Temporal Entities Research Articles

EHPR: Learning evolutionary hierarchy perception representation based on quaternion for temporal knowledge graph completion

Research on temporal knowledge graphs garners attention due to the intricate connection between facts and dynamic temporal factors. However, existing research uses timestamp as auxiliary data for representation learning and directly integrate it into facts, resulting in the inability to capture the intrinsic connections between relations under time evolution. To handle these challenges, we propose the Evolutionary Hierarchy Perception Representation (EHPR), which first leverages the Hamilton product to perform rotational transformations on relation and entity over time, aiming to learn temporal relation and temporal entity with close interactions with time information. Later, EHPR is divided into two modules: (a) Rotating the head entity towards the tail entity using temporal relation through Hamilton product to model complex patterns with quaternion rotation capabilities. (b) Adopting an evolutionary hierarchical factor to capture the differences in modulus distribution between the temporal head entity and the temporal tail entity, aiming to manage the evolutionary hierarchical information between different temporal entities. In this way, EHPR not only utilizes the rich quaternion rotation capabilities to model various relation patterns but also further enables modeling of evolutionary hierarchical patterns through evolutionary hierarchy factors. Experiments show that EHPR achieves remarkable performance on six mature benchmarks compared to state-of-the-art models. Furthermore, we successfully transferred the core idea of EHPR into complex embeddings, showcasing the framework's adaptability. Compared to complex embedding models, EHPR also demonstrates stronger expressive abilities with the Hamilton operator, surpassing the performance of complex Hermitian operator.

Leveraging social computing for epidemic surveillance: A case study

Social media platforms have become a popular source of information for real-time monitoring of events and user behavior. In particular, Twitter provides invaluable information related to diseases and public health to build real-time disease surveillance systems. Effective use of such social media platforms for public health surveillance requires data-driven AI models which are hindered by the difficult, expensive, and time-consuming task of collecting high-quality and large-scale datasets. In this paper, we build and analyze the Epidemic TweetBank (EpiBank) dataset containing 271 million English tweets related to six epidemic-prone diseases COVID19, Flu, Hepatitis, Dengue, Malaria, and HIV/AIDs. For this purpose, we develop a tool of ESS-T (Epidemic Surveillance Study via Twitter) which collects tweets according to provided input parameters and keywords. Also, our tool assigns location to tweets with 95% accuracy value and performs analysis of collected tweets focusing on temporal distribution, spatial patterns, users, entities, sentiment, and misinformation. Leveraging ESS-T, we build two geo-tagged datasets of EpiBank-global and EpiBank-Pak containing 86 million tweets from 190 countries and 2.6 million tweets from Pakistan, respectively. Our spatial analysis of EpiBank-global for COVID19, Malaria, and Dengue indicates that our framework correctly identifies high-risk epidemic-prone countries according to World Health Organization (WHO) statistics.

TELS: Learning time-evolving information and latent semantics using dual quaternion for temporal knowledge graph completion

In temporal knowledge graphs (TKGs), the status of facts is intricately tied to the dynamic and precise nature of temporal factors. Existing research merely treats time as supplementary information, without considering the latent semantic changes caused by the positional changes of entities within specific relations in a temporal context. Furthermore, due to the coarse granularity of timestamps in existing TKGs, the number of multiple relations pattern among entities significantly increases, limiting model performance. This paper proposes a Time-Evolving Information and Latent Semantics model (TELS), which represents facts as dual quaternion embeddings to provide a compact and elegant representation. Specifically, we use timestamp dual quaternions, transforming the entity and relation into temporal entity and temporal relation through dual quaternion multiplication. Besides, we introduce semantic-aware dual quaternion to capture the latent semantics arising from the positional changes of entities within specific relations. Next, TELS consists of two parts: (a) We use semantic-aware dual quaternions to perform transformations on head entity and tail entity respectively through dual quaternion multiplication. Next, we utilize temporal relation to transform head entity to tail entity through dual quaternion multiplication. (b) We adopt an evolutionary hierarchical factor to encapsulate the differences in modulus distribution between the temporal head entity and temporal tail entity. In this way, TELS not only uses dual quaternions to handle key patterns and multiple relations pattern, but also handles evolutionary hierarchical patterns by capturing the modulus distribution differences between temporal entities. Meanwhile, TELS learns semantic-aware dual quaternion embeddings to capture the latent semantics endowed by relations to entities. Empirically, TELS can boost the performance over seven temporal knowledge graph benchmarks.

