The ability of Large Language Models (LLMs) to encode syntactic and semantic structures of language is well examined in NLP. Additionally, analogy identification, in the form of word analogies are extensively studied in the last decade of language modeling literature. In this work we specifically look at how LLMs’ abilities to capture sentence analogies (sentences that convey analogous meaning to each other) vary with LLMs’ abilities to encode syntactic and semantic structures of sentences. Through our analysis, we find that LLMs’ ability to identify sentence analogies is positively correlated with their ability to encode syntactic and semantic structures of sentences. Specifically, we find that the LLMs which capture syntactic structures better, also have higher abilities in identifying sentence analogies.

In the fast-paced domain of Large Language Models (LLMs), the issue of hallucination is a prominent challenge. Despite continuous endeavors to address this concern, it remains a highly active area of research within the LLM landscape. Grasping the intricacies of this problem can be daunting, especially for those new to the field. This tutorial aims to bridge this knowledge gap by introducing the emerging realm of hallucination in LLMs. It will comprehensively explore the key aspects of hallucination, including benchmarking, detection, and mitigation techniques. Furthermore, we will delve into the specific constraints and shortcomings of current approaches, providing valuable insights to guide future research efforts for participants.

Automatic fact verification has received significant attention recently. Contemporary automatic fact-checking systems focus on estimating truthfulness using numerical scores which are not human-interpretable. A human fact-checker generally follows several logical steps to verify a verisimilitude claim and conclude whether it’s truthful or a mere masquerade. Popular fact-checking websites follow a common structure for fact categorization such as half true, half false, false, pants on fire, etc. Therefore, it is necessary to have an aspect-based (delineating which part(s) are true and which are false) explainable system that can assist human fact-checkers in asking relevant questions related to a fact, which can then be validated separately to reach a final verdict. In this paper, we propose a 5W framework (who, what, when, where, and why) for question-answer-based fact explainability. To that end, we present a semi-automatically generated dataset called FACTIFY-5WQA, which consists of 391, 041 facts along with relevant 5W QAs – underscoring our major contribution to this paper. A semantic role labeling system has been utilized to locate 5Ws, which generates QA pairs for claims using a masked language model. Finally, we report a baseline QA system to automatically locate those answers from evidence documents, which can serve as a baseline for future research in the field. Lastly, we propose a robust fact verification system that takes paraphrased claims and automatically validates them. The dataset and the baseline model are available at https: //github.com/ankuranii/acl-5W-QA

Over the past decade, analogies, in the form of word-level analogies, have played a significant role as an intrinsic measure of evaluating the quality of word embedding methods such as word2vec. Modern large language models (LLMs), however, are primarily evaluated on extrinsic measures based on benchmarks such as GLUE and SuperGLUE, and there are only a few investigations on whether LLMs can draw analogies between long texts. In this paper, we present ANALOGICAL, a new benchmark to intrinsically evaluate LLMs across a taxonomy of analogies of long text with six levels of complexity – (i) word, (ii) word vs. sentence, (iii) syntactic, (iv) negation, (v) entailment, and (vi) metaphor. Using thirteen datasets and three different distance measures, we evaluate the abilities of eight LLMs in identifying analogical pairs in the semantic vector space. Our evaluation finds that it is increasingly challenging for LLMs to identify analogies when going up the analogy taxonomy.

