
Oier Lopez de LacalleUniversity of Edinburgh | UoE · Institute for Language, Cognition and Computation (ILCC)
Oier Lopez de Lacalle
PhD (Computer Science)
About
59
Publications
5,404
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
1,438
Citations
Introduction
Skills and Expertise
Additional affiliations
Publications
Publications (59)
Large Language Models (LLMs) exhibit extensive knowledge about the world, but most evaluations have been limited to global or anglocentric subjects. This raises the question of how well these models perform on topics relevant to other cultures, whose presence on the web is not that prominent. To address this gap, we introduce BertaQA, a multiple-ch...
Translate-test is a popular technique to improve the performance of multilingual language models. This approach works by translating the input into English using an external machine translation system, and running inference over the translated input. However, these improvements can be attributed to the use of a separate translation system, which is...
Automatic Short Answer Grading aims to automatically grade short answers authored by students. Recent work has shown that this task can be effectively reformulated as a Natural Language Inference problem. State-of-the-art is defined by the use of large pretrained language models fine-tuned in the domain dataset. But how to quantify the effectivenes...
This chapter landscapes the field of Language Technology (LT) and language- centric AI by assembling a comprehensive state-of-the-art of basic and applied research in the area. It sketches all recent advances in AI, including the most recent deep learning neural technologies. The chapter brings to light not only where language-centric AI as a whole...
Language Models are the core for almost any Natural Language Processing system nowadays. One of their particularities is their contextualized representations, a game changer feature when a disambiguation between word senses is necessary. In this paper we aim to explore to what extent language models are capable of discerning among senses at inferen...
Integrating outside knowledge for reasoning in visio-linguistic tasks such as visual question answering (VQA) is an open problem. Given that pretrained language models have been shown to include world knowledge, we propose to use a unimodal (text-only) train and inference procedure based on automatic off-the-shelf captioning of images and pretraine...
Recent work has shown that NLP tasks such as Relation Extraction (RE) can be recasted as Textual Entailment tasks using verbalizations, with strong performance in zero-shot and few-shot settings thanks to pre-trained entailment models. The fact that relations in current RE datasets are easily verbalized casts doubts on whether entailment would be e...
The current workflow for Information Extraction (IE) analysts involves the definition of the entities/relations of interest and a training corpus with annotated examples. In this demonstration we introduce a new workflow where the analyst directly verbalizes the entities/relations, which are then used by a Textual Entailment model to perform zero-s...
Integrating outside knowledge for reasoning in visio-linguistic tasks such as visual question answering (VQA) is an open problem. Given that pretrained language models have been shown to include world knowledge, we propose to use a unimodal (text-only) train and inference procedure based on automatic off-the-shelf captioning of images and pretraine...
Relation extraction systems require large amounts of labeled examples which are costly to annotate. In this work we reformulate relation extraction as an entailment task, with simple, hand-made, verbalizations of relations produced in less than 15 min per relation. The system relies on a pretrained textual entailment engine which is run as-is (no t...
Generating an image from its textual description requires both a certain level of language understanding and common sense knowledge about the spatial relations of the physical entities being described. In this work, we focus on inferring the spatial relation between entities, a key step in the process of composing scenes based on text. More specifi...
Generating an image from its textual description requires both a certain level of language understanding and common sense knowledge about the spatial relations of the physical entities being described. In this work, we focus on inferring the spatial relation between entities, a key step in the process of composing scenes based on text. More specifi...
Automatic generation of reading comprehension questions is a topic receiving growing interest in the NLP community, but there is currently no consensus on evaluation metrics and many approaches focus on linguistic quality only while ignoring the pedagogic value and appropriateness of questions. This paper overcomes such weaknesses by a new evaluati...
In this paper we present a relation extraction system that given a text extracts pedagogically motivated relation types, as a previous step to obtaining a semantic representation of the text which will make possible to automatically generate questions for reading comprehension. The system maps pedagogically motivated relations with relations from C...
The combination of visual and textual representations has produced excellent results in tasks such as image captioning and visual question answering, but the inference capabilities of multimodal representations are largely untested. In the case of textual representations, inference tasks such as Textual Entailment and Semantic Textual Similarity ha...
