Hemant Purohit

Hossein Salemi, Jitin Krishnan, Hemant Purohit

Attribution theory explains how individuals interpret and attribute others' behavior in a social context by employing personal (dispositional) and impersonal (situational) causality. Large Language Models (LLMs), trained on human-generated corpora, may implicitly mimic this social attribution process in social contexts. However, the extent to which LLMs utilize these causal attributions in their reasoning remains underexplored. Although using reasoning paradigms, such as Chain-of-Thought (CoT), has shown promising results in various tasks, ignoring social attribution in reasoning could lead to biased responses by LLMs in social contexts. In this study, we investigate the impact of incorporating a user's goal as knowledge to infer dispositional causality and message context to infer situational causality on LLM performance. To this end, we introduce a scalable method to mitigate such biases by enriching the instruction prompts for LLMs with two prompt aids using social-attribution knowledge, based on the context and goal of a social media message. This method improves the model performance while reducing the social-attribution bias of the LLM in the reasoning on zero-shot classification tasks for behavior analytics applications. We empirically show the benefits of our method across two tasks-intent detection and theme detection on social media in the disaster domain-when considering the variability of disaster types and multiple languages of social media. Our experiments highlight the biases of three open-source LLMs: Llama3, Mistral, and Gemma, toward social attribution, and show the effectiveness of our mitigation strategies.

Nitin Choudhury, Nikhil Kumar, Aditya Kumar Sinha et al.

Wide exploration on robocall surveillance research is hindered due to limited access to public datasets, due to privacy concerns. In this work, we first curate Robo-SAr, a synthetic robocall dataset designed for robocall surveillance research. Robo-SAr comprises of ~200 unwanted and ~1200 legitimate synthetic robocall samples across three realistic adversarial axes: psycholinguistics-manipulated transcripts, emotion-eliciting speech, and cloned voices. We further propose RoboKA, a Kolmogorov-Arnold Network (KAN)-based multimodal fusion framework designed to model structured nonlinear interactions between acoustic and linguistic cues that characterize diverse adversarial robocall strategies. RoboKA first leverages cross-modal contrastive learning to align latent modality representations and feeds the resulting embeddings to a KAN-projection head for final classification. We benchmark RoboKA against strong unimodal and multimodal baselines in both in-domain and out-of-domain setups, finding RoboKA to surpass all baselines in terms of recall and F1-score.

Siqi Lu, Mirsaleh Bahavarnia, Hiba Baroud et al.

Probabilistic search algorithms, such as Monte Carlo Tree Search (MCTS), have proven very effective in solving sequential decision-making tasks under uncertainty. However, interpreting asymmetric search trees that incorporate bandit-based tree traversal and simulation-based value estimation is difficult for end users based solely on raw tree statistics. While prior work requires hand-crafted formal logic constraints that must be updated when the problem changes, we present a framework that enables large language models (LLMs) to generate evidence-grounded explanations of MCTS decisions from recorded search traces in an end-to-end manner. Our framework maps natural-language questions to a structured set of intent categories, determines whether the existing tree contains sufficient evidence, triggers targeted expansion when needed, and generates explanations using tree statistics such as visit counts, value estimates, and risk information. Experimental results provide the first evidence that LLMs can serve as end-to-end explainers for probabilistic search, without requiring intermediate formal representations.

Chen-Wei Chang, Shailik Sarkar, Hossein Salemi et al.

Scam detection remains a critical challenge in cybersecurity as adversaries craft messages that evade automated filters. We propose a Hierarchical Scam Detection System (HSDS) that combines a lightweight multi-model voting front end with a fine-tuned LLaMA 3.1 8B Instruct back end to improve accuracy and robustness against adversarial attacks. An ensemble of four classifiers provides preliminary predictions through majority vote, and ambiguous cases are escalated to the fine-tuned model, which is optimized with adversarial training to reduce misclassification. Experiments show that this hierarchical design both improves adversarial scam detection and shortens inference time by routing most cases away from the LLM, outperforming traditional machine-learning baselines and proprietary LLM baselines. The findings highlight the effectiveness of a hybrid voting mechanism and adversarial fine-tuning in fortifying LLMs against evolving scam tactics, enhancing the resilience of automated scam detection systems.

