Disambiguation Process Research Articles

Global Coronal Plasma Diagnostics Based on Multislit Extreme-ultraviolet Spectroscopy

Full-disk spectroscopic observations of the solar corona are highly desired to forecast solar eruptions and their impact on planets and to uncover the origin of solar wind. In this paper, we introduce a new multislit design (five slits) to obtain extreme-ultraviolet (EUV) spectra simultaneously. The selected spectrometer wavelength range (184–197 Å) contains several bright EUV lines that can be used for spectral diagnostics. The multislit approach offers an unprecedented way to efficiently obtain the global spectral data but the ambiguity from different slits should be resolved. Using a numerical simulation of the global corona, we primarily concentrate on the optimization of the disambiguation process, with the objective of extracting decomposed spectral information of six primary lines. This subsequently facilitates a comprehensive series of plasma diagnostics, including density (Fe xii 195.12/186.89 Å), Doppler velocity (Fe xii 193.51 Å), line width (Fe xii 193.51 Å), and temperature diagnostics (Fe viii 185.21 Å, Fe x 184.54 Å, Fe xi 188.22 Å, and Fe xii 193.51 Å). We find a good agreement between the forward modeling parameters and the inverted results at the initial eruption stage of a coronal mass ejection, indicating the robustness of the decomposition method and its immense potential for global monitoring of the solar corona.

A graph based named entity disambiguation using clique partitioning and semantic relatedness

Disambiguating name mentions in texts is a crucial task in Natural Language Processing, especially in entity linking. The credibility and efficiency of such systems depend largely on this task. For a given name entity mention in a text, there are many potential candidate entities that may refer to it in the knowledge base. Therefore, it is very difficult to assign the correct candidate from the whole set of candidate entities of this mention. To solve this problem, collective entity disambiguation is a prominent approach. In this paper, we present a novel algorithm called CPSR for collective entity disambiguation, which is based on a graph approach and semantic relatedness. A clique partitioning algorithm is used to find the best clique that contains a set of candidate entities. These candidate entities provide the answers to the corresponding mentions in the disambiguation process. To evaluate our algorithm, we carried out a series of experiments on seven well-known datasets, namely, AIDA/CoNLL2003-TestB, IITB, MSNBC, AQUAINT, ACE2004, Cweb, and Wiki. The Kensho Derived Wikimedia Dataset (KDWD) is used as the knowledge base for our system. From the experimental results, our CPSR algorithm outperforms both the baselines and other well-known state-of-the-art approaches.

Triple alignment-enhanced complex question answering over knowledge bases

Program induction is a crucial paradigm for complex question answering over knowledge bases. In the existing learning framework, the predicted program is required to strictly align (word-by-word) with a gold program, which could cause over penalization for minor deviations. Meanwhile, due to the existence of synonyms, program induction question answering often fails to retrieve answer from the knowledge base because individual function argument cannot perfectly replicate the target argument. To address above misalignment problems, we propose a triple alignment-enhanced complex question answering (TACQA) method by incorporating global token alignment, function alignment, and argument alignment. First, apart from the classical global token alignment, the predicted functions are extracted and aligned separately with the gold functions, enabling efficient learning of implicit structural information related to the query framework of program from input questions. Second, an argument alignment is introduced to correct the ambiguous function arguments, which enhances the disambiguation processing efficiency of multi-argument by optimizing candidate pool construction and similarity calculation. The experiments on KQA Pro show that our method consistently outperforms the SOTA methods, demonstrating the effectiveness of triple alignment processing mechanism for simultaneously addressing function misalignment and argument ambiguity in program induction and further improving the model performance.

