% true rate. Wikipedia covers just 61\% of surfaced subjects, and three model families overlap by only 7.3\% in subject choice. In a capture-trap benchmark inspired by prior analysis of Grokipedia, LLMpedia achieves substantially higher factuality at roughly half the textual similarity to Wikipedia. Unlike Grokipedia, every prompt, artifact, and evaluation verdict is publicly released, making LLMpedia the first fully open parametric encyclopedia -- bridging factuality evaluation and knowledge materialization. All data, code, and a browsable interface are at https://llmpedia.net.

Valentin Knappich, Simon Razniewski, Annemarie Friedrich

Automatic simplification can help laypeople to comprehend complex scientific text. Language models are frequently applied to this task by translating from complex to simple language. In this paper, we describe our system based on Llama 2, which ranked first in the PLABA shared task addressing the simplification of biomedical text. We find that the large portion of shared tokens between input and output leads to weak training signals and conservatively editing models. To mitigate these issues, we propose sentence-level and token-level loss weights. They give higher weight to modified tokens, indicated by edit distance and edit operations, respectively. We conduct an empirical evaluation on the PLABA dataset and find that both approaches lead to simplifications closer to those created by human annotators (+1.8% / +3.5% SARI), simpler language (-1 / -1.1 FKGL) and more edits (1.6x / 1.8x edit distance) compared to the same model fine-tuned with standard cross entropy. We furthermore show that the hyperparameter $λ$ in token-level loss weights can be used to control the edit distance and the simplicity level (FKGL).

Timo Pierre Schrader, Lukas Lange, Tobias Kaminski et al.

The rise of large language models (LLMs) has sparked interest in coding assistants. While general-purpose programming languages are well supported, generating code for domain-specific languages remains a challenging problem for LLMs. In this paper, we focus on the LLM-based generation of code for Answer Set Programming (ASP), a particularly effective approach for finding solutions to combinatorial search problems. The effectiveness of LLMs in ASP code generation is currently hindered by the limited number of examples seen during their initial pre-training phase. In this paper, we introduce a novel ASP-solver-in-the-loop approach for solver-guided instruction-tuning of LLMs to addressing the highly complex semantic parsing task inherent in ASP code generation. Our method only requires problem specifications in natural language and their solutions. Specifically, we sample ASP statements for program continuations from LLMs for unriddling logic puzzles. Leveraging the special property of declarative ASP programming that partial encodings increasingly narrow down the solution space, we categorize them into chosen and rejected instances based on solver feedback. We then apply supervised fine-tuning to train LLMs on the curated data and further improve robustness using a solver-guided search that includes best-of-N sampling. Our experiments demonstrate consistent improvements in two distinct prompting settings on two datasets.

Valentin Knappich, Anna Hätty, Simon Razniewski et al.

Novelty assessment is a critical yet complex task in the examination process for patent acceptance, requiring examiners to determine whether an invention is disclosed in a prior art document. The process involves intricate matching between specific features of a patent claim and passages in the prior art. While prior work has approached novelty prediction primarily as a binary classification task at the claim level, we argue that this formulation is susceptible to spurious correlations and lacks the granularity required for practical application. In this work, we introduce FiNE-Patents (Fine-grained Novelty Examination of Patents), a novel dataset comprising 3,658 first patent claims annotated with fine-grained, feature-level prior art references extracted from European Search Opinion (ESOP) documents. We propose shifting the evaluation paradigm from simple binary classification to a joint retrieval and abstract reasoning task at the feature level, requiring models to identify specific passages from a prior art document that disclose individual claim features, and to identify which features of a claim make it novel. We implement and evaluate LLM-based workflows that decompose claims into features, analyze each feature against prior art, and finally derive a claim-level novelty prediction. Our experiments demonstrate that these workflows outperform embedding-based baselines on passage retrieval and novel feature identification. Furthermore, we show that unlike trained classifiers, LLMs are robust against spurious correlations present in the claim-level novelty classification task. We release the dataset and code to foster further research into transparent and granular patent analysis.