Boundedness supports children's event representations.

Natural languages distinguish between telic predicates that denote events leading to an inherent endpoint (e.g., draw a balloon) and atelic predicates that denote events with no inherent endpoint (e.g., draw balloons). Telicity distinctions in many languages are already partly available to 4-5-year-olds. Here, using exclusively nonlinguistic tasks and a sample of English-speaking children, we ask whether young learners use corresponding temporal notions to characterize event structure-that is, whether children represent events in cognition as bounded temporal entities with a specified endpoint or unbounded temporal units that could in principle extend indefinitely. We find that 4-5-year-old children in our sample compute boundedness during an event categorization task (Experiment 1) and distinguish event boundedness from event completion (Experiment 2). Furthermore, 4-5-year-olds in our sample evaluate interruptions at event endpoints versus midpoints differently-but only for events that are construed as bounded, presumably because in such construals, events truly culminate (Experiment 3). We conclude that young children represent events in terms of foundational and abstract temporal properties. These properties could support the acquisition of linguistic aspectual distinctions and further scaffold the way children conceptualize and process their dynamic experiences. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Ukrainian biaspectuality: An instantiation of compositional aspect in a verbal-aspect language

Aspect, the perfective-imperfective contrast, is a universal phenomenon, part of man’s cognitive organization to reflect objective/subjective reality by conceptualizing referents of verbs and of nominals/NPs standing for participants in situations as temporal entities, residing in speaker-hearers’ heads and interacting between each other. Aspect is instantiated across languages through two archetypes: verbal aspect (VA) – grammatical, as in the Slavic languages, including Ukrainian; compositional aspect (CA) – complex semantico-syntactic, sporadically dependent on pragmatic discourse elements, as in English. The paper explores Ukrainian language data to, first, confirm that CA, realized mainly as a very complex interplay of sentence components, exists not only in CA languages but, albeit peripherally, also in VA languages, including Ukrainian. Second, to find out how the Ukrainian aspect is realized in sentences with biaspectual verbs and particular numbers of situation-participant NPs: three, two, one. The referents of verbs and of nominals/NPs standing for participants in situations in both VA and CA languages are part of the never-ending process of thinking and perpetual resorting to memory and are not some abstract self-contained system of symbols divorced from human cognition. Phrased otherwise, aspect, especially CA, cannot be understood within the domain of traditional grammar and mainstream linguistics with their naivist notions ignoring man’s cognitive capacity and maintaining, inter alia, that nominals/NPs are concrete/physical or abstract entities. The study of matter is ordained to physics. Linguistics must investigate not the material world but how language reflects this world and other possible (imaginable) worlds. A simple analogue is a woman in a mirror: it is not a material object but an image of a woman; likewise, a woman referred to through language is not a material object but a token of a woman. Hence, NP referents of material things are not physical entities but images of such entities, fully describable, and their kineticism is handled by verb referents, whereby the intricate CA mechanism, which is cognitive, can be observed, albeit peripherally, also in VA languages, including Ukrainian. Disclosure Statement The authors reported no potential conflicts of interest. * Corresponding author: Krasimir Kabakčiev, 0000-0002-5529-0872kkabakciev@atiner.gr

A Chinese telemedicine-dialogue dataset annotated for named entities

BackgroundA large collection of dialogues between patients and doctors must be annotated for medical named entities to build intelligence for telemedicine. However, since most patients involved in telemedicine deliver related named entities in informal and long multiword expressions, it is challenging to tag their telemedicine dialogue data. This study aims to address this issue.MethodsWith the telemedicine dialogue dataset for obstetrics and gynecology taken from haodf.com, we developed guidelines and followed a two-round procedure to tag six types of named entities, including disease, symptom, time, pharmaceutical, operation, and examination. Additionally, we developed four deep-learning models based on this dataset to establish a benchmark for named-entity recognition (NER).ResultsThe distilled obstetrics and gynecology dataset contains 2,383 consultations between doctors and patients, of which 13,411 sentences were from doctors, and 17,929 were from patients. With 63,560 named entities in total, the average number of characters per named entity is 4.33. The experimental results suggest that LatticeLSTM performs best on our dataset in terms of accuracy, precision, recall, and F score.ConclusionCompared with other datasets, this dataset offers three novel facets. This study offers intricately tagged long multiword expressions for medical named entities. Second, this study is one of the first attempts to mark temporal entities in a medical dataset. Third, this annotated dataset is balanced across the six types of labels, which we believe will play a considerable role in expanding telemedicine artificial intelligence.