Navigating Real-World Partial Label Learning: Unveiling Fine-Grained Images with Attributes

Partial label learning (PLL), a significant research area, addresses the challenge of annotating each sample with a candidate label set containing the true label when obtaining accurate labels is infeasible. However, existing PLL methods often rely on generic datasets like CIFAR, where annotators can readily differentiate candidate labels and are unlikely to confuse, making it less realistic for real-world partial label applications. In response, our research focuses on a rarely studied problem, PLL on fine-grained images with attributes. And we propose a novel framework called Shared to Learn, Distinct to Disambiguate (SoDisam). Within the candidate label set, the categories may exhibit numerous shared attribute features, posing a challenge in accurately distinguishing them. Rather than perceiving it as an impediment, we capitalize on these shared attributes as definitive sources of supervision. This insight guides us to learn attribute space visual representation to focus on the information from these shared attributes. Moreover, we introduce an attribute attention mechanism tailored to harness the remaining distinct attributes. This mechanism directs the originally holistic feature towards specific regions, capturing corresponding discriminative features. In addition, a dynamic disambiguation module is introduced, continuously adjusting the two aforementioned mechanisms and achieve the final disambiguation process. Extensive experiments demonstrate the effectiveness of our approach on fine-grained partial label datasets. The proposed SoDisam framework not only addresses the challenges associated with fine-grained partial label learning but also provides a more realistic representation of real-world partial label scenarios.

Robust Chinese Short Text Entity Disambiguation Method Based on Feature Fusion and Contrastive Learning

To address the limitations of existing methods of short-text entity disambiguation, specifically in terms of their insufficient feature extraction and reliance on massive training samples, we propose an entity disambiguation model called COLBERT, which fuses LDA-based topic features and BERT-based semantic features, as well as using contrastive learning, to enhance the disambiguation process. Experiments on a publicly available Chinese short-text entity disambiguation dataset show that the proposed model achieves an F1-score of 84.0%, which outperforms the benchmark method by 0.6%. Moreover, our model achieves an F1-score of 74.5% with a limited number of training samples, which is 2.8% higher than the benchmark method. These results demonstrate that our model achieves better effectiveness and robustness and can reduce the burden of data annotation as well as training costs.

Learning multi-granular worker intentions from incomplete visual observations for worker-robot collaboration in construction

Construction is often considered a dangerous and labor-intensive industry, with safety and workforce shortages plaguing the industry globally. Owing to its flexibility, Worker-Robot Collaboration (WRC) could combine the strengths of humans and robots, alleviating human workers from dangerous and tedious tasks and boosting productivity. WRC envisions robots assisting workers appropriately, which necessitates the ability to understand workers' intentions. However, challenges arise when robots recognize workers' intentions in construction. First, the highly unstructured and dynamic construction workplace prevents robots from capturing consistent long-term worker activities, which is detrimental to the information accumulation and disambiguation processes in intention recognition. Second, the long-horizon tasks with implicit steps make it difficult for robots to find a suitable granularity to capture workers' intentions. Considering these challenges, this study seeks to (1) explore whether it is feasible to recognize worker intentions with temporally incomplete observations and (2) investigate which granularity is best suited for recognizing worker intentions. To tackle the first challenge, a data-and-knowledge fusion method is proposed, where the contextual information from visual data is combined with task knowledge to mitigate the information loss due to incomplete observations. To address the second challenge, the best-suited granularity is determined by comparing the recognition accuracy and intention modeling ability of different granularities after defining and recognizing multi-granular worker intentions using a hierarchical task structure and Multi-Task Learning (MTL), respectively. Results show that the proposed method can recognize multi-granular worker intentions with macro-F1 scores higher than 0.85, and that the intermediate activity is the best-suited granularity as it strikes a good balance between intention recognition accuracy and intention modeling capability. This study contributes to the existing body of knowledge by recognizing multi-granular worker intentions and determining the appropriate granularity under temporally incomplete observations. The findings will help facilitate seamless WRC in construction.

