Spanish Insolvency Law 1/2020 of the 5th of May enables individuals to apply for debt waiver under certain conditions. The large number of applications submitted each year places a heavy burden on judges and court officers, who must examine heterogeneous documentation before issuing a ruling. This paper presents an AI-based assistant designed to support the processing of debt waiver cases. The system integrates PDF-to-text conversion, rule-based document classification, large language model (LLM)-based information extraction, and post-processing to consolidate fragmented or duplicated records. A front-end interface provides structured summaries of the application content, and can automatically generate draft rulings. Evaluated on a set of real applications, the system achieves over 92% F1 in document classification and up to 91% F1 in personal data extraction, showing the potential of open-source LLMs to reduce administrative workload and accelerate judicial procedures, while keeping the final decision with the judge.

Abstract Research‐oriented universities often comprise numerous researchers of various types and possess complex research structures that encompass research groups, departments, laboratories, and research institutes. In this situation, understanding the university's strengths and areas of excellence requires careful examination. Additionally, individuals at different levels of governance (e.g., department heads, directors of research institutes, rectors) may seek to establish synergies among researchers to tackle issues such as international project applications or industry technology transfer. University officials and faculty members frequently require the expertise of specific research groups or individuals, but struggle to obtain this information beyond their personal networks. This limits their ability to locate necessary resources effectively. Fortunately, most institutions have databases containing publications that could provide valuable insights into areas of strength within the university. In this article, we present a visual analysis application capable of addressing these questions and assisting management in making informed decisions regarding governance measures such as creating new research institutes. Our system has been evaluated by domain experts, who found it highly beneficial and expressed interest in utilising it regularly.

In this work, we focus on improving the captions generated by image-caption generation systems. We propose a novel re-ranking approach that leverages visual-semantic measures to identify the ideal caption that maximally captures the visual information in the image. Our re-ranker utilizes the Belief Revision framework (Blok et al., 2003) to calibrate the original likelihood of the top-n captions by explicitly exploiting semantic relatedness between the depicted caption and the visual context. Our experiments demonstrate the utility of our approach, where we observe that our re-ranker can enhance the performance of a typical image-captioning system without necessity of any additional training or fine-tuning.

Computing a mapping between two process models is a crucial technique, since it enables reasoning and operating across processes, like providing a similarity score between two processes, or merging different process variants to generate a consolidated process model. In this paper we present a new flexible technique for process model mapping, based on the relaxation labeling constraint satisfaction algorithm. The technique can be instantiated so that different modes are devised, depending on the context. For instance, it can be adapted to the case where one of the mapped process models is incomplete, or it can be used to ground an adaptable similarity measure between process models. The approach has been implemented inside the open platform NLP4BPM, providing a visualization of the performed mappings and computed similarity scores. The experimental results witness the flexibility and usefulness of the technique proposed.

Ahmed Sabir, Francesc Moreno, Lluís Padró. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

The Business Process Management (BPM) field focuses in the coordination of labor so that organizational processes are smoothly executed in a way that products and services are properly delivered. At the same time, NLP has reached a maturity level that enables its widespread application in many contexts, thanks to publicly available frameworks. In this position paper, we show how NLP has potential in raising the benefits of BPM practices at different levels. Instead of being exhaustive, we show selected key challenges were a successful application of NLP techniques would facilitate the automation of particular tasks that nowadays require a significant effort to accomplish. Finally, we report on applications that consider both the process perspective and its enhancement through NLP.

Xavier Carreras, Lluís Padró, Lei Zhang, Achim Rettinger, Zhixing Li, Esteban García-Cuesta, Željko Agić, Božo Bekavac, Blaz Fortuna, Tadej Štajner. Proceedings of the Demonstrations at the 14th Conference of the European Chapter of the Association for Computational Linguistics. 2014.

This article presents an ensemble parse approach to detecting and selecting high-quality linguistic analyses output by a hand-crafted HPSG grammar of Spanish implemented in the LKB system. The approach uses full agreement (i.e., exact syntactic match) along with a MaxEnt parse selection model and a statistical dependency parser trained on the same data. The ultimate goal is to develop a hybrid corpus annotation methodology that combines fully automatic annotation and manual parse selection, in order to make the annotation task more efficient while maintaining high accuracy and the high degree of consistency necessary for any foreseen uses of a treebank.

This work is focused on research in machine learning for coreference resolution. Coreference resolution is a natural language processing task that consists of determining the expressions in a discourse that refer to the same entity. The main contributions of this article are (i) a new approach to coreference resolution based on constraint satisfaction, using a hypergraph to represent the problem and solving it by relaxation labeling; and (ii) research towards improving coreference resolution performance using world knowledge extracted from Wikipedia. The developed approach is able to use an entity-mention classification model with more expressiveness than the pair-based ones, and overcome the weaknesses of previous approaches in the state of the art such as linking contradictions, classifications without context, and lack of information evaluating pairs. Furthermore, the approach allows the incorporation of new information by adding constraints, and research has been done in order to use world knowledge to improve performances. RelaxCor, the implementation of the approach, achieved results at the state-of-the-art level, and participated in international competitions: SemEval-2010 and CoNLL-2011. RelaxCor achieved second place in CoNLL-2011.

We present a robust approach for linking already existing lexical/semantic hierarchies. We used a constraint satisfaction algorithm (relaxation labeling) to select - among a set of candidates- the node in a target taxonomy that bests matches each node in a source taxonomy. In particular, we use it to map the nominal part of WordNet 1.5 onto WordNet 1.6, with a very high precision and a very low remaining ambiguity.

We present an algorithm that automatically learns context constraints using statistical decision trees. We then use the acquired constraints in a flexible POS tagger. The tagger is able to use information of any degree: n-grams, automatically learned context constraints, linguistically motivated manually written constraints, etc. The sources and kinds of constraints are unrestricted, and the language model can be easily extended, improving the results. The tagger has been tested and evaluated on the WSJ corpus.