Chen-Wei Chang, Shailik Sarkar, Shutonu Mitra et al.

Can we trust Large Language Models (LLMs) to accurately predict scam? This paper investigates the vulnerabilities of LLMs when facing adversarial scam messages for the task of scam detection. We addressed this issue by creating a comprehensive dataset with fine-grained labels of scam messages, including both original and adversarial scam messages. The dataset extended traditional binary classes for the scam detection task into more nuanced scam types. Our analysis showed how adversarial examples took advantage of vulnerabilities of a LLM, leading to high misclassification rate. We evaluated the performance of LLMs on these adversarial scam messages and proposed strategies to improve their robustness.

Yasas Senarath, Hemant Purohit

User behavior on online platforms is evolving, reflecting real-world changes in how people post, whether it's helpful messages or hate speech. Models that learn to capture this content can experience a decrease in performance over time due to data drift, which can lead to ineffective behavioral analytics systems. However, fine-tuning such a model over time with new data can be detrimental due to catastrophic forgetting. Replay-based approaches in continual learning offer a simple yet efficient method to update such models, minimizing forgetting by maintaining a buffer of important training instances from past learned tasks. However, the main limitation of this approach is the fixed size of the buffer. External knowledge bases can be utilized to overcome this limitation through data augmentation. We propose a novel augmentation-based approach to incorporate external knowledge in the replay-based continual learning framework. We evaluate several strategies with three datasets from prior studies related to deviant behavior classification to assess the integration of external knowledge in continual learning and demonstrate that augmentation helps outperform baseline replay-based approaches.

Rahul Pandey, Ziwei Zhu, Hemant Purohit

Effective labeled data collection plays a critical role in developing and fine-tuning robust streaming analytics systems. However, continuously labeling documents to filter relevant information poses significant challenges like limited labeling budget or lack of high-quality labels. There is a need for efficient human-in-the-loop machine learning (HITL-ML) design to improve streaming analytics systems. One particular HITL- ML approach is online active learning, which involves iteratively selecting a small set of the most informative documents for labeling to enhance the ML model performance. The performance of such algorithms can get affected due to human errors in labeling. To address these challenges, we propose ORIS, a method to perform Online active learning using Reinforcement learning-based Inclusive Sampling of documents for labeling. ORIS aims to create a novel Deep Q-Network-based strategy to sample incoming documents that minimize human errors in labeling and enhance the ML model performance. We evaluate the ORIS method on emotion recognition tasks, and it outperforms traditional baselines in terms of both human labeling performance and the ML model performance.

Geoffrey Pettet, Hunter Baxter, Sayyed Mohsen Vazirizade et al.

Designing effective emergency response management (ERM) systems to respond to incidents such as road accidents is a major problem faced by communities. In addition to responding to frequent incidents each day (about 240 million emergency medical services calls and over 5 million road accidents in the US each year), these systems also support response during natural hazards. Recently, there has been a consistent interest in building decision support and optimization tools that can help emergency responders provide more efficient and effective response. This includes a number of principled subsystems that implement early incident detection, incident likelihood forecasting and strategic resource allocation and dispatch policies. In this paper, we highlight the key challenges and provide an overview of the approach developed by our team in collaboration with our community partners.

Yasas Senarath, Ayan Mukhopadhyay, Sayyed Mohsen Vazirizade et al.