On a BiCIKL to Wikidata: Harmonizing the chaotic universe of natural history collectors

People play a key role in science and have been getting increased recognition for this work, initially by the efforts of libraries to harmonize attribution of authors of creative works, such as in the Virtual International Authority File (VIAF). Now, in the realm of scientific publishing, the international ORCID (Open Researcher and Contributor ID) organization allows scientists to mint globally unique and self-maintained identifiers for themselves, which tie all of their scientific output together. This has been fairly successful for formal publications, but less so for other scientific contributions, such as samples or specimens collected from organisms to serve as vouchers or references for long-term study. Such specimens may end up in natural history collections, where their metadata and, in particular, their taxonomic identity is refined over the years by experts. As the cost of digitization has decreased and the methods become more refined, digital versions of these specimens are now being published online in vast numbers, making attribution with identifiers such as ORCID easier. However, many of these specimens are only linked to the people who worked on them by name strings, often using a variety of syntaxes, transliterations and abbreviations. Recently, some collections have made an effort to disambiguate these names by enriching them with an explicit link to a persistent identifier, such as an ORCID (Groom et al. 2022, Little et al. 2022, von Mering et al. 2022, Meeus et al. 2023). In the European Union-funded BiCIKL (Biodiversity Community Integrated Knowledge) project, one of the goals is to develop better workflows that aid and streamline this enrichment process. This approach focuses primarily on Wikidata, as a fall-back resource for cases where ORCID identifiers are not available. Wikidata is suitable for people who are unable to register for an ORCID identifier, but also acts as a broker to harmonize between other authority sources such as the fore-mentioned VIAF. In this presentation, we will show the results achieved so far in this BiCIKL effort. First, we will provide a landscape analysis of different person identifier enrichment efforts so far in the natural history sector. Infrastructures such as the Global Biodiversity Information Facility (GBIF), the related Bionomia and the Botany Pilot (Güntsch et al. 2021) will play a key role. Building on those results, we will identify improved methods to link name strings to unique biographical records in resources such as Wikidata and ORCID. The focus will primarily be on properties of these records that make them more identifiable as collectors of specimens, and that can systematically be harnessed to rank multiple possible candidates or reject false positives. Finally, these findings will be incorporated into workflows that streamline the disambiguation and enrichment process. These workflows will build on existing pipelines and platforms, such as Bionomia and the Botany Pilot, focusing in particular on data roundtripping from these resources into other infrastructures, such as local collection management systems, but also specimen data aggregators such as GBIF and the Distributed System of Scientific Collections (DiSSCo).

SPEEDSER: A Possibilistic System for Query Disambiguation, Expansion and Translation

We design, implement and assess in this paper a new architecture of a possibilistic mono- and cross-language information retrieval (IR/CLIR) system. The latter is useful to experiment query disambiguation, expansion and translation processes in both IR and CLIR frameworks. We take advantage of possibility theory to overcome the problems of query disambiguation and expansion in an uncertain and imprecise IR/CLIR context. We investigate the impact of combining possibilistic query disambiguation with expansion on IR/CLIR efficiency. A co-occurrence graph representation is exploited to quantify the similarity between query terms and their semantically close words (expansion task) or between query terms and their possible meanings (disambiguation task). We extend the possibilistic mono-language query disambiguation approach to a cross-language framework. We conduct a set of experiments using ROMANSEVAL data collection, the French–English parallel text corpus Europarl and the CLEF-2003 French–English IR/CLIR test collection. Results highlight some statistically significant improvements of our possibilistic approaches when compared to some state-of-the-art IR/CLIR works.

Non-ambiguous and simplified quantum teleportation protocol

In this study, a new version of the quantum teleportation protocol is presented, which does not require a Bell state measurement (BSM) module on the sender side (Alice), a unitary transform to reconstruct the teleported state on the receiver side (Bob), neither a disambiguation process through two classic bits that travel through a classic disambiguation channel located between sender and receiver. The corresponding theoretical deduction of the protocol, as well as the experimental verification of its operation for several examples of qubits through implementation on an optical table, complete the present study. Both the theoretical and experimental outcomes show a marked superiority in the performance of the new protocol over the original version, with more simplicity and lower implementation costs, and identical fidelity in its most complete version.