Tuan-Phong Nguyen, Simon Razniewski, Gerhard Weikum

Despite recent progress, large language models (LLMs) still face the challenge of appropriately reacting to the intricacies of social and cultural conventions. This paper presents MANGO, a methodology for distilling high-accuracy, high-recall assertions of cultural knowledge. We judiciously and iteratively prompt LLMs for this purpose from two entry points, concepts and cultures. Outputs are consolidated via clustering and generative summarization. Running the MANGO method with GPT-3.5 as underlying LLM yields 167K high-accuracy assertions for 30K concepts and 11K cultures, surpassing prior resources by a large margin in quality and size. In an extrinsic evaluation for intercultural dialogues, we explore augmenting dialogue systems with cultural knowledge assertions. Notably, despite LLMs inherently possessing cultural knowledge, we find that adding knowledge from MANGO improves the overall quality, specificity, and cultural sensitivity of dialogue responses, as judged by human annotators. Data and code are available for download.

Yujia Hu, Tuan-Phong Nguyen, Shrestha Ghosh et al.

Large language models (LLMs) have majorly advanced NLP and AI, and next to their ability to perform a wide range of procedural tasks, a major success factor is their internalized factual knowledge. Since Petroni et al. (2019), analyzing this knowledge has gained attention. However, most approaches investigate one question at a time via modest-sized pre-defined samples, introducing an ``availability bias'' (Tversky&Kahnemann, 1973) that prevents the analysis of knowledge (or beliefs) of LLMs beyond the experimenter's predisposition. To address this challenge, we propose a novel methodology to comprehensively materialize an LLM's factual knowledge through recursive querying and result consolidation. Our approach is a milestone for LLM research, for the first time providing constructive insights into the scope and structure of LLM knowledge (or beliefs). As a prototype, we build GPTKB, a knowledge base (KB) comprising 101 million relational triples for over 2.9 million entities from GPT-4o-mini. We use GPTKB to exemplarily analyze GPT-4o-mini's factual knowledge in terms of scale, accuracy, bias, cutoff and consistency, at the same time. GPTKB is accessible at https://gptkb.org

Timo Pierre Schrader, Lukas Lange, Simon Razniewski et al.

Reasoning is key to many decision making processes. It requires consolidating a set of rule-like premises that are often associated with degrees of uncertainty and observations to draw conclusions. In this work, we address both the case where premises are specified as numeric probabilistic rules and situations in which humans state their estimates using words expressing degrees of certainty. Existing probabilistic reasoning datasets simplify the task, e.g., by requiring the model to only rank textual alternatives, by including only binary random variables, or by making use of a limited set of templates that result in less varied text. In this work, we present QUITE, a question answering dataset of real-world Bayesian reasoning scenarios with categorical random variables and complex relationships. QUITE provides high-quality natural language verbalizations of premises together with evidence statements and expects the answer to a question in the form of an estimated probability. We conduct an extensive set of experiments, finding that logic-based models outperform out-of-the-box large language models on all reasoning types (causal, evidential, and explaining-away). Our results provide evidence that neuro-symbolic models are a promising direction for improving complex reasoning. We release QUITE and code for training and experiments on Github.

Yujia Hu, Tuan-Phong Nguyen, Shrestha Ghosh et al.

Language models are powerful tools, yet their factual knowledge is still poorly understood, and inaccessible to ad-hoc browsing and scalable statistical analysis. This demonstration introduces GPTKB v1.5, a densely interlinked 100-million-triple knowledge base (KB) built for $14,000 from GPT-4.1, using the GPTKB methodology for massive-recursive LLM knowledge materialization (Hu et al., ACL 2025). The demonstration experience focuses on three use cases: (1) link-traversal-based LLM knowledge exploration, (2) SPARQL-based structured LLM knowledge querying, (3) comparative exploration of the strengths and weaknesses of LLM knowledge. Massive-recursive LLM knowledge materialization is a groundbreaking opportunity both for the research area of systematic analysis of LLM knowledge, as well as for automated KB construction. The GPTKB demonstrator is accessible at https://gptkb.org.