Emergency response is highly dependent on the time of incident reporting. Unfortunately, the traditional approach to receiving incident reports (e.g., calling 911 in the USA) has time delays. Crowdsourcing platforms such as Waze provide an opportunity for early identification of incidents. However, detecting incidents from crowdsourced data streams is difficult due to the challenges of noise and uncertainty associated with such data. Further, simply optimizing over detection accuracy can compromise spatial-temporal localization of the inference, thereby making such approaches infeasible for real-world deployment. This paper presents a novel problem formulation and solution approach for practitioner-centered incident detection using crowdsourced data by using emergency response management as a case-study. The proposed approach CROME (Crowdsourced Multi-objective Event Detection) quantifies the relationship between the performance metrics of incident classification (e.g., F1 score) and the requirements of model practitioners (e.g., 1 km. radius for incident detection). First, we show how crowdsourced reports, ground-truth historical data, and other relevant determinants such as traffic and weather can be used together in a Convolutional Neural Network (CNN) architecture for early detection of emergency incidents. Then, we use a Pareto optimization-based approach to optimize the output of the CNN in tandem with practitioner-centric parameters to balance detection accuracy and spatial-temporal localization. Finally, we demonstrate the applicability of this approach using crowdsourced data from Waze and traffic accident reports from Nashville, TN, USA. Our experiments demonstrate that the proposed approach outperforms existing approaches in incident detection while simultaneously optimizing the needs for real-world deployment and usability.

Jitin Krishnan, Antonios Anastasopoulos, Hemant Purohit et al.

Transliteration is very common on social media, but transliterated text is not adequately handled by modern neural models for various NLP tasks. In this work, we combine data augmentation approaches with a Teacher-Student training scheme to address this issue in a cross-lingual transfer setting for fine-tuning state-of-the-art pre-trained multilingual language models such as mBERT and XLM-R. We evaluate our method on transliterated Hindi and Malayalam, also introducing new datasets for benchmarking on real-world scenarios: one on sentiment classification in transliterated Malayalam, and another on crisis tweet classification in transliterated Hindi and Malayalam (related to the 2013 North India and 2018 Kerala floods). Our method yielded an average improvement of +5.6% on mBERT and +4.7% on XLM-R in F1 scores over their strong baselines.

Jitin Krishnan, Antonios Anastasopoulos, Hemant Purohit et al.

Predicting user intent and detecting the corresponding slots from text are two key problems in Natural Language Understanding (NLU). In the context of zero-shot learning, this task is typically approached by either using representations from pre-trained multilingual transformers such as mBERT, or by machine translating the source data into the known target language and then fine-tuning. Our work focuses on a particular scenario where the target language is unknown during training. To this goal, we propose a novel method to augment the monolingual source data using multilingual code-switching via random translations to enhance a transformer's language neutrality when fine-tuning it for a downstream task. This method also helps discover novel insights on how code-switching with different language families around the world impact the performance on the target language. Experiments on the benchmark dataset of MultiATIS++ yielded an average improvement of +4.2% in accuracy for intent task and +1.8% in F1 for slot task using our method over the state-of-the-art across 8 different languages. Furthermore, we present an application of our method for crisis informatics using a new human-annotated tweet dataset of slot filling in English and Haitian Creole, collected during Haiti earthquake disaster.

Yasas Senarath, Saideep Nannapaneni, Hemant Purohit et al.

The number of emergencies have increased over the years with the growth in urbanization. This pattern has overwhelmed the emergency services with limited resources and demands the optimization of response processes. It is partly due to traditional `reactive' approach of emergency services to collect data about incidents, where a source initiates a call to the emergency number (e.g., 911 in U.S.), delaying and limiting the potentially optimal response. Crowdsourcing platforms such as Waze provides an opportunity to develop a rapid, `proactive' approach to collect data about incidents through crowd-generated observational reports. However, the reliability of reporting sources and spatio-temporal uncertainty of the reported incidents challenge the design of such a proactive approach. Thus, this paper presents a novel method for emergency incident detection using noisy crowdsourced Waze data. We propose a principled computational framework based on Bayesian theory to model the uncertainty in the reliability of crowd-generated reports and their integration across space and time to detect incidents. Extensive experiments using data collected from Waze and the official reported incidents in Nashville, Tenessee in the U.S. show our method can outperform strong baselines for both F1-score and AUC. The application of this work provides an extensible framework to incorporate different noisy data sources for proactive incident detection to improve and optimize emergency response operations in our communities.