Exact Learning of Qualitative Constraint Networks from Membership Queries

A Qualitative Constraint Network (QCN) is a constraint graph representing problems under qualitative temporal or spatial relations. More formally, a QCN includes a set of entities and a list of qualitative constraints defining the possible scenarios between these entities. Qualitative constraints are expressed as disjunctions of binary relations capturing the (incomplete) knowledge between the involved entities. QCNs effectively represent various real-world applications, including scheduling and planning, configuration, and Geographic Information Systems (GIS). It is, however, challenging to elicit, from the user, the QCN representing a given problem. To overcome this difficulty in practice, we propose a new algorithm for learning, through membership queries, a QCN from a non-expert. Membership queries are asked to elicit temporal or spatial relationships between pairs of temporal or spatial entities. To improve the time performance of our learning algorithm, constraint propagation and ordering heuristics are enforced. The goal is to reduce the number of membership queries needed to reach the target QCN. We conducted several experiments on randomly generated temporal and spatial QCN instances to assess the practical effect of constraint propagation and ordering heuristics. The results of the experiments are encouraging and promising.

Events and objects are similar cognitive entities

Logico-semantic theories have long noted parallels between the linguistic representation of temporal entities (events) and spatial entities (objects): bounded (or telic) predicates such as fix a car resemble count nouns such as sandcastle because they are “atoms” that have well-defined boundaries, contain discrete minimal parts and cannot be divided arbitrarily. By contrast, unbounded (or atelic) phrases such as drive a car resemble mass nouns such as sand in that they are unspecified for atomic features. Here, we demonstrate for the first time the parallels in the perceptual-cognitive representation of events and objects even in entirely non-linguistic tasks. Specifically, after viewers form categories of bounded or unbounded events, they can extend the category to objects or substances respectively (Experiments 1 and 2). Furthermore, in a training study, people successfully learn event-to-object mappings that respect atomicity (i.e., grouping bounded events with objects and unbounded events with substances) but fail to acquire the opposite, atomicity-violating mappings (Experiment 3). Finally, viewers can spontaneously draw connections between events and objects without any prior training (Experiment 4). These striking similarities between the mental representation of events and objects have implications for current theories of event cognition, as well as the relationship between language and thought.

How to distinguish between semantically close cases: A case study of Mordvin illative and lative

In this paper I discuss the distribution of GOAL-cases in Erzya and Moksha (Mordvin, Uralic). Both languages have two productive GOAL-cases, namely illative and lative, but only one productive case for every other spatial relation (LOCATION, SOURCE, PATH). Furthermore, the two GOAL-cases do not seem to exhibit any clear division of labor (e.g., marking inner and outer GOALS). Previous studies have proposed diff erent solutions, but they do not cover all the occurring variation, and thus are notgeneral enough. In my study I propose that the best way to account for this phenomenon is to investigate the similarities in the language-external situations that are coded with each case. To accomplish this, I adopt the Cognitive Linguistics framework, and look at the conceptualizations of situations where each case is used. I assess the types of nouns used as Landmarks and divide them into eight categories: 0D objects, 2D bounded and unbounded entities, 3D bounded and unbounded entities, institutions, abstract entities, and temporal entities. The bulk of this paper treats the fi rst fi ve categories, i.e. spatial relations. I show that there are clear preferences of coding GOAL with one or the other case when the Landmark represents a certain category. Then I show that similar preferences are at work when coding GOALS with other types of Landmarks as well. After establishing these preferences, I go on to propose a solution to the variation of GOAL-marking in Mordvin languages based on the specifi city of the relation between the Trajector and the Landmark.

