Across countries, a noteworthy paradigm shift towards a more sustainable and environmentally responsible economy is underway. However, this positive transition is accompanied by an upsurge in greenwashing, where companies make exaggerated claims about their environmental commitments. To address this challenge and protect consumers, initiatives have emerged to substantiate green claims. With the proliferation of environmental and scientific assertions, a critical need arises for automated methods to detect and validate these claims at scale. In this paper, we introduce EnClaim, a transformer network architecture augmented with stylistic features for automatically detecting claims from open web documents or social media posts. The proposed model considers various linguistic stylistic features in conjunction with language models to predict whether a given statement constitutes a claim. We have rigorously evaluated the model using multiple open datasets. Our initial findings indicate that incorporating stylistic vectors alongside the BERT-based language model enhances the overall effectiveness of environmental claim detection.

The escalating prevalence of food safety incidents within the food supply chain necessitates immediate action to protect consumers. These incidents encompass a spectrum of issues, including food product contamination and deliberate food and feed adulteration for economic gain leading to outbreaks and recalls. Understanding the origins and pathways of contamination is imperative for prevention and mitigation. In this paper, we introduce FORCE Foodborne disease Outbreak and ReCall Event extraction from openweb). Our proposed model leverages a multi-tasking sequence labeling architecture in conjunction with transformer-based document embeddings. We have compiled a substantial annotated corpus comprising relevant articles published between 2011 and 2023 to train and evaluate the model. The dataset will be publicly released with the paper. The event detection model demonstrates fair accuracy in identifying food-related incidents and outbreaks associated with organizations, as assessed through cross-validation techniques.

In this paper we propose a framework for automatic translation of English text to American Sign Language (ASL) which leverages a linguistically informed transformer model to translate English sentences into ASL gloss sequences. These glosses are then associated with respective ASL videos, effectively representing English text in ASL. To facilitate experimentation, we create an English-ASL parallel dataset on banking domain.Our preliminary results demonstrated that the linguistically informed transformer model achieves a 97.83% ROUGE-L score for text-to-gloss translation on the ASLG-PC12 dataset. Furthermore, fine-tuning the transformer model on the banking domain dataset yields an 89.47% ROUGE-L score when fine-tuned on ASLG-PC12 + banking domain dataset. These results demonstrate the effectiveness of the linguistically informed model for both general and domain-specific translations. To facilitate parallel dataset generation in banking-domain, we choose ASL despite having limited benchmarks and data corpus compared to some of the other sign languages.

Obtaining demand trends for products is an essential aspect of supply chain planning. It helps in generating scenarios for simulation before actual demands start pouring in. Presently, experts obtain this number manually from different News sources. In this paper, we have presented methods that can automate the information acquisition process. We have presented a joint framework that performs information extraction and sentiment analysis to acquire demand related information from business text documents. The proposed system leverages a TwinBERT-based deep neural network model to first extract product information for which demand is associated and then identify the respective sentiment polarity. The articles are also subjected to causal analytics, that, together yield rich contextual information about reasons for rise or fall of demand of various products. The enriched information is targeted for the decision-makers, analysts and knowledge workers. We have exhaustively evaluated our proposed models with datasets curated and annotated for two different domains namely, automobile sector and housing. The proposed model outperforms the existing baseline systems.

The number of assistive technologies available for dyslexia in Bangla is low and most of them do not use multisensory teaching methods. As a solution, a computer-based audio-visual system Dy-poThon is proposed to teach sentence reading in Bangla. It incorporates the multisensory teaching method through three activities, listening, reading, and writing, checks the reading and writing ability of the user and tracks the response time. A criteria-based evaluation was conducted with 28 special educators to evaluate Dy-poThon. Content, efficiency, ease of use and aesthetics are evaluated using a standardised questionnaire. The result suggests that Dy-poThon is useful for teaching Bangla sentence-reading.

Advanced neural network architectures have provided several opportunities to develop systems to automatically capture information from domain-specific unstructured text sources. The FinSim4-ESG shared task, collocated with the FinNLP workshop, proposed two sub-tasks. In sub-task1, the challenge was to design systems that could utilize contextual word embeddings along with sustainability resources to elaborate an ESG taxonomy. In the second sub-task, participants were asked to design a system that could classify sentences into sustainable or unsustainable sentences. In this paper, we utilize semantic similarity features along with BERT embeddings to segregate domain terms into a fixed number of class labels. The proposed model not only considers the contextual BERT embeddings but also incorporates Word2Vec, cosine, and Jaccard similarity which gives word-level importance to the model. For sentence classification, several linguistic elements along with BERT embeddings were used as classification features. We have shown a detailed ablation study for the proposed models.

Automatic extraction of cause-effect relationships from natural language texts is a challenging open problem in Artificial Intelligence. Most of the early attempts at its solution used manually constructed linguistic and syntactic rules on restricted domain data sets. With the advent of big data, and the recent popularization of deep learning, the paradigm to tackle this problem has slowly shifted. In this work we proposed a transformer based architecture to automatically detect causal sentences from textual mentions and then identify the corresponding cause-effect relations. We describe our submission to the FinCausal 2022 shared task based on this method. Our model achieves a F1-score of 0.99 for the Task-1 and F1-score of 0.60 for Task-2 on the shared task data set on financial documents.