Rahul Pandey, Hemant Purohit, Carlos Castillo et al.

High-quality human annotations are necessary for creating effective machine learning-driven stream processing systems. We study hybrid stream processing systems based on a Human-In-The-Loop Machine Learning (HITL-ML) paradigm, in which one or many human annotators and an automatic classifier (trained at least partially by the human annotators) label an incoming stream of instances. This is typical of many near-real-time social media analytics and web applications, including annotating social media posts during emergencies by digital volunteer groups. From a practical perspective, low-quality human annotations result in wrong labels for retraining automated classifiers and indirectly contribute to the creation of inaccurate classifiers. Considering human annotation as a psychological process allows us to address these limitations. We show that human annotation quality is dependent on the ordering of instances shown to annotators and can be improved by local changes in the instance sequence/order provided to the annotators, yielding a more accurate annotation of the stream. We adapt a theoretically-motivated human error framework of mistakes and slips for the human annotation task to study the effect of ordering instances (i.e., an "annotation schedule"). Further, we propose an error-avoidance approach to the active learning paradigm for stream processing applications robust to these likely human errors (in the form of slips) when deciding a human annotation schedule. We support the human error framework using crowdsourcing experiments and evaluate the proposed algorithm against standard baselines for active learning via extensive experimentation on classification tasks of filtering relevant social media posts during natural disasters.

Jitin Krishnan, Patrick Coronado, Hemant Purohit et al.

We build a common-knowledge concept recognition system for a Systems Engineer's Virtual Assistant (SEVA) which can be used for downstream tasks such as relation extraction, knowledge graph construction, and question-answering. The problem is formulated as a token classification task similar to named entity extraction. With the help of a domain expert and text processing methods, we construct a dataset annotated at the word-level by carefully defining a labelling scheme to train a sequence model to recognize systems engineering concepts. We use a pre-trained language model and fine-tune it with the labeled dataset of concepts. In addition, we also create some essential datasets for information such as abbreviations and definitions from the systems engineering domain. Finally, we construct a simple knowledge graph using these extracted concepts along with some hyponym relations.

Jitin Krishnan, Hemant Purohit, Huzefa Rangwala

During the onset of a disaster event, filtering relevant information from the social web data is challenging due to its sparse availability and practical limitations in labeling datasets of an ongoing crisis. In this paper, we hypothesize that unsupervised domain adaptation through multi-task learning can be a useful framework to leverage data from past crisis events for training efficient information filtering models during the sudden onset of a new crisis. We present a novel method to classify relevant tweets during an ongoing crisis without seeing any new examples, using the publicly available dataset of TREC incident streams. Specifically, we construct a customized multi-task architecture with a multi-domain discriminator for crisis analytics: multi-task domain adversarial attention network. This model consists of dedicated attention layers for each task to provide model interpretability; critical for real-word applications. As deep networks struggle with sparse datasets, we show that this can be improved by sharing a base layer for multi-task learning and domain adversarial training. Evaluation of domain adaptation for crisis events is performed by choosing a target event as the test set and training on the rest. Our results show that the multi-task model outperformed its single task counterpart. For the qualitative evaluation of interpretability, we show that the attention layer can be used as a guide to explain the model predictions and empower emergency services for exploring accountability of the model, by showcasing the words in a tweet that are deemed important in the classification process. Finally, we show a practical implication of our work by providing a use-case for the COVID-19 pandemic.