With the rise of prolific ChatGPT, the risk and consequences of AI-generated text has increased alarmingly. This triggered a series of events, including an open letter, signed by thousands of researchers and tech leaders in March 2023, demanding a six-month moratorium on the training of AI systems more sophisticated than GPT-4. To address the inevitable question of ownership attribution for AI-generated artifacts, the US Copyright Office released a statement stating that “if the content is traditional elements of authorship produced by a machine, the work lacks human authorship and the office will not register it for copyright”. Furthermore, both the US and the EU governments have recently drafted their initial proposals regarding the regulatory framework for AI. Given this cynosural spotlight on generative AI, AI-generated text detection (AGTD) has emerged as a topic that has already received immediate attention in research, with some initial methods having been proposed, soon followed by the emergence of techniques to bypass detection. This paper introduces the Counter Turing Test (CT2), a benchmark consisting of techniques aiming to offer a comprehensive evaluation of the robustness of existing AGTD techniques. Our empirical findings unequivocally highlight the fragility of the proposed AGTD methods under scrutiny. Amidst the extensive deliberations on policy-making for regulating AI development, it is of utmost importance to assess the detectability of content generated by LLMs. Thus, to establish a quantifiable spectrum facilitating the evaluation and ranking of LLMs according to their detectability levels, we propose the AI Detectability Index (ADI). We conduct a thorough examination of 15 contemporary LLMs, empirically demonstrating that larger LLMs tend to have a lower ADI, indicating they are less detectable compared to smaller LLMs. We firmly believe that ADI holds significant value as a tool for the wider NLP community, with the potential to serve as a rubric in AI-related policy-making.

The recent advancements in Large Language Models (LLMs) have garnered widespread acclaim for their remarkable emerging capabilities. However, the issue of hallucination has parallelly emerged as a by-product, posing significant concerns. While some recent endeavors have been made to identify and mitigate different types of hallucination, there has been a limited emphasis on the nuanced categorization of hallucination and associated mitigation methods. To address this gap, we offer a fine-grained discourse on profiling hallucination based on its degree, orientation, and category, along with offering strategies for alleviation. As such, we define two overarching orientations of hallucination: (i) factual mirage (FM) and (ii) silver lining (SL). To provide a more comprehensive understanding, both orientations are further sub-categorized into intrinsic and extrinsic, with three degrees of severity - (i) mild, (ii) moderate, and (iii) alarming. We also meticulously categorize hallucination into six types: (i) acronym ambiguity, (ii) numeric nuisance, (iii) generated golem, (iv) virtual voice, (v) geographic erratum, and (vi) time wrap. Furthermore, we curate HallucInation eLiciTation (HILT), a publicly available dataset comprising of 75,000 samples generated using 15 contemporary LLMs along with human annotations for the aforementioned categories. Finally, to establish a method for quantifying and to offer a comparative spectrum that allows us to evaluate and rank LLMs based on their vulnerability to producing hallucinations, we propose Hallucination Vulnerability Index (HVI). Amidst the extensive deliberations on policy-making for regulating AI development, it is of utmost importance to assess and measure which LLM is more vulnerable towards hallucination. We firmly believe that HVI holds significant value as a tool for the wider NLP community, with the potential to serve as a rubric in AI-related policy-making. In conclusion, we propose two solution strategies for mitigating hallucinations.

Combating disinformation is one of the burning societal crises - about 67% of the American population believes that disinformation produces a lot of uncertainty, and 10% of them knowingly propagate disinformation. Evidence shows that disinformation can manipulate democratic processes and public opinion, causing disruption in the share market, panic and anxiety in society, and even death during crises. Therefore, disinformation should be identified promptly and, if possible, mitigated. With approximately 3.2 billion images and 720,000 hours of video shared online daily on social media platforms, scalable detection of multimodal disinformation requires efficient fact verification. Despite progress in automatic text-based fact verification (e.g., FEVER, LIAR), the research community lacks substantial effort in multimodal fact verification. To address this gap, we introduce FACTIFY 3M, a dataset of 3 million samples that pushes the boundaries of the domain of fact verification via a multimodal fake news dataset, in addition to offering explainability through the concept of 5W question-answering. Salient features of the dataset include: (i) textual claims, (ii) ChatGPT-generated paraphrased claims, (iii) associated images, (iv) stable diffusion-generated additional images (i.e., visual paraphrases), (v) pixel-level image heatmap to foster image-text explainability of the claim, (vi) 5W QA pairs, and (vii) adversarial fake news stories.

