Maneesh Bilalpur

Arushi Sharma, Abhibha Gupta, Maneesh Bilalpur

To advance argumentative stance prediction as a multimodal problem, the First Shared Task in Multimodal Argument Mining hosted stance prediction in crucial social topics of gun control and abortion. Our exploratory study attempts to evaluate the necessity of images for stance prediction in tweets and compare out-of-the-box text-based large-language models (LLM) in few-shot settings against fine-tuned unimodal and multimodal models. Our work suggests an ensemble of fine-tuned text-based language models (0.817 F1-score) outperforms both the multimodal (0.677 F1-score) and text-based few-shot prediction using a recent state-of-the-art LLM (0.550 F1-score). In addition to the differences in performance, our findings suggest that the multimodal models tend to perform better when image content is summarized as natural language over their native pixel structure and, using in-context examples improves few-shot performance of LLMs.

Maneesh Bilalpur, Megan Hamm, Young Ji Lee et al.

Storytelling is a powerful form of communication and may provide insights into factors contributing to gaps in healthcare outcomes. To determine whether Large Language Models (LLMs) can identify potential underlying factors and avenues for intervention, we performed topic-aware hierarchical summarization of narratives from African American (AA) storytellers. Fifty transcribed stories of AA experiences were used to identify topics in their experience using the Latent Dirichlet Allocation (LDA) technique. Stories about a given topic were summarized using an open-source LLM-based hierarchical summarization approach. Topic summaries were generated by summarizing across story summaries for each story that addressed a given topic. Generated topic summaries were rated for fabrication, accuracy, comprehensiveness, and usefulness by the GPT4 model, and the model's reliability was validated against the original story summaries by two domain experts. 26 topics were identified in the fifty AA stories. The GPT4 ratings suggest that topic summaries were free from fabrication, highly accurate, comprehensive, and useful. The reliability of GPT ratings compared to expert assessments showed moderate to high agreement. Our approach identified AA experience-relevant topics such as health behaviors, interactions with medical team members, caregiving and symptom management, among others. Such insights could help researchers identify potential factors and interventions by learning from unstructured narratives in an efficient manner-leveraging the communicative power of storytelling. The use of LDA and LLMs to identify and summarize the experience of AA individuals suggests a variety of possible avenues for health research and possible clinical improvements to support patients and caregivers, thereby ultimately improving health outcomes.

Maneesh Bilalpur, Saurabh Hinduja, Sonish Sivarajkumar et al.

The classical approach to detecting depression from vision emphasizes interpretable features, such as facial expression, and classifiers such as the Support Vector Machine (SVM). With the advent of deep learning, there has been a shift in feature representations and classification approaches. Contemporary approaches use learnt features from general-purpose vision models such as VGGNet to train machine learning models. Little is known about how classical and deep approaches compare in depression detection with respect to accuracy, fairness, and generalizability, especially across contexts. To address these questions, we compared classical and deep approaches to the detection of depression in the visual modality in two different contexts: Mother-child interactions in the TPOT database and patient-clinician interviews in the Pitt database. In the former, depression was operationalized as a history of depression per the DSM and current or recent clinically significant symptoms. In the latter, all participants met initial criteria for depression per DSM, and depression was reassessed over the course of treatment. The classical approach included handcrafted features with SVM classifiers. Learnt features were turn-level embeddings from the FMAE-IAT that were combined with Multi-Layer Perceptron classifiers. The classical approach achieved higher accuracy in both contexts. It was also significantly fairer than the deep approach in the patient-clinician context. Cross-context generalizability was modest at best for both approaches, which suggests that depression may be context-specific.

Saurabh Hinduja, Gurmeet Kaur, Maneesh Bilalpur et al.

Subject-exclusive cross-validation is the standard evaluation protocol for facial Action Unit (AU) detection, yet reported improvements are often small. We show that cross-validation itself introduces measurable stochastic variance. On BP4D+, repeated 3-fold subject-exclusive splits produce an empirical noise floor of $\pm 0.065$ in average F1, with substantially larger variation for low-prevalence AUs. Operating-point metrics such as F1 fluctuate more than threshold-independent measures such as AUC, and model ranking can change under different fold assignments. We further evaluate cross-dataset robustness using a Leave-One-Dataset-Out (LODO) protocol across five AU datasets. LODO removes partition randomness and exposes domain-level instability that is not visible under single-dataset cross-validation. Together, these results suggest that gains often reported in cross-fold validation may fall within protocol variance. Leave-one-dataset-out cross-validation yields more stable and interpretable findings

Maneesh Bilalpur, Mert Inan, Dorsa Zeinali et al.