Jitin Krishnan, Hemant Purohit, Huzefa Rangwala

Domain adaptation approaches seek to learn from a source domain and generalize it to an unseen target domain. At present, the state-of-the-art unsupervised domain adaptation approaches for subjective text classification problems leverage unlabeled target data along with labeled source data. In this paper, we propose a novel method for domain adaptation of single-task text classification problems based on a simple but effective idea of diversity-based generalization that does not require unlabeled target data but still matches the state-of-the-art in performance. Diversity plays the role of promoting the model to better generalize and be indiscriminate towards domain shift by forcing the model not to rely on same features for prediction. We apply this concept on the most explainable component of neural networks, the attention layer. To generate sufficient diversity, we create a multi-head attention model and infuse a diversity constraint between the attention heads such that each head will learn differently. We further expand upon our model by tri-training and designing a procedure with an additional diversity constraint between the attention heads of the tri-trained classifiers. Extensive evaluation using the standard benchmark dataset of Amazon reviews and a newly constructed dataset of Crisis events shows that our fully unsupervised method matches with the competing baselines that uses unlabeled target data. Our results demonstrate that machine learning architectures that ensure sufficient diversity can generalize better; encouraging future research to design ubiquitously usable learning models without using unlabeled target data.

Xiaojie Guo, Amir Alipour-Fanid, Lingfei Wu et al.

At present, object recognition studies are mostly conducted in a closed lab setting with classes in test phase typically in training phase. However, real-world problem is far more challenging because: i) new classes unseen in the training phase can appear when predicting; ii) discriminative features need to evolve when new classes emerge in real time; and iii) instances in new classes may not follow the "independent and identically distributed" (iid) assumption. Most existing work only aims to detect the unknown classes and is incapable of continuing to learn newer classes. Although a few methods consider both detecting and including new classes, all are based on the predefined handcrafted features that cannot evolve and are out-of-date for characterizing emerging classes. Thus, to address the above challenges, we propose a novel generic end-to-end framework consisting of a dynamic cascade of classifiers that incrementally learn their dynamic and inherent features. The proposed method injects dynamic elements into the system by detecting instances from unknown classes, while at the same time incrementally updating the model to include the new classes. The resulting cascade tree grows by adding a new leaf node classifier once a new class is detected, and the discriminative features are updated via an end-to-end learning strategy. Experiments on two real-world datasets demonstrate that our proposed method outperforms existing state-of-the-art methods.

Rahul Pandey, Carlos Castillo, Hemant Purohit

High-quality human annotations are necessary to create effective machine learning systems for social media. Low-quality human annotations indirectly contribute to the creation of inaccurate or biased learning systems. We show that human annotation quality is dependent on the ordering of instances shown to annotators (referred as 'annotation schedule'), and can be improved by local changes in the instance ordering provided to the annotators, yielding a more accurate annotation of the data stream for efficient real-time social media analytics. We propose an error-mitigating active learning algorithm that is robust with respect to some cases of human errors when deciding an annotation schedule. We validate the human error model and evaluate the proposed algorithm against strong baselines by experimenting on classification tasks of relevant social media posts during crises. According to these experiments, considering the order in which data instances are presented to human annotators leads to both an increase in accuracy for machine learning and awareness toward some potential biases in human learning that may affect the automated classifier.

Habib Karbasian, Hemant Purohit, Rajat Handa et al.

Social media provides a mechanism for people to engage with social causes across a range of issues. It also provides a strategic tool to those looking to advance a cause to exchange, promote or publicize their ideas. In such instances, AI can be either an asset if used appropriately or a barrier. One of the key issues for a workforce diversity campaign is to understand in real-time who is participating - specifically, whether the participants are individuals or organizations, and in case of individuals, whether they are male or female. In this paper, we present a study to demonstrate a case for AI for social good that develops a model to infer in real-time the different user types participating in a cause-driven hashtag campaign on Twitter, ILookLikeAnEngineer (ILLAE). A generic framework is devised to classify a Twitter user into three classes: organization, male and female in a real-time manner. The framework is tested against two datasets (ILLAE and a general dataset) and outperforms the baseline binary classifiers for categorizing organization/individual and male/female. The proposed model can be applied to future social cause-driven campaigns to get real-time insights on the macro-level social behavior of participants.