In this paper we introduce vSTS, a new dataset for measuring textual similarity of sentences using multimodal information. The dataset is comprised by images along with its respectively textual captions. We describe the dataset both quantitatively and qualitatively, and claim that it is a valid gold standard for measuring automatic multimodal textu...
UKB is an open source collection of programs for performing, among other tasks, knowledge-based Word Sense Disambiguation (WSD). Since it was released in 2009 it has been often used out-of-the-box in sub-optimal settings. We show that nine years later it is the state-of-the-art on knowledge-based WSD. This case shows the pitfalls of releasing open...
Laburpena: Lan honetan erantzunen kalifikazio automatikorako lehen urratsak azaltzen ditugu. Urrats horiek azaltzeko irakaskuntzaren domeinuko bi sistema deskribatzen ditugu: bata, erantzunak hierarkikoki antolatzen eta kalifikatzen dituena; eta, bestea, kalifikazioaren azalpena emateko asmoz, erantzunaren zatiak erreferentzia baten aurka alderatze...
We introduce a distantly supervised event extraction approach that extracts complex event templates from microblogs. We show that this near real-time data source is more challenging than news because it contains information that is both approximate (e.g., with values that are close but different from the gold truth) and ambiguous (due to the brevit...
Word Sense Disambiguation (WSD) systems automatically choose the intended meaning of a word in context. In this article we present a WSD algorithm based on random walks over large Lexical Knowledge Bases (LKB). We show that our algorithm performs better than other graph-based methods when run on a graph built from WordNet and eXtended WordNet. Our...
Relation Extraction methods based on Distant Supervision rely on true tuples to retrieve noisy mentions, which are then used to train traditional supervised relation extraction methods. In this paper we analyze the sources of noise in the mentions, and explore simple methods to filter out noisy mentions. The results show that a combination of menti...
Large amounts of digital cultural heritage (CH) information have become available over the past years, requiring more powerful exploration systems than just a search box. The PATHS system aims to provide an environment in which users can successfully explore a large, unknown collection through two modalities: following existing paths to learn about...
Digitisation of the cultural heritage means that a significant amount of material is now available through online digital library portals. However, the vast quantity of cultural heritage material can also be overwhelming for many users who lack knowledge of the collections, subject knowledge and the specialist language used to describe this content...
In this paper we present a novel approach to learning semantic models for multiple domains, which we use to categorize Wikipedia pages and to perform domain Word Sense Disambiguation (WSD). In order to learn a semantic model for each domain we first extract relevant terms from the texts in the domain and then use these terms to initialize a random...
This document describes the prelimi-nary release of the integrated Kyoto sys-tem for specific domain WSD. The sys-tem uses concept miners (Tybots) to ex-tract domain-related terms and produces a domain-related thesaurus, followed by knowledge-based WSD based on word-net graphs (UKB). The resulting system can be applied to any language with a lexica...
Laburpena: Gure hizkuntza anbiguoa da. Hitz batek hainbat interpretazio ditu agertzen den testuinguruaren arabera, eta zein adiera hartzen duen asmatzea ez da lan erraza, nahiz eta guk era naturalean egin. konputazio-metodoak erabiliz hitzen agerpenei adiera egokia ematea hitzaren adiera-desanbiguazioa (HAD) deritzo. HAD automatikoa ezagutzan oina-...
This paper explores the application of knowledge- based Word Sense Disambiguation systems to spe- cific domains, based on our state-of-the-art graph- based WSD system that uses the information in WordNet. Evaluation was performed over a pub- licly available domain-specific dataset of 41 words related to Sports and Finance, comprising exam- ples dra...
The lack of positive results on super- vised domain adaptation for WSD have cast some doubts on the utility of hand- tagging general corpora and thus devel- oping generic supervised WSD systems. In this paper we show for the first time that our WSD system trained on a general source corpus (BNC) and the target corpus, obtains up to 22% error reduct...
Domain portability and adaptation of NLP components and Word Sense Disambiguation systems present new challenges. The diffi- culties found by supervised systems to adapt might change the way we assess the strengths and weaknesses of supervised and knowledge- based WSD systems. Unfortunately, all ex- isting evaluation datasets for specific domains a...
Domain portability and adaptation of NLP components and Word Sense Disambiguation systems present new challenges. The difficulties found by supervised systems to adapt might change the way we assess the strengths and weaknesses of supervised and knowledgebased WSD systems. Unfortunately, all existing evaluation datasets for specific domains are lex...