Sneha Singhania, Simon Razniewski, Gerhard Weikum

Methods for relation extraction from text mostly focus on high precision, at the cost of limited recall. High recall is crucial, though, to populate long lists of object entities that stand in a specific relation with a given subject. Cues for relevant objects can be spread across many passages in long texts. This poses the challenge of extracting long lists from long texts. We present the L3X method which tackles the problem in two stages: (1) recall-oriented generation using a large language model (LLM) with judicious techniques for retrieval augmentation, and (2) precision-oriented scrutinization to validate or prune candidates. Our L3X method outperforms LLM-only generations by a substantial margin.

Timo Pierre Schrader, Simon Razniewski, Lukas Lange et al.

Understanding causality is a core aspect of intelligence. The Event Causality Identification with Causal News Corpus Shared Task addresses two aspects of this challenge: Subtask 1 aims at detecting causal relationships in texts, and Subtask 2 requires identifying signal words and the spans that refer to the cause or effect, respectively. Our system, which is based on pre-trained transformers, stacked sequence tagging, and synthetic data augmentation, ranks third in Subtask 1 and wins Subtask 2 with an F1 score of 72.8, corresponding to a margin of 13 pp. to the second-best system.

Shrestha Ghosh, Luca Giordano, Yujia Hu et al.

LLMs are remarkable artifacts that have revolutionized a range of NLP and AI tasks. A significant contributor is their factual knowledge, which, to date, remains poorly understood, and is usually analyzed from biased samples. In this paper, we take a deep tour into the factual knowledge (or beliefs) of a frontier LLM, based on GPTKB v1.5 (Hu et al., 2025a), a recursively elicited set of 100 million beliefs of one of the strongest currently available frontier LLMs, GPT-4.1. We find that the models' factual knowledge differs quite significantly from established knowledge bases, and that its accuracy is significantly lower than indicated by previous benchmarks. We also find that inconsistency, ambiguity and hallucinations are major issues, shedding light on future research opportunities concerning factual LLM knowledge.

Luca Giordano, Simon Razniewski

Large Language Models (LLMs) encode substantial factual knowledge, yet measuring and systematizing this knowledge remains challenging. Converting it into structured format, for example through recursive extraction approaches such as the GPTKB methodology (Hu et al., 2025b), is still underexplored. Key open questions include whether such extraction can terminate, whether its outputs are reproducible, and how robust they are to variations. We systematically study LLM knowledge materialization using miniGPTKBs (domain-specific, tractable subcrawls), analyzing termination, reproducibility, and robustness across three categories of metrics: yield, lexical similarity, and semantic similarity. We experiment with four variations (seed, language, randomness, model) and three illustrative domains (from history, entertainment, and finance). Our findings show (i) high termination rates, though model-dependent; (ii) mixed reproducibility; and (iii) robustness that varies by perturbation type: high for seeds and temperature, lower for languages and models. These results suggest that LLM knowledge materialization can reliably surface core knowledge, while also revealing important limitations.

Yu Yao, Salil Bhatnagar, Markus Mazzola et al.

Rare, yet critical, scenarios pose a significant challenge in testing and evaluating autonomous driving planners. Relying solely on real-world driving scenes requires collecting massive datasets to capture these scenarios. While automatic generation of traffic scenarios appears promising, data-driven models require extensive training data and often lack fine-grained control over the output. Moreover, generating novel scenarios from scratch can introduce a distributional shift from the original training scenes which undermines the validity of evaluations especially for learning-based planners. To sidestep this, recent work proposes to generate challenging scenarios by augmenting original scenarios from the test set. However, this involves the manual augmentation of scenarios by domain experts. An approach that is unable to meet the demands for scale in the evaluation of self-driving systems. Therefore, this paper introduces a novel LLM-agent based framework for augmenting real-world traffic scenarios using natural language descriptions, addressing the limitations of existing methods. A key innovation is the use of an agentic design, enabling fine-grained control over the output and maintaining high performance even with smaller, cost-effective LLMs. Extensive human expert evaluation demonstrates our framework's ability to accurately adhere to user intent, generating high quality augmented scenarios comparable to those created manually.

Valentin Knappich, Annemarie Friedrich, Anna Hätty et al.