Extracting longitudinal anticancer treatments at scale using deep natural language processing and temporal reasoning.

e18747 Background: Accurate longitudinal cancer treatments are vital for establishing primary endpoints such as outcome as well as for the investigation of adverse events. However, many longitudinal therapeutic regimens are not well captured in structured electronic health records (EHRs). Thus, their recognition in unstructured data such as clinical notes is critical to gain an accurate description of the real-world patient treatment journey. Here, we demonstrate a scalable approach to extract high-quality longitudinal cancer treatments from lung cancer patients' clinical notes using a Bidirectional Long Short Term Memory (BiLSTM) and Conditional Random Fields (CRF) based natural language processing (NLP) pipeline. Methods: The lung cancer (LC) cohort of 4,698 patients was curated from the Mount Sinai Healthcare system (2003-2020). Two domain experts developed a structured framework of entities and semantics that captured treatment and its temporality. The framework included therapy type (chemotherapy, targeted therapy, immunotherapy, etc.), status (on, off, hold, planned, etc.) and temporal reasoning entities and relations (admin_date, duration, etc.) We pre-annotated 149 FDA-approved cancer drugs and longitudinal timelines of treatment on the training corpus. A NLP pipeline was implemented with BiLSTM-CRF-based deep learning models to train and then apply the resulting models to the clinical notes of LC cohort. A postprocessor was developed to subsequently post-coordinate and refine the output. We performed both cross-evaluation and independent evaluation to assess the pipeline performance. Results: We applied the NLP pipeline to the 853,755 clinical notes, and identified 1,155 distinct entities for 194 cancer generic drugs, including 74 chemotherapy drugs, 21 immunotherapy drugs, and 99 targeted therapy drugs. We identified chemotherapy, immunotherapy, or targeted therapy data for 3,509 patients in the LC cohort from the clinical notes. Compared to only 2,395 patients with cancer treatments in structured EHR, this pipeline identified cancer treatments from notes for additional 2,303 patients who did not have any available cancer treatment data in the structured EHR. Our evaluation schema indicates that the longitudinal cancer drug recognition pipeline delivers strong performance (named entity recognization for drugs and temporal: F1 = 95%; drug-temporal relation recognition: F1 = 90%). Conclusions: We developed a high-performance BiLSTM-CRF based NLP pipeline to recognize longitudinal cancer treatments. The pipeline recovers and encodes as twice as many patients with cancer treatments compared with structured EHR. Our study indicates deep NLP with temporal reasoning could substantially accelerate the extraction of treatment profiles at scale. The pipeline is adjustable and can be applied across different cancers.

Temporal Design Patterns for Digital Phenotype Cohort Selection in Critical Care: Systematic Literature Assessment and Qualitative Synthesis.

BackgroundInclusion criteria for observational studies frequently contain temporal entities and relations. The use of digital phenotypes to create cohorts in electronic health record–based observational studies requires rich functionality to capture these temporal entities and relations. However, such functionality is not usually available or requires complex database queries and specialized expertise to build them.ObjectiveThe purpose of this study is to systematically assess observational studies reported in critical care literature to capture design requirements and functionalities for a graphical temporal abstraction-based digital phenotyping tool.MethodsWe iteratively extracted attributes describing patients, interventions, and clinical outcomes. We qualitatively synthesized studies, identifying all temporal and nontemporal entities and relations.ResultsWe extracted data from 28 primary studies and 367 temporal and nontemporal entities. We generated a synthesis of entities, relations, and design patterns.ConclusionsWe report on the observed types of clinical temporal entities and their relations as well as design requirements for a temporal abstraction-based digital phenotyping system. The results can be used to inform the development of such a system.

Chronotopic information interaction: integrating temporal and spatial structure for historical indexing and interactive search

Abstract Domain-based learning and research are important applications driving the development of exploratory search systems. A wealth of historical information about events from around the world resides within documents on the web, yet contemporary search engines do not take advantage of the closely integrated temporal and spatial information found within these web pages for indexing and design of search user interfaces. This gap limits the use of the web as a resource for historical and geohistorical information seeking. In this article, we propose chronotopic information interaction as a new interaction concept for web search that explicitly links temporal and spatial entities to keywords using a space–time grid index and a paired search user interface. The space–time grid index allows different modes of interaction between spatial, temporal, and keyword-based views in the search user interface. We demonstrate the use of the space–time grid index and chronotopic information interaction concept with the development of Pteraform, a prototype of a search engine that enables users to explore information in the English version of Wikipedia through a geohistorical lens.

The Persistence of Society-Driven Engagement in Swedish Cinema: A Locational Analysis, 1936-2016

This article underlines the importance of grass-roots civic society involvement in cinema exhibition in villages and towns in Sweden throughout the twentieth century. We investigate geographic patterns of cinema exhibition and ownership over time in a mixed rural cum (post)industrial region in the central part of Sweden known as Bergslagen. In a systematic comparison across temporal and geographic entities our analysis details the extent of civic society engagement in cinema. Our analysis suggests that civic societies contributed to the high number of cinemas in rural and small town Sweden as well as to the persistence of a surprisingly evenly distributed and decentralized pattern of cinema exhibition up until this day.

