SRSohini Roychowdhury

Sohini Roychowdhury

*This is an algorithmically generated author record based on publicly available information.

Sohini Roychowdhury has a total of 48 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    LO-FAR: A Cost-Aware Local Filter for Sparse Feature Ranking in Industrial Ad Recommendation

    by Egemen Erbayat, Luis Duque, Sohini Roychowdhury, Mohammad Amin, Srihari Reddy

  • Open Access

    Hallucination-minimized Data-to-answer Framework for Financial Decision-makers

    by Sohini Roychowdhury, Andres Alvarez, Brian Moore, Marko Krema, Maria Paz Gelpi, Federico Martin Rodriguez, Angel Rodriguez, Jose Ramon Cabrejas, Pablo Martinez Serrano, Punit Agrawal, et al.

    Large Language Models (LLMs) have been applied to build several automation and personalized question-answering prototypes so far. However, scaling such prototypes to robust products with minimized hallucinations or fake responses still remains an open challenge, especially in niche data-table heavy domains such as financial decision making. In this work, we present a novel Langchain-based framework that transforms data tables into hierarchical textual data chunks to enable a wide variety of actionable question answering. First, the user-queries are classified by intention followed by automated retrieval of the most relevant data chunks to generate customized LLM prompts per query. Next, the custom prompts and their responses undergo multi-metric scoring to assess for hallucinations and response confidence. The proposed system is optimized with user-query intention classification, advanced prompting, data scaling capabilities and it achieves over 90% confidence scores for a variety of user-queries responses ranging from {What, Where, Why, How, predict, trend, anomalies, exceptions} that are crucial for financial decision making applications. The proposed data to answers framework can be extended to other analytical domains such as sales and payroll to ensure optimal hallucination control guardrails.

  • Open Access

    NUMSnet: Nested-U Multi-Class Segmentation Network for 3D Medical Image Stacks

    by Sohini Roychowdhury

    The semantic segmentation of 3D medical image stacks enables accurate volumetric reconstructions, computer-aided diagnostics and follow-up treatment planning. In this work, we present a novel variant of the Unet model, called the NUMSnet, that transmits pixel neighborhood features across scans through nested layers to achieve accurate multi-class semantic segmentation with minimal training data. We analyzed the semantic segmentation performance of the NUMSnet model in comparison with several Unet model variants in the segmentation of 3–7 regions of interest using only 5–10% of images for training per Lung-CT and Heart-CT volumetric image stack. The proposed NUMSnet model achieves up to 20% improvement in segmentation recall, with 2–9% improvement in Dice scores for Lung-CT stacks and 2.5–16% improvement in Dice scores for Heart-CT stacks when compared to the Unet++ model. The NUMSnet model needs to be trained with ordered images around the central scan of each volumetric stack. The propagation of image feature information from the six nested layers of the Unet++ model are found to have better computation and segmentation performance than the propagation of fewer hidden layers or all ten up-sampling layers in a Unet++ model. The NUMSnet model achieves comparable segmentation performance to previous works while being trained on as few as 5–10% of the images from 3D stacks. In addition, transfer learning allows faster convergence of the NUMSnet model for multi-class semantic segmentation from pathology in Lung-CT images to cardiac segmentation in Heart-CT stacks. Thus, the proposed model can standardize multi-class semantic segmentation for a variety of volumetric image stacks with a minimal training dataset. This can significantly reduce the cost, time and inter-observer variability associated with computer-aided detection and treatment.