Word embeddings, i.e., semantically meaningful vector representation of words, are largely influenced by the distributional hypothesis You shall know a word by the company it keeps (Harris, 1954), whereas modern prediction- based neural network embeddings rely on de- sign choices and hyperparameter optimization. Word embeddings like Word2Vec, GloVe etc. well capture the contextuality and real-world analogies but contemporary convolution-based image embeddings such as VGGNet, AlexNet, etc. do not capture contextual knowledge. The popular king-queen analogy does not hold true for most commonly used vision embeddings. In this paper, we introduce a pre-trained joint embedding (JE), named IMAGINATOR, trained on 21K distinct image objects. JE is a way to encode multimodal data into a vec- tor space where the text modality serves as the grounding key, which the complementary modality (in this case, the image) is anchored with. IMAGINATOR encapsulates three in- dividual representations: (i) object-object co- location, (ii) word-object co-location, and (iii) word-object correlation. These three ways cap- ture complementary aspects of the two modal- ities which are further combined to obtain the final object-word JEs. Generated JEs are intrinsically evaluated to assess how well they capture the contextual- ity and real-world analogies. We also evalu- ate pre-trained IMAGINATOR JEs on three downstream tasks: (i) image captioning, (ii) Im- age2Tweet, and (iii) text-based image retrieval. IMAGINATOR establishes a new standard on the aforementioned downstream tasks by out- performing the current SoTA on all the selected tasks. The code is available at https:// github.com/varunakk/IMAGINATOR.

Conversational Agents (CAs) powered with deep language models (DLMs) have shown tremendous promise in the domain of mental health. Prominently, the CAs have been used to provide informational or therapeutic services (e.g., cognitive behavioral therapy) to patients. However, the utility of CAs to assist in mental health triaging has not been explored in the existing work as it requires a controlled generation of follow-up questions (FQs), which are often initiated and guided by the mental health professionals (MHPs) in clinical settings. In the context of ‘depression’, our experiments show that DLMs coupled with process knowledge in a mental health questionnaire generate 12.54% and 9.37% better FQs based on similarity and longest common subsequence matches to questions in the PHQ-9 dataset respectively, when compared with DLMs without process knowledge support. Despite coupling with process knowledge, we find that DLMs are still prone to hallucination, i.e., generating redundant, irrelevant, and unsafe FQs. We demonstrate the challenge of using existing datasets to train a DLM for generating FQs that adhere to clinical process knowledge. To address this limitation, we prepared an extended PHQ-9 based dataset, PRIMATE, in collaboration with MHPs. PRIMATE contains annotations regarding whether a particular question in the PHQ-9 dataset has already been answered in the user’s initial description of the mental health condition. We used PRIMATE to train a DLM in a supervised setting to identify which of the PHQ-9 questions can be answered directly from the user’s post and which ones would require more information from the user. Using performance analysis based on MCC scores, we show that PRIMATE is appropriate for identifying questions in PHQ-9 that could guide generative DLMs towards controlled FQ generation (with minimal hallucination) suitable for aiding triaging. The dataset created as a part of this research can be obtained from https://github.com/primate-mh/Primate2022

Recent work uses a Siamese Network, initialized with BioWordVec embeddings (distributed word embeddings), for predicting synonymy among biomedical terms to automate a part of the UMLS (Unified Medical Language System) Metathesaurus construction process. We evaluate the use of contextualized word embeddings extracted from nine different biomedical BERT-based models for synonym prediction in the UMLS by replacing BioWordVec embeddings with embeddings extracted from each biomedical BERT model using different feature extraction methods. Finally, we conduct a thorough grid search, which prior work lacks, to find the best set of hyperparameters. Surprisingly, we find that Siamese Networks initialized with BioWordVec embeddings still out perform the Siamese Networks initialized with embedding extracted from biomedical BERT model.