Patent claims define the scope of protection for an invention. If there are ambiguities in a claim, it is rejected by the patent office. In the US, this is referred to as indefiniteness (35 U.S.C § 112(b)) and is among the most frequent reasons for patent application rejection. The development of automatic methods for patent definiteness examination has the potential to make patent drafting and examination more efficient, but no annotated dataset has been published to date. We introduce PEDANTIC (Patent Definiteness Examination Corpus), a novel dataset of 14k US patent claims from patent applications relating to Natural Language Processing (NLP), annotated with reasons for indefiniteness. We construct PEDANTIC using a fully automatic pipeline that retrieves office action documents from the USPTO and uses Large Language Models (LLMs) to extract the reasons for indefiniteness. A human validation study confirms the pipeline's accuracy in generating high-quality annotations. To gain insight beyond binary classification metrics, we implement an LLM-as-Judge evaluation that compares the free-form reasoning of every model-cited reason with every examiner-cited reason. We show that LLM agents based on Qwen 2.5 32B and 72B struggle to outperform logistic regression baselines on definiteness prediction, even though they often correctly identify the underlying reasons. PEDANTIC provides a valuable resource for patent AI researchers, enabling the development of advanced examination models. We will publicly release the dataset and code.

Hiba Arnaout, Simon Razniewski

We examine the ability of large language models (LLMs) to generate salient (interesting) negative statements about real-world entities; an emerging research topic of the last few years. We probe the LLMs using zero- and k-shot unconstrained probes, and compare with traditional methods for negation generation, i.e., pattern-based textual extractions and knowledge-graph-based inferences, as well as crowdsourced gold statements. We measure the correctness and salience of the generated lists about subjects from different domains. Our evaluation shows that guided probes do in fact improve the quality of generated negatives, compared to the zero-shot variant. Nevertheless, using both prompts, LLMs still struggle with the notion of factuality of negatives, frequently generating many ambiguous statements, or statements with negative keywords but a positive meaning.

Simon Razniewski, Hiba Arnaout, Shrestha Ghosh et al.

General-purpose knowledge bases (KBs) are a cornerstone of knowledge-centric AI. Many of them are constructed pragmatically from Web sources, and are thus far from complete. This poses challenges for the consumption as well as the curation of their content. While several surveys target the problem of completing incomplete KBs, the first problem is arguably to know whether and where the KB is incomplete in the first place, and to which degree. In this survey we discuss how knowledge about completeness, recall, and negation in KBs can be expressed, extracted, and inferred. We cover (i) the logical foundations of knowledge representation and querying under partial closed-world semantics; (ii) the estimation of this information via statistical patterns; (iii) the extraction of information about recall from KBs and text; (iv) the identification of interesting negative statements; and (v) relaxed notions of relative recall. This survey is targeted at two types of audiences: (1) practitioners who are interested in tracking KB quality, focusing extraction efforts, and building quality-aware downstream applications; and (2) data management, knowledge base and semantic web researchers who wish to understand the state of the art of knowledge bases beyond the open-world assumption. Consequently, our survey presents both fundamental methodologies and their working, and gives practice-oriented recommendations on how to choose between different approaches for a problem at hand.

Tuan-Phong Nguyen, Simon Razniewski

Starting from the COMET methodology by Bosselut et al. (2019), generating commonsense knowledge directly from pre-trained language models has recently received significant attention. Surprisingly, up to now no materialized resource of commonsense knowledge generated this way is publicly available. This paper fills this gap, and uses the materialized resources to perform a detailed analysis of the potential of this approach in terms of precision and recall. Furthermore, we identify common problem cases, and outline use cases enabled by materialized resources. We posit that the availability of these resources is important for the advancement of the field, as it enables an off-the-shelf-use of the resulting knowledge, as well as further analyses on its strengths and weaknesses.

Tuan-Phong Nguyen, Simon Razniewski, Julien Romero et al.