This article focuses on Word Sense Disambiguation (WSD), which is a Natural Language Processing task that is thought to be important for many Language Technology applications, such as Information Retrieval, Information Extraction, or Machine Translation. One of the main issues preventing the deployment of WSD technology is the lack of training exam...
In this paper we explore robustness and domain adaptation issues for Word Sense Disambiguation (WSD) using Singular Value Decomposition (SVD) and unlabeled data. We focus on the semi-supervised domain adaptation scenario, where we train on the source corpus and test on the target corpus, and try to improve results using unlabeled data. Our method y...
This paper presents a first attempt of an application-drivenevaluation exercise of WSD. We used a CLIR testbed from the Cross Lingual Evaluation Forum. The expansion, indexing and retrieval strategies where fixed by the organizers. The participants had to return both the topics and documents tagged with WordNet 1.6 word senses. The organization pro...
This paper describes the joint submission of two systems to the all-words WSD sub-task of SemEval-2007 task 17. The main goal of this work was to build a competitive unsupervised system by combining heterogeneous algorithms. As a secondary goal, we explored the integration of unsupervised predictions into a supervised system by different means.
This work describes the University of the Basque Country system (UBC-ALM) for lexical sample and all-words WSD subtasks of SemEval-2007 task 17, where it per-formed in the second and fifth positions re-spectively. The system is based on a com-bination of k-Nearest Neighbor classifiers, with each classifier learning from a distinct set of features:...
This paper presents a first attempt of an application-driven evaluation exercise of WSD. We used a CLIR testbed from the Cross Lingual Evaluation Forum. The expansion, indexing and retrieval strategies where fixed by the organizers. The participants had to return both the topics and documents tagged with WordNet 1.6 word senses. The organization pr...
This work describes the University of the Basque Country system (UBC-ALM) for lexical sample and all-words WSD subtasks of SemEval-2007 task 17, where it performed in the second and fifth positions respectively. The system is based on a combination of k-Nearest Neighbor classifiers, with each classifier learning from a distinct set of features: loc...
This paper explores the use of two graph algorithms for unsupervised induction and tagging of nominal word senses based on corpora. Our main contribution is the op- timization of the free parameters of those algorithms and its evaluation against pub- licly available gold standards. We present a thorough evaluation comprising super- vised and unsupe...
Véronis (2004) has recently proposed an innovative unsupervised algorithm for word sense disambiguation based on small-world graphs called HyperLex. This paper explores two sides of the algorithm. First, we extend Véronis' work by optimizing the free parameters (on a set of words which is different to the target set). Second, given that the empiric...
This paper explores the split of features sets in order to obtain better wsd systems through combinations of classifiers learned over each of the split feature sets. Our results show that only k-NN is able to profit from the combination of split features, and that simple voting is not enough for that. Instead we propose combining all k-NN subsystem...
Topic signatures are context vectors built for concepts. They can be automatically acquired for any concept hierarchy using simple methods. This paper explores the correlation between a distributional-based semantic similarity based on topic signatures and several hierarchy-based similarities. We show that topic signatures can be used to approximat...
Topic signatures are context vectors built for word senses and concepts. They can be automatically acquired from the web for any concept hierarchy using the "monosemous relative" method. Topic signatures have been shown to be useful in Word Sense Disambiguation, for modeling similarity between word senses, classifying new terms in hierarchies and a...
This paper presents the results of a set of methods to cluster WordNet word senses. The methods rely on different information sources: confusion matrixes from Senseval-2 Word Sense Disambiguation systems, translation similarities, hand-tagged examples of the target word senses and examples obtained automatically from the web for the target word sen...
This paper describes a European project called PATHS (Personalized Access To cultural Heritage Spaces) that aims to support information exploration and discovery through digital cultural heritage collections. Significant amounts of cultural heritage material are now available through online digital library portals, wich can also be overwhelming for...
This paper describes our submissions for the Slot Filling task of TAC-KBP 2011. The sys-tem takes as baseline the one we developed for the 2010 edition (Intxaurrondo et al., 2010), which is based on distant supervision. We did a straightforward implementation, trained us-ing snippets of the document collection con-taining both entity and filler fro...