  • Open Access

    NUMSnet: Nested-U Multi-class Segmentation network for 3D Medical Image Stacks

    by Sohini Roychowdhury

    Semantic segmentation for medical 3D image stacks enables accurate volumetric reconstructions, computer-aided diagnostics and follow up treatment planning. In this work, we present a novel variant of the Unet model called the NUMSnet that transmits pixel neighborhood features across scans through nested layers to achieve accurate multi-class semantic segmentations with minimal training data. We analyze the semantic segmentation performance of the NUMSnet model in comparison with several Unet model variants to segment 3-7 regions of interest using only 10% of images for training per Lung-CT and Heart-CT volumetric image stacks. The proposed NUMSnet model achieves up to 20% improvement in segmentation recall with 4-9% improvement in Dice scores for Lung-CT stacks and 2.5-10% improvement in Dice scores for Heart-CT stacks when compared to the Unet++ model. The NUMSnet model needs to be trained by ordered images around the central scan of each volumetric stack. Propagation of image feature information from the 6 nested layers of the Unet++ model are found to have better computation and segmentation performances than propagation of all up-sampling layers in a Unet++ model. The NUMSnet model achieves comparable segmentation performances to existing works, while being trained on as low as 5\% of the training images. Also, transfer learning allows faster convergence of the NUMSnet model for multi-class semantic segmentation from pathology in Lung-CT images to cardiac segmentations in Heart-CT stacks. Thus, the proposed model can standardize multi-class semantic segmentation on a variety of volumetric image stacks with minimal training dataset. This can significantly reduce the cost, time and inter-observer variabilities associated with computer-aided detections and treatment.


  • Open Access

    QU-net++: Image Quality Detection Framework for Segmentation of Medical 3D Image Stacks

    by Sohini Roychowdhury

    Automated segmentation of pathological regions of interest aids medical image diagnostics and follow-up care. However, accurate pathological segmentations require high quality of annotated data that can be both cost and time intensive to generate. In this work, we propose an automated two-step method that detects a minimal image subset required to train segmentation models by evaluating the quality of medical images from 3D image stacks using a U-net++ model. These images that represent a lack of quality training can then be annotated and used to fully train a U-net-based segmentation model. The proposed QU-net++ model detects this lack of quality training based on the disagreement in segmentations produced from the final two output layers. The proposed model isolates around 10% of the slices per 3D image stack and can scale across imaging modalities to segment cysts in OCT images and ground glass opacity (GGO) in lung CT images with Dice scores in the range 0.56-0.72. Thus, the proposed method can be applied for cost effective multi-modal pathology segmentation tasks.

  • Open Access

    OPAM: Online Purchasing-behavior Analysis using Machine learning

    by Sohini Roychowdhury, Ebrahim Alareqi, Wenxi Li

    Customer purchasing behavior analysis plays a key role in developing insightful communication strategies between online vendors and their customers. To support the recent increase in online shopping trends, in this work, we present a customer purchasing behavior analysis system using supervised, unsupervised and semi-supervised learning methods. The proposed system analyzes session and user-journey level purchasing behaviors to identify customer categories/clusters that can be useful for targeted consumer insights at scale. We observe higher sensitivity to the design of online shopping portals for session-level purchasing prediction with accuracy/recall in range 91-98%/73-99%, respectively. The user-journey level analysis demonstrates five unique user clusters, wherein New Shoppers are most predictable and Impulsive Shoppers are most unique with low viewing and high carting behaviors for purchases. Further, cluster transformation metrics and partial label learning demonstrates the robustness of each user cluster to new/unlabelled events. Thus, customer clusters can aid strategic targeted nudge models.


  • Open Access

    SISE-PC: Semi-supervised Image Subsampling for Explainable Pathology

    by Sohini Roychowdhury, Kwok Sun Tang, Mohith Ashok, Anoop Sanka

    Although automated pathology classification using deep learning (DL) has proved to be predictively efficient, DL methods are found to be data and compute cost intensive. In this work, we aim to reduce DL training costs by pre-training a Resnet feature extractor using SimCLR contrastive loss for latent encoding of OCT images. We propose a novel active learning framework that identifies a minimal sub-sampled dataset containing the most uncertain OCT image samples using label propagation on the SimCLR latent encodings. The pre-trained Resnet model is then fine-tuned with the labelled minimal sub-sampled data and the underlying pathological sites are visually explained. Our framework identifies upto 2% of OCT images to be most uncertain that need prioritized specialist attention and that can fine-tune a Resnet model to achieve upto 97% classification accuracy. The proposed method can be extended to other medical images to minimize prediction costs.