Commonsense knowledge (CSK) about concepts and their properties is helpful for AI applications. Prior works, such as ConceptNet, have compiled large CSK collections. However, they are restricted in their expressiveness to subject-predicate-object (SPO) triples with simple concepts for S and strings for P and O. This paper presents a method called ASCENT++ to automatically build a large-scale knowledge base (KB) of CSK assertions, with refined expressiveness and both better precision and recall than prior works. ASCENT++ goes beyond SPO triples by capturing composite concepts with subgroups and aspects, and by refining assertions with semantic facets. The latter is essential to express the temporal and spatial validity of assertions and further qualifiers. Furthermore, ASCENT++ combines open information extraction (OpenIE) with judicious cleaning and ranking by typicality and saliency scores. For high coverage, our method taps into the large-scale crawl C4 with broad web contents. The evaluation with human judgments shows the superior quality of the ASCENT++ KB, and an extrinsic evaluation for QA-support tasks underlines the benefits of ASCENT++. A web interface, data, and code can be accessed at https://ascentpp.mpi-inf.mpg.de/.

Sneha Singhania, Simon Razniewski, Gerhard Weikum

This paper presents a new task of predicting the coverage of a text document for relation extraction (RE): does the document contain many relational tuples for a given entity? Coverage predictions are useful in selecting the best documents for knowledge base construction with large input corpora. To study this problem, we present a dataset of 31,366 diverse documents for 520 entities. We analyze the correlation of document coverage with features like length, entity mention frequency, Alexa rank, language complexity and information retrieval scores. Each of these features has only moderate predictive power. We employ methods combining features with statistical models like TF-IDF and language models like BERT. The model combining features and BERT, HERB, achieves an F1 score of up to 46%. We demonstrate the utility of coverage predictions on two use cases: KB construction and claim refutation.

Simon Razniewski, Andrew Yates, Nora Kassner et al.

Pre-trained language models (LMs) have recently gained attention for their potential as an alternative to (or proxy for) explicit knowledge bases (KBs). In this position paper, we examine this hypothesis, identify strengths and limitations of both LMs and KBs, and discuss the complementary nature of the two paradigms. In particular, we offer qualitative arguments that latent LMs are not suitable as a substitute for explicit KBs, but could play a major role for augmenting and curating KBs.

Tuan-Phong Nguyen, Simon Razniewski, Gerhard Weikum

ASCENT is a fully automated methodology for extracting and consolidating commonsense assertions from web contents (Nguyen et al., WWW 2021). It advances traditional triple-based commonsense knowledge representation by capturing semantic facets like locations and purposes, and composite concepts, i.e., subgroups and related aspects of subjects. In this demo, we present a web portal that allows users to understand its construction process, explore its content, and observe its impact in the use case of question answering. The demo website and an introductory video are both available online.

Simon Razniewski

Compiling commonsense knowledge is traditionally an AI topic approached by manual labor. Recent advances in web data processing have enabled automated approaches. In this demonstration we will showcase three systems for automated commonsense knowledge base construction, highlighting each time one aspect of specific interest to the data management community. (i) We use Quasimodo to illustrate knowledge extraction systems engineering, (ii) Dice to illustrate the role that schema constraints play in cleaning fuzzy commonsense knowledge, and (iii) Ascent to illustrate the relevance of conceptual modelling. The demos are available online at https://quasimodo.r2.enst.fr, https://dice.mpi-inf.mpg.de and ascent.mpi-inf.mpg.de.

Tuan-Phong Nguyen, Simon Razniewski, Gerhard Weikum

Commonsense knowledge (CSK) about concepts and their properties is useful for AI applications such as robust chatbots. Prior works like ConceptNet, TupleKB and others compiled large CSK collections, but are restricted in their expressiveness to subject-predicate-object (SPO) triples with simple concepts for S and monolithic strings for P and O. Also, these projects have either prioritized precision or recall, but hardly reconcile these complementary goals. This paper presents a methodology, called Ascent, to automatically build a large-scale knowledge base (KB) of CSK assertions, with advanced expressiveness and both better precision and recall than prior works. Ascent goes beyond triples by capturing composite concepts with subgroups and aspects, and by refining assertions with semantic facets. The latter are important to express temporal and spatial validity of assertions and further qualifiers. Ascent combines open information extraction with judicious cleaning using language models. Intrinsic evaluation shows the superior size and quality of the Ascent KB, and an extrinsic evaluation for QA-support tasks underlines the benefits of Ascent. A web interface, data and code can be found at https://ascent.mpi-inf.mpg.de/.