Temporal information extraction from mental health records to identify duration of untreated psychosis

BackgroundDuration of untreated psychosis (DUP) is an important clinical construct in the field of mental health, as longer DUP can be associated with worse intervention outcomes. DUP estimation requires knowledge about when psychosis symptoms first started (symptom onset), and when psychosis treatment was initiated. Electronic health records (EHRs) represent a useful resource for retrospective clinical studies on DUP, but the core information underlying this construct is most likely to lie in free text, meaning it is not readily available for clinical research. Natural Language Processing (NLP) is a means to addressing this problem by automatically extracting relevant information in a structured form. As a first step, it is important to identify appropriate documents, i.e., those that are likely to include the information of interest. Next, temporal information extraction methods are needed to identify time references for early psychosis symptoms. This NLP challenge requires solving three different tasks: time expression extraction, symptom extraction, and temporal “linking”. In this study, we focus on the first step, using two relevant EHR datasets.ResultsWe applied a rule-based NLP system for time expression extraction that we had previously adapted to a corpus of mental health EHRs from patients with a diagnosis of schizophrenia (first referrals). We extended this work by applying this NLP system to a larger set of documents and patients, to identify additional texts that would be relevant for our long-term goal, and developed a new corpus from a subset of these new texts (early intervention services). Furthermore, we added normalized value annotations (“2011–05”) to the annotated time expressions (“May 2011”) in both corpora. The finalized corpora were used for further NLP development and evaluation, with promising results (normalization accuracy 71–86%). To highlight the specificities of our annotation task, we also applied the final adapted NLP system to a different temporally annotated clinical corpus.ConclusionsDeveloping domain-specific methods is crucial to address complex NLP tasks such as symptom onset extraction and retrospective calculation of duration of a preclinical syndrome. To the best of our knowledge, this is the first clinical text resource annotated for temporal entities in the mental health domain.

The Musical Structure of Time in the Brain: Repetition, Rhythm, and Harmony in fMRI During Rest and Passive Movie Viewing.

Space generally overshadows time in the construction of theories in cognitive neuroscience. In this paper, we pivot from the spatial axes to the temporal, analyzing fMRI image series to reveal structures in time rather than space. To determine affinities among global brain patterns at different times, core concepts in network analysis (derived from graph theory) were applied temporally, as relations among brain images at every time point during an fMRI scanning epoch. To explore the temporal structures observed through this adaptation of network analysis, data from 180 subjects in the Human Connectome Project were examined, during two experimental conditions: passive movie viewing and rest. The temporal brain, like the spatial brain, exhibits a modular structure, where “modules” are intermittent (distributed in time). These temporal entities are here referred to as themes. Short sequences of themes – motifs – were studied in sequences from 4 to 11 s in length. Many motifs repeated at constant intervals, and are therefore rhythmic; rhythms, converted to frequencies, were often harmonic. We speculate that the structure and interaction of these global oscillations underwrites the capacity to experience and navigate a world which is both recognizably stable and noticeably changing at every moment – a temporal world. In its temporal structure, this brain-constituted world resembles music.

Why Monday is notin front ofTuesday: On the uses of English and Finnish front adpositions in sequence metaphors of time

AbstractI analyze English and Finnishfrontadpositions commonly used insequencemetaphors of time (The days ahead of Christmas are busy). Asequencemetaphor represents the order of temporal entities as their different positions on a path along which they are moving, a temporally earlier entity being ‘ahead of’ a later one, i.e. further advanced in the direction of motion. In English, the motion-relatedfrontprepositionahead ofis conventional insequencemetaphors, as opposed toin front of, which is used only occasionally and often results in ill-formedness (*Monday is in front of Tuesday). In Finnish, the dedicated two-mover adpositionedellä‘ahead of [two-mover]’ is conventional insequencemetaphors, while the generalfrontadpositionedessäis not. Based on an analysis of these adpositions in different scenarios of spatial motion, I argue that the adpositions common insequencemetaphors (ahead of, edellä) evoke a motion frame of reference (motion FoR). In the motion FoR,frontis adjacent to the direction where the moving Ground (the entity designated by the complement of the adposition) is headed. In contrast, adpositions that bear a strong association with the standard (intrinsic or relative) FoRs (in front ofandedessä) are less felicitous insequencemetaphors. Together, the two languages demonstrate how a metaphorical motion scenario in asequencemetaphor is grammatically coded as a stable, unchanging array, where both participants are steadily moving in the same direction.