Automated Word Sense Disambiguation Using WordNet Ontology

Automatic word sense disambiguation is a major challenge in natural language processing domain. In recent years, many of supervised and knowledge-based approaches were developed to solve this problem. The use of a sense inventory as a knowledge background to disambiguate words is a very useful technique rather than supervised approaches, with its need of a large pre-trained text corpus. This paper proposes a new approach to disambiguate words in text based on WordNet ontology, as sense inventory. The authors introduce a new technique called Gloss+ for word-sense disambiguation (WSD), which is based on using of the glosses of WordNet synonyms of the target word and the local context in which this word is used. This technique exploits a special behavior of the polysemous synsets. This behavior is detected during the disambiguation process and is used to improve the results obtained. In the experiment part, the authors compare the proposed approach to the methodologies which use synonyms or glosses only.

Learning about the meanings of ambiguous words: evidence from a word-meaning priming paradigm with short narratives.

Fluent language comprehension requires people to rapidly activate and integrate context-appropriate word meanings. This process is challenging for meanings of ambiguous words that are comparatively lower in frequency (e.g., the “bird” meaning of “crane”). Priming experiments have shown that recent experience makes such subordinate (less frequent) word meanings more readily available at the next encounter. These experiments used lists of unconnected sentences in which each ambiguity was disambiguated locally by neighbouring words. In natural language, however, disambiguation may occur via more distant contextual cues, embedded in longer, connected communicative contexts. In the present experiment, participants (N = 51) listened to 3-sentence narratives that ended in an ambiguous prime. Cues to disambiguation were relatively distant from the prime; the first sentence of each narrative established a situational context congruent with the subordinate meaning of the prime, but the remainder of the narrative did not provide disambiguating information. Following a short delay, primed subordinate meanings were more readily available (compared with an unprimed control), as assessed by responses in a word association task related to the primed meaning. This work confirms that listeners reliably disambiguate spoken ambiguous words on the basis of cues from wider narrative contexts, and that they retain information about the outcome of these disambiguation processes to inform subsequent encounters of the same word form.

No service is an island: towards an ecosystem approach to mental health service evaluation.

No service is an island: towards an ecosystem approach to mental health service evaluation.

Exploring Representations for Singular and Multi-Concept Relations for Biomedical Named Entity Normalization.

Since the rise of the COVID-19 pandemic, peer-reviewed biomedical repositories have experienced a surge in chemical and disease related queries. These queries have a wide variety of naming conventions and nomenclatures from trademark and generic, to chemical composition mentions. Normalizing or disambiguating these mentions within texts provides researchers and data-curators with more relevant articles returned by their search query. Named entity normalization aims to automate this disambiguation process by linking entity mentions onto their appropriate candidate concepts within a biomedical knowledge base or ontology. We explore several term embedding aggregation techniques in addition to how the term's context affects evaluation performance. We also evaluate our embedding approaches for normalizing term instances containing one or many relations within unstructured texts.

The prevalence and impact of university affiliation discrepancies between four bibliographic databases—Scopus, Web of Science, Dimensions, and Microsoft Academic

Abstract Research managers benchmarking universities against international peers face the problem of affiliation disambiguation. Different databases have taken separate approaches to this problem and discrepancies exist between them. Bibliometric data sources typically conduct a disambiguation process that unifies variant institutional names and those of its subunits so that researchers can then search all records from that institution using a single unified name. This study examined affiliation discrepancies between Scopus, Web of Science (WoS), Dimensions, and Microsoft Academic for 18 Arab universities over a 5-year period. We confirmed that digital object identifiers (DOIs) are suitable for extracting comparable scholarly material across databases and quantified the affiliation discrepancies between them. A substantial share of records assigned to the selected universities in any one database were not assigned to the same university in another. The share of discrepancy was higher in the larger databases (Dimensions and Microsoft Academic). The smaller, more selective databases (Scopus and especially WoS) tended to agree to a greater degree with affiliations in the other databases. Manual examination of affiliation discrepancies showed that they were caused by a mixture of missing affiliations, unification differences, and assignation of records to the wrong institution.