Eligibility criteria in the clinical trials specify the characteristics that a patient must or must not possess in order to be treated according to a standard clinical care guideline. As the process of manual eligibility determination is time-consuming, automatic structuring of the eligibility criteria into various semantic categories or aspects is the need of the hour. Existing methods use hand-crafted rules and feature-based statistical machine learning methods to dynamically induce semantic aspects. However, in order to deal with paucity of aspect-annotated clinical trials data, we propose a novel weakly-supervised co-training based method which can exploit a large pool of unlabeled criteria sentences to augment the limited supervised training data, and consequently enhance the performance. Experiments with 0.2M criteria sentences show that the proposed approach outperforms the competitive supervised baselines by 12% in terms of micro-averaged F1 score for all the aspects. Probing deeper into analysis, we observe domain-specific information boosts up the performance by a significant margin.

Performance of standard text analytics algorithms are known to be substantially degraded on consumer generated data, which are often very noisy. These algorithms also do not work well on enterprise data which has a very different nature from News repositories, storybooks or Wikipedia data. Text cleaning is a mandatory step which aims at noise removal and correction to improve performance. However, enterprise data need special cleaning methods since it contains many domain terms which appear to be noise against a standard dictionary, but in reality are not so. In this work we present detailed analysis of characteristics of enterprise data and suggest unsupervised methods for cleaning these repositories after domain terms have been automatically segregated from true noise terms. Noise terms are thereafter corrected in a contextual fashion. The effectiveness of the method is established through careful manual evaluation of error corrections over several standard data sets, including those available for hate speech detection, where there is deliberate distortion to avoid detection. We also share results to show enhancement in classification accuracy after noise correction.

In this paper, we demonstrate a system for the automatic extraction and curation of crime-related information from multi-source digitally published News articles collected over a period of five years. We have leveraged the use of deep convolution recurrent neural network model to analyze crime articles to extract different crime related entities and events. The proposed methods are not restricted to detecting known crimes only but contribute actively towards maintaining an updated crime ontology. We have done experiments with a collection of 5000 crime-reporting News articles span over time, and multiple sources. The end-product of our experiments is a crime-register that contains details of crime committed across geographies and time. This register can be further utilized for analytical and reporting purposes.

In this paper we present a qualitatively enhanced deep convolution recurrent neural network for computing the quality of a text in an automatic essay scoring task. The novelty of the work lies in the fact that instead of considering only the word and sentence representation of a text, we try to augment the different complex linguistic, cognitive and psycological features associated within a text document along with a hierarchical convolution recurrent neural network framework. Our preliminary investigation shows that incorporation of such qualitative feature vectors along with standard word/sentence embeddings can give us better understanding about improving the overall evaluation of the input essays.

In this paper we have proposed a linguistically informed recursive neural network architecture for automatic extraction of cause-effect relations from text. These relations can be expressed in arbitrarily complex ways. The architecture uses word level embeddings and other linguistic features to detect causal events and their effects mentioned within a sentence. The extracted events and their relations are used to build a causal-graph after clustering and appropriate generalization, which is then used for predictive purposes. We have evaluated the performance of the proposed extraction model with respect to two baseline systems,one a rule-based classifier, and the other a conditional random field (CRF) based supervised model. We have also compared our results with related work reported in the past by other authors on SEMEVAL data set, and found that the proposed bi-directional LSTM model enhanced with an additional linguistic layer performs better. We have also worked extensively on creating new annotated datasets from publicly available data, which we are willing to share with the community.

In this paper, we have explored web-based evidence gathering and different linguistic features to automatically extract drug names from tweets and further classify such tweets into Adverse Drug Events or not. We have evaluated our proposed models with the dataset as released by the SMM4H workshop shared Task-1 and Task-3 respectively. Our evaluation results shows that the proposed model achieved good results, with Precision, Recall and F-scores of 78.5%, 88% and 82.9% respectively for Task1 and 33.2%, 54.7% and 41.3% for Task3.

Any real world events or trends that can affect the company’s growth trajectory can be considered as risk. There has been a growing need to automatically identify, extract and analyze risk related statements from news events. In this demonstration, we will present a risk analytics framework that processes enterprise project management reports in the form of textual data and news documents and classify them into valid and invalid risk categories. The framework also extracts information from the text pertaining to the different categories of risks like their possible cause and impacts. Accordingly, we have used machine learning based techniques and studied different linguistic features like n-gram, POS, dependency, future timing, uncertainty factors in texts and their various combinations. A manual annotation study from management experts using risk descriptions collected for a specific organization was conducted to evaluate the framework. The evaluation showed promising results for automated risk analysis and identification.

In this paper we have developed an open-source online computational framework that can be used by different research groups to conduct reading researches on Indian language texts. The framework can be used to develop a large annotated Indian language text comprehension data from different user based experiments. The novelty in this framework lies in the fact that it brings different empirical data-collection techniques for text comprehension under one roof. The framework has been customized specifically to address language particularities for Indian languages. It will also offer many types of automatic analysis on the data at different levels such as full text, sentence and word level. To address the subjectivity of text difficulty perception, the framework allows to capture user background against multiple factors. The assimilated data can be automatically cross referenced against varying strata of readers.

Machine transliteration is used in a number of NLP applications ranging from machine translation and information retrieval to input mechanisms for non-roman scripts. Many popular Input Method Editors for Indian languages, like Baraha, Akshara, Quillpad etc, use back-transliteration as a mechanism to allow users to input text in a number of Indian language. The lack of a standard dataset to evaluate these systems makes it difficult to make any meaningful comparisons of their relative accuracies. In this paper, we describe the methodology for the creation of a dataset of ~2500 transliterated sentence pairs each in Bangla, Hindi and Telugu. The data was collected across three different modes from a total of 60 users. We believe that this dataset will prove useful not only for the evaluation and training of back-transliteration systems but also help in the linguistic analysis of the process of transliterating Indian languages from native scripts to Roman.