Enabling Spatio-Temporal Search in Open Data

Intuitively, most datasets found on governmental Open Data portals are organized by spatio-temporal criteria, that is, single datasets provide data for a certain region, valid for a certain time period. Likewise, for many use cases (such as, for instance, data journalism and fact checking) a pre-dominant need is to scope down the relevant datasets to a particular period or region. Rich spatio-temporal annotations are therefore a crucial need to enable semantic search for (and across) Open Data portals along those dimensions, yet – to the best of our knowledge – no working solution exists. To this end, we (i) present a scalable approach to construct a spatio-temporal knowledge graph that hierarchically structures geographical as well as temporal entities, (ii) annotate a large corpus of tabular datasets from open data portals with entities from this knowledge graph, and (iii) enable structured, spatio-temporal search and querying over Open Data catalogs, both via a search interface as well as via a SPARQL endpoint, available at data.wu.ac.at/odgraphsearch/.

Interval vs. Point Temporal Logic Model Checking

In recent years, model checking with interval temporal logics is emerging as a viable alternative to model checking with standard point-based temporal logics, such as LTL, CTL, CTL*, and the like. The behavior of the system is modeled by means of (finite) Kripke structures, as usual. However, while temporal logics which are interpreted “point-wise” describe how the system evolves state-by-state, and predicate properties of system states, those which are interpreted “interval-wise” express properties of computation stretches, spanning a sequence of states. A proposition letter is assumed to hold over a computation stretch (interval) if and only if it holds over each component state (homogeneity assumption). A natural question arises: is there any advantage in replacing points by intervals as the primary temporal entities, or is it just a matter of taste? In this article, we study the expressiveness of Halpern and Shoham’s interval temporal logic (HS) in model checking, in comparison with those of LTL, CTL, and CTL*. To this end, we consider three semantic variants of HS: the state-based one, introduced by Montanari et al. in [30, 34], that allows time to branch both in the past and in the future, the computation-tree-based one, that allows time to branch in the future only, and the trace-based variant, that disallows time to branch. These variants are compared among themselves and to the aforementioned standard logics, getting a complete picture. In particular, we show that HS with trace-based semantics is equivalent to LTL (but at least exponentially more succinct), HS with computation-tree-based semantics is equivalent to finitary CTL*, and HS with state-based semantics is incomparable with all of them (LTL, CTL, and CTL*).

On Temporal Aspects in Cross-Cultural e-Collaboration Between Finland and Japan Research Teams

Time is an essential dimension in cross-cultural e-collaboration among research project teams. Understanding temporal aspects and project dynamics in cross-cultural research e-collaboration and related processes can improve team members' skills in cross-cultural communication and increase their cultural competence. The present case cultures are Finnish and Japanese, and the case universities are the University of Jyväskylä (Finland) and Keio University (Japan). Three issues are addressed in this article. First, cultural dimensions and time models in the cross-cultural e-collaboration context are discussed. Second, temporal aspects related to e-collaboration activities are introduced. Third, formal, ontological approaches for identifying and describing temporal entities in cross-cultural e-collaboration are presented and examples of applications are given. The objectives of this article are (1) to deepen the knowledge and understanding of temporal aspects (informal and formal) in a cross-cultural e-collaboration environment (CCeCE) and (2) to create know-how for designing CCeCE-like systems.

Enabling Spatio-Temporal Search in Open Data

Intuitively, most datasets found on governmental Open Data portals are organized by spatio-temporal criteria, that is, single datasets provide data for a certain region, valid for a certain time period. Likewise, for many use cases (such as, for instance, data journalism and fact checking) a pre-dominant need is to scope down the relevant datasets to a particular period or region. Rich spatio-temporal annotations are therefore a crucial need to enable semantic search for (and across) Open Data portals along those dimensions, yet - to the best of our knowledge - no working solution exists. To this end, we (i) present a scalable approach to construct a spatio-temporal knowledge graph that hierarchically structures geographical as well as temporal entities, (ii) annotate a large corpus of tabular datasets from open data portals with entities from this knowledge graph, and (iii) enable structured, spatio-temporal search and querying over Open Data catalogs, both via a search interface as well as via a SPARQL endpoint, available at data.wu.ac.at/odgraphsearch/.