Siddharth Krishnan

Avisha Kumar, Kunal Kotkar, Kelly Jiang et al.

While deep learning has catalyzed breakthroughs across numerous domains, its broader adoption in clinical settings is inhibited by the costly and time-intensive nature of data acquisition and annotation. To further facilitate medical machine learning, we present an ultrasound dataset of 10,223 Brightness-mode (B-mode) images consisting of sagittal slices of porcine spinal cords (N=25) before and after a contusion injury. We additionally benchmark the performance metrics of several state-of-the-art object detection algorithms to localize the site of injury and semantic segmentation models to label the anatomy for comparison and creation of task-specific architectures. Finally, we evaluate the zero-shot generalization capabilities of the segmentation models on human ultrasound spinal cord images to determine whether training on our porcine dataset is sufficient for accurately interpreting human data. Our results show that the YOLOv8 detection model outperforms all evaluated models for injury localization, achieving a mean Average Precision (mAP50-95) score of 0.606. Segmentation metrics indicate that the DeepLabv3 segmentation model achieves the highest accuracy on unseen porcine anatomy, with a Mean Dice score of 0.587, while SAMed achieves the highest Mean Dice score generalizing to human anatomy (0.445). To the best of our knowledge, this is the largest annotated dataset of spinal cord ultrasound images made publicly available to researchers and medical professionals, as well as the first public report of object detection and segmentation architectures to assess anatomical markers in the spinal cord for methodology development and clinical applications.

Rakeen Rouf, Trupti Bavalatti, Osama Ahmed et al.

Novel research aimed at text-to-image (T2I) generative AI safety often relies on publicly available datasets for training and evaluation, making the quality and composition of these datasets crucial. This paper presents a comprehensive review of the key datasets used in the T2I research, detailing their collection methods, compositions, semantic and syntactic diversity of prompts and the quality, coverage, and distribution of harm types in the datasets. By highlighting the strengths and limitations of the datasets, this study enables researchers to find the most relevant datasets for a use case, critically assess the downstream impacts of their work given the dataset distribution, particularly regarding model safety and ethical considerations, and also identify the gaps in dataset coverage and quality that future research may address.

Joshua Melton, Shannon Reid, Gabriel Terejanu et al.

The high volume and rapid evolution of content on social media present major challenges for studying the stance of social media users. In this work, we develop a two stage stance labeling method that utilizes the user-hashtag bipartite graph and the user-user interaction graph. In the first stage, a simple and efficient heuristic for stance labeling uses the user-hashtag bipartite graph to iteratively update the stance association of user and hashtag nodes via a label propagation mechanism. This set of soft labels is then integrated with the user-user interaction graph to train a graph neural network (GNN) model using semi-supervised learning. We evaluate this method on two large-scale datasets containing tweets related to climate change from June 2021 to June 2022 and gun control from January 2022 to January 2023. Our experiments demonstrate that enriching text-based embeddings of users with network information from the user interaction graph using our semi-supervised GNN method outperforms both classifiers trained on user textual embeddings and zero-shot classification using LLMs such as GPT4. We discuss the need for integrating nuanced understanding from social science with the scalability of computational methods to better understand how polarization on social media occurs for divisive issues such as climate change and gun control.

Joshua Melton, Arunkumar Bagavathi, Siddharth Krishnan

Online hate speech on social media has become a fast-growing problem in recent times. Nefarious groups have developed large content delivery networks across several main-stream (Twitter and Facebook) and fringe (Gab, 4chan, 8chan, etc.) outlets to deliver cascades of hate messages directed both at individuals and communities. Thus addressing these issues has become a top priority for large-scale social media outlets. Three key challenges in automated detection and classification of hateful content are the lack of clearly labeled data, evolving vocabulary and lexicon - hashtags, emojis, etc. - and the lack of baseline models for fringe outlets such as Gab. In this work, we propose a novel framework with three major contributions. (a) We engineer an ensemble of deep learning models that combines the strengths of state-of-the-art approaches, (b) we incorporate a tuning factor into this framework that leverages transfer learning to conduct automated hate speech classification on unlabeled datasets, like Gab, and (c) we develop a weak supervised learning methodology that allows our framework to train on unlabeled data. Our ensemble models achieve an 83% hate recall on the HON dataset, surpassing the performance of the state-of-the-art deep models. We demonstrate that weak supervised training in combination with classifier tuning significantly increases model performance on unlabeled data from Gab, achieving a hate recall of 67%.

Arunkumar Bagavathi, Siddharth Krishnan, Sanjay Subrahmanyan et al.

South Indian classical music (Carnatic music) is best consumed through live concerts. A carnatic recital requires meticulous planning accounting for several parameters like the performers' repertoire, composition variety, musical versatility, thematic structure, the recital's arrangement, etc. to ensure that the audience have a comprehensive listening experience. In this work, we present ragamAI a novel machine learning framework that utilizes the tonic nuances and musical structures in the carnatic music to generate a concert recital that melodically captures the entire range in an octave. Utilizing the underlying idea of playlist and session-based recommender models, the proposed model studies the mathematical structure present in past concerts and recommends relevant items for the playlist/concert. ragamAI ensembles recommendations given by multiple models to learn user idea and past preference of sequences in concerts to extract recommendations. Our experiments on a vast collection of concert show that our model performs 25%-50% better than baseline models. ragamAI's applications are two-fold. 1) it will assist musicians to customize their performance with the necessary variety required to sustain the interest of the audience for the entirety of the concert 2) it will generate carefully curated lists of south Indian classical music so that the listener can discover the wide range of melody that the musical system can offer.

Arunkumar Bagavathi, Pedram Bashiri, Shannon Reid et al.

Online social media, periodically serves as a platform for cascading polarizing topics of conversation. The inherent community structure present in online social networks (homophily) and the advent of fringe outlets like Gab have created online "echo chambers" that amplify the effects of polarization, which fuels detrimental behavior. Recently, in October 2018, Gab made headlines when it was revealed that Robert Bowers, the individual behind the Pittsburgh Synagogue massacre, was an active member of this social media site and used it to express his anti-Semitic views and discuss conspiracy theories. Thus to address the need of automated data-driven analyses of such fringe outlets, this research proposes novel methods to discover topics that are prevalent in Gab and how they cascade within the network. Specifically, using approximately 34 million posts, and 3.7 million cascading conversation threads with close to 300k users; we demonstrate that there are essentially five cascading patterns that manifest in Gab and the most "viral" ones begin with an echo-chamber pattern and grow out to the entire network. Also, we empirically show, through two models viz. Susceptible-Infected and Bass, how the cascades structurally evolve from one of the five patterns to the other based on the topic of the conversation with upto 84% accuracy.