  • Open Access

    Automated Augmentation with Reinforcement Learning and GANs for Robust Identification of Traffic Signs using Front Camera Images

    by Sohini Roy Chowdhury, Lars Tornberg, Robin Halvfordsson, Jonatan Nordh, Adam Suhren Gustafsson, Joel Wall, Mattias Westerberg, Adam Wirehed, Louis Tilloy, Zhanying Hu, et al.

    Traffic sign identification using camera images from vehicles plays a critical role in autonomous driving and path planning. However, the front camera images can be distorted due to blurriness, lighting variations and vandalism which can lead to degradation of detection performances. As a solution, machine learning models must be trained with data from multiple domains, and collecting and labeling more data in each new domain is time consuming and expensive. In this work, we present an end-to-end framework to augment traffic sign training data using optimal reinforcement learning policies and a variety of Generative Adversarial Network (GAN) models, that can then be used to train traffic sign detector modules. Our automated augmenter enables learning from transformed nightime, poor lighting, and varying degrees of occlusions using the LISA Traffic Sign and BDD-Nexar dataset. The proposed method enables mapping training data from one domain to another, thereby improving traffic sign detection precision/recall from 0.70/0.66 to 0.83/0.71 for nighttime images.

  • Segmentation of Intra-Retinal Cysts From Optical Coherence Tomography Images Using a Fully Convolutional Neural Network Model

    by G. N. Girish, Bibhash Thakur, Sohini Roy Chowdhury, Abhishek R. Kothari, Jeny Rajan

    Optical Coherence Tomography (OCT) is an imaging modality that is used extensively for ophthalmic diagnosis, near-histological visualization and quantification of retinal abnormalities such as cysts, exudates, retinal layer disorganization, etc. Intra-retinal cysts (IRCs) occur in several macular disorders such as, diabetic macular edema, retinal vascular disorders, age-related macular degeneration, and inflammatory disorders. Automated segmentation of IRCs poses challenges owing to variations in the acquisition system scan intensities, speckle noise, and imaging artifacts. Several segmentation methods have been proposed in the literature for IRC segmentation on vendor-specific OCT images that lack generalizability across imaging systems. In this work, we propose a fully convolutional network (FCN) model for vendor-independent IRC segmentation. The proposed method counteracts image noise variabilities and trains FCN models on OCT sub-images from the OPTIMA cyst segmentation challenge dataset (with four different vendor-specific images, namely, Cirrus, Nidek, Spectralis, and Topcon). Further, optimal data augmentation and model hyper-parametrization is shown to prevent over-fitting for IRC area segmentation. The proposed method is evaluated on the test data set with a recall/precision rate of 0.66/0.79 across imaging vendors. The Dice correlation coefficient of the proposed method outperforms that of the published algorithms in the OPTIMA cyst segmentation challenge with a Dice rate of 0.71 across the vendors.


  • Open Access

    Azure-Based Smart Monitoring System for Anemia-Like Pallor

    by Sohini Roychowdhury, Paul Hage, Joseph Vasquez

    Increasing costs of diagnostic healthcare have necessitated the development of hardware independent non-invasive Point of Care (POC) systems. Although anemia prevalence rates in global populations vary between 10% and 60% in various demographic groups, smart monitoring systems have not yet been developed for screening and tracking anemia-like pallor. In this work, we present two cloud platform-hosted POC applications that are directed towards smart monitoring of anemia-like pallor through eye and tongue pallor site images. The applications consist of a front-end graphical user interface (GUI) module and two different back-end image processing and machine learning modules. Both applications are hosted on a browser accessible tomcat server ported to an Azure Virtual Machine (VM). We observe that the first application spatially segments regions of interest from pallor site images with higher pallor classification accuracy and relatively longer processing times when compared to the lesser accurate yet faster second application. Also, both applications achieve 65%–98% accuracy in separating normal images from images with pallor or abnormalities. The optimized front-end module is significantly light-weight with a run-through time ratio of 10−5 with respect to the back-end modules. Thus, the proposed applications are portable and hardware independent, allowing for their use in pallor monitoring and screening tasks.