Addressing the critical shortage of mental health resources for effective screening, diagnosis, and treatment remains a significant challenge. This scarcity underscores the need for innovative solutions, particularly in enhancing the accessibility and efficacy of therapeutic support. Embodied agents with advanced interactive capabilities emerge as a promising and cost-effective supplement to traditional caregiving methods. Crucial to these agents' effectiveness is their ability to simulate non-verbal behaviors, like backchannels, that are pivotal in establishing rapport and understanding in therapeutic contexts but remain under-explored. To improve the rapport-building capabilities of embodied agents we annotated backchannel smiles in videos of intimate face-to-face conversations over topics such as mental health, illness, and relationships. We hypothesized that both speaker and listener behaviors affect the duration and intensity of backchannel smiles. Using cues from speech prosody and language along with the demographics of the speaker and listener, we found them to contain significant predictors of the intensity of backchannel smiles. Based on our findings, we introduce backchannel smile production in embodied agents as a generation problem. Our attention-based generative model suggests that listener information offers performance improvements over the baseline speaker-centric generation approach. Conditioned generation using the significant predictors of smile intensity provides statistically significant improvements in empirical measures of generation quality. Our user study by transferring generated smiles to an embodied agent suggests that agent with backchannel smiles is perceived to be more human-like and is an attractive alternative for non-personal conversations over agent without backchannel smiles.

Sonish Sivarajkumar, Hang Zhang, Yuelyu Ji et al.

Electronic health records (EHRs) are rich clinical data sources but complex repositories of patient data, spanning structured elements (demographics, vitals, lab results, codes), unstructured clinical notes and other modalities of data. Harnessing this heterogeneity is critical for improving patient outcomes. Recent advances in large language models (LLMs) have enabled foundation models that can learn from multiple data modalities and support clinical tasks. However, most current approaches simply serialize numeric EHR data into text, which risks losing temporal and quantitative detail. We introduce Generative Deep Patient (GDP), a multimodal foundation model that natively encodes structured EHR time-series via a CNN-Transformer encoder and fuses it with unstructured EHRs through cross-modal attention into a LLaMA-based decoder. GDP is trained in two stages: (1) generative pretraining, where it learns to produce clinical narratives from raw patient timelines while also performing masked feature prediction (MFP) and next time-step prediction (NTP) to capture temporal dynamics; and (2) multi-task fine-tuning for clinically meaningful predictions (e.g., heart failure, type 2 diabetes, 30-day readmission). In clinical prediction, GDP demonstrated superior performance on MIMIC-IV: heart failure AUROC = 0.923, type 2 diabetes AUROC = 0.817, and 30-day readmission AUROC = 0.627. For narrative generation, GDP achieved ROUGE-L = 0.135 and BERTScore-F1 = 0.545. In a blinded human evaluation, GDP-Instruct scored highest on faithfulness, fluency, and overall clinical utility, suggesting reduced hospital documentation workload without sacrificing accuracy. Our results demonstrate that a single multimodal foundation model can both predict clinically actionable events and generate high-quality clinical narratives. Furthermore, GDP's flexible architecture can be extended to additional modalities.

Monika Gahalawat, Maneesh Bilalpur, Raul Fernandez Rojas et al.

Depression is a debilitating mood disorder negatively impacting millions worldwide. While researchers have explored multiple verbal and non-verbal behavioural cues for automated depression assessment, head motion has received little attention thus far. Further, the common practice of validating machine learning models via a single dataset can limit model generalisability. This work examines the effectiveness and generalisability of models utilising elementary head motion units, termed kinemes, for depression severity estimation. Specifically, we consider three depression datasets from different western cultures (German: AVEC2013, Australian: Blackdog and American: Pitt datasets) with varied contextual and recording settings to investigate the generalisability of the derived kineme patterns via two methods: (i) k-fold cross-validation over individual/multiple datasets, and (ii) model reuse on other datasets. Evaluating classification and regression performance with classical machine learning methods, our results show that: (1) head motion patterns are efficient biomarkers for estimating depression severity, achieving highly competitive performance for both classification and regression tasks on a variety of datasets, including achieving the second best Mean Absolute Error (MAE) on the AVEC2013 dataset, and (2) kineme-based features are more generalisable than (a) raw head motion descriptors for binary severity classification, and (b) other visual behavioural cues for severity estimation (regression).