Principle-Based Approach for the De-Identification of Code-Mixed Electronic Health Records

Code-mixing is a phenomenon where at least two languages are combined in a hybrid manner in the context of a single conversation. The use of mixed language is widespread in multilingual and multicultural countries and poses significant challenges for the development of automated language processing tools. In Taiwan’s electronic health record (EHR) systems, unstructured EHR texts are usually represented in a mixture of English and Chinese which increases the difficulty for de-identification and synthetization of protected health information (PHI). We explored this problem by applying several state-of-the-art pre-trained mono- and multilingual language models and propose to exploit the principle-based approach (PBA) for the tasks of PHI recognition and resynthesis on a code-mixed EHR corpus annotated with 6 main categories and 25 subcategories of PHIs. A hierarchical principle slot schema is defined in the PBA to encode knowledge of code-mixed PHIs and utilize slots to learn from the training set to assemble principles for recognizing PHI mentions and synthesizing surrogates simultaneously. In addition, a semantic disambiguation process is implemented to disambiguate ambiguous PHI categories in the de-identification process and to dynamically extend the knowledge encoded in PBA during the knowledge augmentation process. The experiment results demonstrate that the proposed method can achieve the best micro- and macro-F-scores in comparison to the other mono- and multilingual language models fine-tuned on our code-mixed corpus.

Concept-Based Approach to Word-Sense Disambiguation

The task of automatically determining the correct sense of a polysemous word has remained a challenge to this day. In our research, we introduce Concept-Based Disambiguation (CBD), a novel framework that utilizes recent semantic analysis techniques to represent both the context of the word and its senses in a high-dimensional space of natural concepts. The concepts are retrieved from a vast encyclopedic resource, thus enriching the disambiguation process with large amounts of domain-specific knowledge. In such concept-based spaces, more comprehensive measures can be applied in order to pick the right sense. Additionally, we introduce a novel representation scheme, denoted anchored representation, that builds a more specific text representation associated with an anchoring word. We evaluate our framework and show that the anchored representation is more suitable to the task of word-sense disambiguation (WSD). Additionally, we show that our system is superior to state-of-the-art methods when evaluated on domain-specific corpora, and competitive with recent methods when evaluated on a general corpus.

Is Your Collection Ambiguous?

The natural history specimens of the world have been documented on paper labels, often physically attached to the specimen itself. As we transcribe these data to make them digital and more useful for analysis, we make interpretations. Sometimes these interpretations are trivial, because the label is unambiguous, but often the meaning is not so clear, even if it is easily read. One key element that suffers from considerable ambiguity is people’s names. Though a person is indivisible, their name can change, is rarely unique and can be written in many ways. Yet knowing the people associated with data is incredibly useful. Data on people can be used to validate other data, simplify data capture, link together data across domains, reduce duplication-of-effort and facilitate data-gap-analysis. In addition, people data enable the discovery of individuals unique to our collections, the collective charting of the history of scientific researchers and the provision of credit to the people who deserve it (Groom et al. 2020). We foresee a future where the people associated with collections are not ambiguous, are shared globally, and data of all kinds are linked through the people who generate them. The TDWG People in Biodiversity Data Task Group is therefore working on a guide to the disambiguation of people in natural history collections. The ultimate goal is to connect the various strings of characters on specimen labels and other documentation to persistent identifiers (PIDs) that unambiguously link a name “string” to the identity of a person. In working towards this goal, 150 volunteers in the Bionomia project have linked 21 million specimens to persistent identifiers for their collectors and determiners. An additional 2 million specimens with links to identifiers for people have already emerged directly from collections that make use of the recently ratified Darwin Core terms recordedByID and identifiedByID. Furthermore, the CETAF Botany Pilot conducted among a group of European herbaria and museums has connected over 1.4 million specimens to disambiguated collectors (Güntsch et al. 2021). Still, given the estimated 2 billion (Ariño 2010) natural history specimens globally, there is much more disambiguation to be done. The process of disambiguation starts with a trigger, which is often the transcription of a specimen’s label data. Unambiguous identification of the collector may facilitate this transcription, as it offers knowledge of their biographical details and collecting habits, allowing us to infer missing information such as collecting date or locality. Another trigger might be the flagging of inconsistent data during data entry or resulting from data quality processes, revealing for instance that multiple collectors have been conflated. A disambiguation trigger is followed by the gathering of data, then the evaluation of the results and finally by the documentation of the new information. Disambiguation is not always straightforward and there are many pitfalls. It requires access to biographical data, and identifiers to be minted. In the case of living people, they have to cooperate with being disambiguated and we have to follow legal and ethical guidelines. In the case of dead people, particularly those long dead, disambiguation may require considerable research. We will present the progress made by the People in Biodiversity Data Task Group and their recommendations for disambiguation in collections. We want to encourage other institutions to engage with a global effort of linking people to persistent identifiers to collaboratively improve all collection data.

