2,919 research outputs found
Predicting Correlations Between Lexical Alignments and Semantic Inferences
No full textWhile there is a strong intuition that word alignments (e.g. synonymy, hyperonymy) play a relevant role in recognizing text-to-text semantic inferences (e.g. textual entailment, semantic similarity), this intuition is often not reflected in the system performances and there is a general need of a deeper comprehension of the role of lexical resources. This paper provides an empirical analysis of the dependencies between data-sets, lexical resources and algorithms that are commonly used in text-to-text inference tasks. We define a resource impact index , based on lexical alignments between pairs of texts, and
show that such index is significantly correlated with the performance of different textual entailment algorithms. The result is an operational, algorithm-independent, procedure for predicting the performance
of a class of available RTE algorithms
Estimating Lexical Resources Impact in Text-to-Text Inference Tasks
No full textThis paper provides an empirical analysis of both the datasets and the
lexical resources that are commonly used in text-to-text inference tasks (e.g. textual entailment, semantic similarity). According to the analysis, we define an index for the impact of a lexical resource, and we show that such index significantly correlates with the performance of a textual entailment system
FBK-HLT-NLP at SemEval-2016 Task 2: A Multitask, Deep Learning Approach for Interpretable Semantic Textual Similarity
No full textWe present the system developed at FBK for the SemEval 2016 Shared Task 2 ”Interpretable Semantic Textual Similarity” as well as the results of the submitted runs. We use a single neural network classification model for predicting the alignment at chunk level, the relation type of the alignment and the similarity scores. Our best run was ranked as first in one the subtracks (i.e. raw input data, Student Answers), among 12 runs submitted, and the approach proved to be very robust across the different datasets
Comparing Machine Learning and Deep Learning Approaches on NLP Tasks for the Italian Language
No full textWe present a comparison between deep learning and traditional machine learning methods for various NLP tasks in Italian. We carried on experiments using available datasets (e.g., from the Evalita shared tasks) on two sequence tagging tasks (i.e., named entities recognition and nominal entities recognition) and four classification tasks (i.e., lexical relations among words, semantic relations among sentences, sentiment analysis and text classification). We show that deep learning approaches outperform traditional machine learning algorithms in sequence tagging, while for classification tasks that heavily rely on semantics approaches based on feature engineering are still competitive. We think that a similar analysis could be carried out for other languages to provide an assessment of machine learning / deep learning models across different languages
A Chatbot (Juno) Prototype to Deploy a Behavioral Activation Intervention to Pregnant Women: Qualitative Evaluation Using a Multiple Case Study
No full textBackground: Despite the increasing focus on perinatal care, preventive digital interventions are still scarce. Furthermore, the literature suggests that the design and development of these interventions are mainly conducted through a top-down approach that limitedly accounts for direct end user perspectives. Objective: Building from a previous co-design study, this study aimed to qualitatively evaluate pregnant women's experiences with a chatbot (Juno) prototype designed to deploy a preventive behavioral activation intervention. Methods: Using a multiple-case study design, the research aims to uncover similarities and differences in participants' perceptions of the chatbot while also exploring women's desires for improvement and technological advancements in chatbot-based interventions in perinatal mental health. Five pregnant women interacted weekly with the chatbot, operationalized in Telegram, following a 6-week intervention. Self-report questionnaires were administered at baseline and postintervention time points. About 10-14 days after concluding interactions with Juno, women participated in a semistructured interview focused on (1) their personal experience with Juno, (2) user experience and user engagement, and (3) their opinions on future technological advancements. Interview transcripts, comprising 15 questions, were qualitatively evaluated and compared. Finally, a text-mining analysis of transcripts was performed. Results: Similarities and differences have emerged regarding women's experiences with Juno, appreciating its esthetic but highlighting technical issues and desiring clearer guidance. They found the content useful and pertinent to pregnancy but differed on when they deemed it most helpful. Women expressed interest in receiving increasingly personalized responses and in future integration with existing health care systems for better support. Accordingly, they generally viewed Juno as an effective momentary support but emphasized the need for human interaction in mental health care, particularly if increasingly personalized. Further concerns included overreliance on chatbots when seeking psychological support and the importance of clearly educating users on the chatbot's limitations. Conclusions: Overall, the results highlighted both the positive aspects and the shortcomings of the chatbot-based intervention, providing insight into its refinement and future developments. However, women stressed the need to balance technological support with human interactions, particularly when the intervention involves beyond preventive mental health context, to favor a greater and more reliable monitoring
FBK-HLT: A New Framework for Semantic Textual Similarity
No full textThis paper reports the description and perfor- mance of our system, FBK-HLT, participat- ing in the SemEval 2015, Task #2 “Semantic Textual Similarity”, English subtask. We sub- mitted three runs with different hypothesis in combining typical features (lexical similarity, string similarity, word n-grams, etc) with syn- tactic structure features, resulting in different sets of features. The results evaluated on both STS 2014 and 2015 datasets prove our hypoth- esis of building a STS system taking into con- sideration of syntactic information. We out- perform the best system on STS 2014 datasets and achieve a very competitive result to the best system on STS 2015 datasets
How to Use Gazetteers for Entity Recognition with Neural Models
No full textAlthough the use of end-to-end neural architectures has been proven to be effective on several sequence labeling tasks, the use of gazetteers in these architectures is still rather unexplored. We investigate several options, aiming at exploiting gazetteers to extract relevant features, and then at integrating these
features in a neural model for entity recognition. We provide experimental evidences on two datasets (named entities and nominal entities) and two languages (English and Italian), showing that extracting features from a rich model of the gazetteer and then concatenating such features with the input embeddings of a
neural model is the best strategy in all our experimental settings, significantly outperforming more conventional approaches
- …