Maneesh Bilalpur, Seyed Mostafa Kia, Mohan Kankanhalli et al.

This work explores the utility of implicit behavioral cues, namely, Electroencephalogram (EEG) signals and eye movements for gender recognition (GR) and emotion recognition (ER) from psychophysical behavior. Specifically, the examined cues are acquired via low-cost, off-the-shelf sensors. 28 users (14 male) recognized emotions from unoccluded (no mask) and partially occluded (eye or mouth masked) emotive faces; their EEG responses contained gender-specific differences, while their eye movements were characteristic of the perceived facial emotions. Experimental results reveal that (a) reliable GR and ER is achievable with EEG and eye features, (b) differential cognitive processing of negative emotions is observed for females and (c) eye gaze-based gender differences manifest under partial face occlusion, as typified by the eye and mouth mask conditions.

Viral Parekh, Maneesh Bilalpur, Sharavan Kumar et al.

We examine if EEG-based cognitive load (CL) estimation is generalizable across the character, spatial pattern, bar graph and pie chart-based visualizations for the nback~task. CL is estimated via two recent approaches: (a) Deep convolutional neural network, and (b) Proximal support vector machines. Experiments reveal that CL estimation suffers across visualizations motivating the need for effective machine learning techniques to benchmark visual interface usability for a given analytic task.

Maneesh Bilalpur, Mohan Kankanhalli, Stefan Winkler et al.

Data Visualization has been receiving growing attention recently, with ubiquitous smart devices designed to render information in a variety of ways. However, while evaluations of visual tools for their interpretability and intuitiveness have been commonplace, not much research has been devoted to other forms of data rendering, eg, sonification. This work is the first to automatically estimate the cognitive load induced by different acoustic parameters considered for sonification in prior studies. We examine cognitive load via (a) perceptual data-sound mapping accuracies of users for the different acoustic parameters, (b) cognitive workload impressions explicitly reported by users, and (c) their implicit EEG responses compiled during the mapping task. Our main findings are that (i) low cognitive load-inducing (ie, more intuitive) acoustic parameters correspond to higher mapping accuracies, (ii) EEG spectral power analysis reveals higher $α$ band power for low cognitive load parameters, implying a congruent relationship between explicit and implicit user responses, and (iii) Cognitive load classification with EEG features achieves a peak F1-score of 0.64, confirming that reliable workload estimation is achievable with user EEG data compiled using wearable sensors.

Maneesh Bilalpur, Seyed Mostafa Kia, Manisha Chawla et al.

We examine the utility of implicit user behavioral signals captured using low-cost, off-the-shelf devices for anonymous gender and emotion recognition. A user study designed to examine male and female sensitivity to facial emotions confirms that females recognize (especially negative) emotions quicker and more accurately than men, mirroring prior findings. Implicit viewer responses in the form of EEG brain signals and eye movements are then examined for existence of (a) emotion and gender-specific patterns from event-related potentials (ERPs) and fixation distributions and (b) emotion and gender discriminability. Experiments reveal that (i) Gender and emotion-specific differences are observable from ERPs, (ii) multiple similarities exist between explicit responses gathered from users and their implicit behavioral signals, and (iii) Significantly above-chance ($\approx$70%) gender recognition is achievable on comparing emotion-specific EEG responses-- gender differences are encoded best for anger and disgust. Also, fairly modest valence (positive vs negative emotion) recognition is achieved with EEG and eye-based features.

Maneesh Bilalpur, Seyed Mostafa Kia, Tat-Seng Chua et al.

We examine the utility of implicit behavioral cues in the form of EEG brain signals and eye movements for gender recognition (GR) and emotion recognition (ER). Specifically, the examined cues are acquired via low-cost, off-the-shelf sensors. We asked 28 viewers (14 female) to recognize emotions from unoccluded (no mask) as well as partially occluded (eye and mouth masked) emotive faces. Obtained experimental results reveal that (a) reliable GR and ER is achievable with EEG and eye features, (b) differential cognitive processing especially for negative emotions is observed for males and females and (c) some of these cognitive differences manifest under partial face occlusion, as typified by the eye and mouth mask conditions.