The Measure of the Archive: The Ro­bustness of Network Analysis in Early Modern Correspondence

Network analysis of historical correspondence can be a fruitful way to address historical research questions, and has been increasingly used in historical studies over the past decade. As with many areas of quantitative humanities research, the reliability of the results are often called into question, given that such approaches require ’hard data’ as input, yet almost inevitably use datasets with partial or missing records. Other disciplines using network analysis have conducted robustness experiments designed to test the impact of data loss or error on their results. In order to test how this missing data might affect our own area of research, we conducted a number of experiments designed to simulate the impact of the kinds of loss often seen in historical correspondence data, including random document loss, missing years, and errors in the disambiguation and de-duplication process. The results show that most network centrality measures maintain robustness until a very large proportion of the data (60% or more) is removed. Some measures showed a linear change in robustness, while others remained high and then fell off sharply. Only one, transitivity (local clustering coefficient) was significantly impacted throughout. We tested a range of data loss scenarios (random single letters, folio books of manuscript letters, catalogues, and entire years) and a range of commonly used network metrics. In addition, we tested the robustness of more complex network analysis results in the literature that combine several network metrics to highlight individuals in the network, and found that the same types of individuals would have likely been highlighted even with 50% random letter loss. Alongside the article is a web application, built using Shiny, which will calculate robustness measures for a user-uploaded network dataset. We conclude that researchers working with similar historical correspondence datasets might be able to consider network analysis results to be robust in most cases, rather than work on the assumption that missing data would lead to very different findings or results.

Scientometric analysis of the Chaos journal (1991–2019): From descriptive statistics to complex networks viewpoints

We performed a scientometric analysis of Chaos papers from 1991 to 2019, applying a careful disambiguation process for identifying the authors correctly. First, we used standard scientometric tools based on descriptive statistics. This analysis enabled us to compute productivity and the degree of collaboration. The evolution in the number of authors, countries, and topics per article has an increasing trend. An analysis of the citations considering their temporal mean number exhibits a growing tendency in time. Second, we dealt with Lotka-Zipf's law considering the rank distributions of 15 datasets. We found that the sum of Crossref citations by country was the only dataset for which the power-law was the only plausible distribution. Next, we examined the networks of authors, countries, and topics, going from the simplest case of undirected and unweighted networks to the general case of weighted and directed networks and assigning a weight to the individual nodes. Based on the networks' topology and features, we introduced diversity, collaboration, influence, and productivity measures and found a significant increase in the diversity of all the considered networks (authors, countries, and topics) but manifesting a very different network structure. The computation of typical network quantities combined with the communities' identification reveals the presence of several hubs and the existence of various communities that encompass nodes of all the continents in the case of countries. Finally, using the most general networks, it was possible to compute influence and productivity indexes to find the USA, China, and Germany's leadership inside the network.

Working memory and lexical ambiguity resolution in Cantonese Chinese.

The present study examined how working memory functions in the underlying mechanism of the lexical disambiguation process (in activation approach or in inhibition approach). We recruited sixty native Cantonese listeners to participate in two experimental tasks: (a) a Cantonese-version reading span task to measure their working memory (WM) capacity and (b) a standard cross-modal priming task to measure the lexical disambiguation time. The results revealed that (1) the underlying mechanism of the disambiguation process seemed favorable for an inhibition approach and (2) the frequency of the individual meanings of the ambiguous words and the numbers of their meanings might interact with the WM capacity during lexical access, particularly for the low-WM span group.