Gerhard Weikum, Luna Dong, Simon Razniewski et al.

Equipping machines with comprehensive knowledge of the world's entities and their relationships has been a long-standing goal of AI. Over the last decade, large-scale knowledge bases, also known as knowledge graphs, have been automatically constructed from web contents and text sources, and have become a key asset for search engines. This machine knowledge can be harnessed to semantically interpret textual phrases in news, social media and web tables, and contributes to question answering, natural language processing and data analytics. This article surveys fundamental concepts and practical methods for creating and curating large knowledge bases. It covers models and methods for discovering and canonicalizing entities and their semantic types and organizing them into clean taxonomies. On top of this, the article discusses the automatic extraction of entity-centric properties. To support the long-term life-cycle and the quality assurance of machine knowledge, the article presents methods for constructing open schemas and for knowledge curation. Case studies on academic projects and industrial knowledge graphs complement the survey of concepts and methods.

Jesse Josua Benjamin, Claudia Müller-Birn, Simon Razniewski

The global infrastructure of the Web, designed as an open and transparent system, has a significant impact on our society. However, algorithmic systems of corporate entities that neglect those principles increasingly populated the Web. Typical representatives of these algorithmic systems are recommender systems that influence our society both on a scale of global politics and during mundane shopping decisions. Recently, such recommender systems have come under critique for how they may strengthen existing or even generate new kinds of biases. To this end, designers and engineers are increasingly urged to make the functioning and purpose of recommender systems more transparent. Our research relates to the discourse of algorithm awareness, that reconsiders the role of algorithm visibility in interface design. We conducted online experiments with 105 participants using MTurk for the recommender system Recoin, a gadget for Wikidata. In these experiments, we presented users with one of a set of three different designs of Recoin's user interface, each of them exhibiting a varying degree of explainability and interactivity. Our findings include a positive correlation between comprehension of and trust in an algorithmic system in our interactive redesign. However, our results are not conclusive yet, and suggest that the measures of comprehension, fairness, accuracy and trust are not yet exhaustive for the empirical study of algorithm awareness. Our qualitative insights provide a first indication for further measures. Our study participants, for example, were less concerned with the details of understanding an algorithmic calculation than with who or what is judging the result of the algorithm.

Diego Calvanese, Julien Corman, Davide Lanti et al.

Counting answers to a query is an operation supported by virtually all database management systems. In this paper we focus on counting answers over a Knowledge Base (KB), which may be viewed as a database enriched with background knowledge about the domain under consideration. In particular, we place our work in the context of Ontology-Mediated Query Answering/Ontology-based Data Access (OMQA/OBDA), where the language used for the ontology is a member of the DL-Lite family and the data is a (usually virtual) set of assertions. We study the data complexity of query answering, for different members of the DL-Lite family that include number restrictions, and for variants of conjunctive queries with counting that differ with respect to their shape (connected, branching, rooted). We improve upon existing results by providing a PTIME and coNP lower bounds, and upper bounds in PTIME and LOGSPACE. For the latter case, we define a novel query rewriting technique into first-order logic with counting.

Shrestha Ghosh, Simon Razniewski, Gerhard Weikum

Predicate constraints of general-purpose knowledge bases (KBs) like Wikidata, DBpedia and Freebase are often limited to subproperty, domain and range constraints. In this demo we showcase CounQER, a system that illustrates the alignment of counting predicates, like staffSize, and enumerating predicates, like workInstitution^{-1} . In the demonstration session, attendees can inspect these alignments, and will learn about the importance of these alignments for KB question answering and curation. CounQER is available at https://counqer.mpi-inf.mpg.de/spo.