Existing studies on using social media for deriving mental health status of users focus on the depression detection task. However, for case management and referral to psychiatrists, health-care workers require practical and scalable depressive disorder screening and triage system. This study aims to design and evaluate a decision support system (DSS) to reliably determine the depressive triage level by capturing fine-grained depressive symptoms expressed in user tweets through the emulation of the Patient Health Questionnaire-9 (PHQ-9) that is routinely used in clinical practice. The reliable detection of depressive symptoms from tweets is challenging because the 280-character limit on tweets incentivizes the use of creative artifacts in the utterances and figurative usage contributes to effective expression. We propose a novel BERT based robust multi-task learning framework to accurately identify the depressive symptoms using the auxiliary task of figurative usage detection. Specifically, our proposed novel task sharing mechanism,co-task aware attention, enables automatic selection of optimal information across the BERT lay-ers and tasks by soft-sharing of parameters. Our results show that modeling figurative usage can demonstrably improve the model’s robustness and reliability for distinguishing the depression symptoms.

One of the cardinal tasks in achieving robust medical question answering systems is textual entailment. The existing approaches make use of an ensemble of pre-trained language models or data augmentation, often to clock higher numbers on the validation metrics. However, two major shortcomings impede higher success in identifying entailment: (1) understanding the focus/intent of the question and (2) ability to utilize the real-world background knowledge to capture the con-text beyond the sentence. In this paper, we present a novel Medical Knowledge-Enriched Textual Entailment framework that allows the model to acquire a semantic and global representation of the input medical text with the help of a relevant domain-specific knowledge graph. We evaluate our framework on the benchmark MEDIQA-RQE dataset and manifest that the use of knowledge-enriched dual-encoding mechanism help in achieving an absolute improvement of 8.27% over SOTA language models.

In recent past, social media has emerged as an active platform in the context of healthcare and medicine. In this paper, we present a study where medical user’s opinions on health-related issues are analyzed to capture the medical sentiment at a blog level. The medical sentiments can be studied in various facets such as medical condition, treatment, and medication that characterize the overall health status of the user. Considering these facets, we treat analysis of this information as a multi-task classification problem. In this paper, we adopt a novel adversarial learning approach for our multi-task learning framework to learn the sentiment’s strengths expressed in a medical blog. Our evaluation shows promising results for our target tasks.

This work addresses challenges arising from extracting entities from textual data, including the high cost of data annotation, model accuracy, selecting appropriate evaluation criteria, and the overall quality of annotation. We present a framework that integrates Entity Set Expansion (ESE) and Active Learning (AL) to reduce the annotation cost of sparse data and provide an online evaluation method as feedback. This incremental and interactive learning framework allows for rapid annotation and subsequent extraction of sparse data while maintaining high accuracy. We evaluate our framework on three publicly available datasets and show that it drastically reduces the cost of sparse entity annotation by an average of 85% and 45% to reach 0.9 and 1.0 F-Scores respectively. Moreover, the method exhibited robust performance across all datasets.

Extracting location names from informal and unstructured social media data requires the identification of referent boundaries and partitioning compound names. Variability, particularly systematic variability in location names (Carroll, 1983), challenges the identification task. Some of this variability can be anticipated as operations within a statistical language model, in this case drawn from gazetteers such as OpenStreetMap (OSM), Geonames, and DBpedia. This permits evaluation of an observed n-gram in Twitter targeted text as a legitimate location name variant from the same location-context. Using n-gram statistics and location-related dictionaries, our Location Name Extraction tool (LNEx) handles abbreviations and automatically filters and augments the location names in gazetteers (handling name contractions and auxiliary contents) to help detect the boundaries of multi-word location names and thereby delimit them in texts. We evaluated our approach on 4,500 event-specific tweets from three targeted streams to compare the performance of LNEx against that of ten state-of-the-art taggers that rely on standard semantic, syntactic and/or orthographic features. LNEx improved the average F-Score by 33-179%, outperforming all taggers. Further, LNEx is capable of stream processing.