Shrestha Ghosh, Simon Razniewski, Gerhard Weikum

Knowledge Bases (KBs) contain a wealth of structured information about entities and predicates. This paper focuses on set-valued predicates, i.e., the relationship between an entity and a set of entities. In KBs, this information is often represented in two formats: (i) via counting predicates such as numberOfChildren and staffSize, that store aggregated integers, and (ii) via enumerating predicates such as parentOf and worksFor, that store individual set memberships. Both formats are typically complementary: unlike enumerating predicates, counting predicates do not give away individuals, but are more likely informative towards the true set size, thus this coexistence could enable interesting applications in question answering and KB curation. In this paper we aim at uncovering this hidden knowledge. We proceed in two steps. (i) We identify set-valued predicates from a given KB predicates via statistical and embedding-based features. (ii) We link counting predicates and enumerating predicates by a combination of co-occurrence, correlation and textual relatedness metrics. We analyze the prevalence of count information in four prominent knowledge bases, and show that our linking method achieves up to 0.55 F1 score in set predicate identification versus 0.40 F1 score of a random selection, and normalized discounted gains of up to 0.84 at position 1 and 0.75 at position 3 in relevant predicate alignments. Our predicate alignments are showcased in a demonstration system available at https://counqer.mpi-inf.mpg.de/spo.

Hiba Arnaout, Simon Razniewski, Gerhard Weikum et al.

Knowledge bases (KBs) about notable entities and their properties are an important asset in applications such as search, question answering and dialogue. All popular KBs capture virtually only positive statements, and abstain from taking any stance on statements not stored in the KB. This paper makes the case for explicitly stating salient statements that do not hold. Negative statements are useful to overcome limitations of question answering systems that are mainly geared for positive questions; they can also contribute to informative summaries of entities. Due to the abundance of such invalid statements, any effort to compile them needs to address ranking by saliency. We present a statisticalinference method for compiling and ranking negative statements, based on expectations from positive statements of related entities in peer groups. Experimental results, with a variety of datasets, show that the method can effectively discover notable negative statements, and extrinsic studies underline their usefulness for entity summarization. Datasets and code are released as resources for further research.

Yohan Chalier, Simon Razniewski, Gerhard Weikum

Commonsense knowledge (CSK) supports a variety of AI applications, from visual understanding to chatbots. Prior works on acquiring CSK, such as ConceptNet, have compiled statements that associate concepts, like everyday objects or activities, with properties that hold for most or some instances of the concept. Each concept is treated in isolation from other concepts, and the only quantitative measure (or ranking) of properties is a confidence score that the statement is valid. This paper aims to overcome these limitations by introducing a multi-faceted model of CSK statements and methods for joint reasoning over sets of inter-related statements. Our model captures four different dimensions of CSK statements: plausibility, typicality, remarkability and salience, with scoring and ranking along each dimension. For example, hyenas drinking water is typical but not salient, whereas hyenas eating carcasses is salient. For reasoning and ranking, we develop a method with soft constraints, to couple the inference over concepts that are related in in a taxonomic hierarchy. The reasoning is cast into an integer linear programming (ILP), and we leverage the theory of reduction costs of a relaxed LP to compute informative rankings. This methodology is applied to several large CSK collections. Our evaluation shows that we can consolidate these inputs into much cleaner and more expressive knowledge. Results are available at https://dice.mpi-inf.mpg.de.

Sreyasi Nag Chowdhury, Simon Razniewski, Gerhard Weikum

Multimodal contents have become commonplace on the Internet today, manifested as news articles, social media posts, and personal or business blog posts. Among the various kinds of media (images, videos, graphics, icons, audio) used in such multimodal stories, images are the most popular. The selection of images from a collection - either author's personal photo album, or web repositories - and their meticulous placement within a text, builds a succinct multimodal commentary for digital consumption. In this paper we present a system that automates the process of selecting relevant images for a story and placing them at contextual paragraphs within the story for a multimodal narration. We leverage automatic object recognition, user-provided tags, and commonsense knowledge, and use an unsupervised combinatorial optimization to solve the selection and placement problems seamlessly as a single unit.

Julien Romero, Simon Razniewski, Koninika Pal et al.

Commonsense knowledge about object properties, human behavior and general concepts is crucial for robust AI applications. However, automatic acquisition of this knowledge is challenging because of sparseness and bias in online sources. This paper presents Quasimodo, a methodology and tool suite for distilling commonsense properties from non-standard web sources. We devise novel ways of tapping into search-engine query logs and QA forums, and combining the resulting candidate assertions with statistical cues from encyclopedias, books and image tags in a corroboration step. Unlike prior work on commonsense knowledge bases, Quasimodo focuses on salient properties that are typically associated with certain objects or concepts. Extensive evaluations, including extrinsic use-case studies, show that Quasimodo provides better coverage than state-of-the-art baselines with comparable quality.

Cuong Xuan Chu, Simon Razniewski, Gerhard Weikum

Taxonomies are important building blocks of structured knowledge bases, and their construction from text sources and Wikipedia has received much attention. In this paper we focus on the construction of taxonomies for fictional domains, using noisy category systems from fan wikis or text extraction as input. Such fictional domains are archetypes of entity universes that are poorly covered by Wikipedia, such as also enterprise-specific knowledge bases or highly specialized verticals. Our fiction-targeted approach, called TiFi, consists of three phases: (i) category cleaning, by identifying candidate categories that truly represent classes in the domain of interest, (ii) edge cleaning, by selecting subcategory relationships that correspond to class subsumption, and (iii) top-level construction, by mapping classes onto a subset of high-level WordNet categories. A comprehensive evaluation shows that TiFi is able to construct taxonomies for a diverse range of fictional domains such as Lord of the Rings, The Simpsons or Greek Mythology with very high precision and that it outperforms state-of-the-art baselines for taxonomy induction by a substantial margin.

Paramita Mirza, Simon Razniewski, Fariz Darari et al.

Information extraction traditionally focuses on extracting relations between identifiable entities, such as <Monterey, locatedIn, California>. Yet, texts often also contain Counting information, stating that a subject is in a specific relation with a number of objects, without mentioning the objects themselves, for example, "California is divided into 58 counties". Such counting quantifiers can help in a variety of tasks such as query answering or knowledge base curation, but are neglected by prior work. This paper develops the first full-fledged system for extracting counting information from text, called CINEX. We employ distant supervision using fact counts from a knowledge base as training seeds, and develop novel techniques for dealing with several challenges: (i) non-maximal training seeds due to the incompleteness of knowledge bases, (ii) sparse and skewed observations in text sources, and (iii) high diversity of linguistic patterns. Experiments with five human-evaluated relations show that CINEX can achieve 60% average precision for extracting counting information. In a large-scale experiment, we demonstrate the potential for knowledge base enrichment by applying CINEX to 2,474 frequent relations in Wikidata. CINEX can assert the existence of 2.5M facts for 110 distinct relations, which is 28% more than the existing Wikidata facts for these relations.

Simon Razniewski, Vevake Balaraman, Werner Nutt

In knowledge bases such as Wikidata, it is possible to assert a large set of properties for entities, ranging from generic ones such as name and place of birth to highly profession-specific or background-specific ones such as doctoral advisor or medical condition. Determining a preference or ranking in this large set is a challenge in tasks such as prioritisation of edits or natural-language generation. Most previous approaches to ranking knowledge base properties are purely data-driven, that is, as we show, mistake frequency for interestingness. In this work, we have developed a human-annotated dataset of 350 preference judgments among pairs of knowledge base properties for fixed entities. From this set, we isolate a subset of pairs for which humans show a high level of agreement (87.5% on average). We show, however, that baseline and state-of-the-art techniques achieve only 61.3% precision in predicting human preferences for this subset. We then analyze what contributes to one property being rated as more important than another one, and identify that at least three factors play a role, namely (i) general frequency, (ii) applicability to similar entities and (iii) semantic similarity between property and entity. We experimentally analyze the contribution of each factor and show that a combination of techniques addressing all the three factors achieves 74% precision on the task. The dataset is available at www.kaggle.com/srazniewski/wikidatapropertyranking.

Paramita Mirza, Simon Razniewski, Fariz Darari et al.

Information extraction (IE) from text has largely focused on relations between individual entities, such as who has won which award. However, some facts are never fully mentioned, and no IE method has perfect recall. Thus, it is beneficial to also tap contents about the cardinalities of these relations, for example, how many awards someone has won. We introduce this novel problem of extracting cardinalities and discusses the specific challenges that set it apart from standard IE. We present a distant supervision method using conditional random fields. A preliminary evaluation results in precision between 3% and 55%, depending on the difficulty of relations.