Kowshik Thopalli

Kowshik Thopalli

Auto-generated profile, based on publicly available information

Kowshik Thopalli

Lawrence Livermore National Laboratory, Livermore, CA, USA

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
6 October 2018
Verified
Book Chapter
Closed access
PDF file
ASAnirudh Som
Anirudh Som
KTKowshik Thopalli
Kowshik Thopalli
KRKarthikeyan Natesan Ramamurthy
Karthikeyan Natesan Ramamurthy
VVVinay Venkataraman
Vinay Venkataraman
ASAnkita Shukla
Ankita Shukla
PTPavan Turaga
Pavan Turaga

25 January 2023 in IEEE Access
Verified
Journal Article
Open access
KTKowshik Thopalli
Kowshik Thopalli
RARushil Anirudh
Rushil Anirudh
PTPavan Turaga
Pavan Turaga
JTJayaraman J. Thiagarajan
Jayaraman J. Thiagarajan
Unsupervised domain adaptation (UDA) aims to transfer and adapt knowledge from a labeled source domain to an unlabeled target domain. Traditionally, geometry-based alignment methods, e.g., Orthogonal Procrustes Alignment (OPA), formed an important class of solutions to this problem. Despite their mathematical tractability, they rarely produce effective adaptation performance with the recent benchmarks. Instead, state-of-the-art approaches rely on sophisticated distribution alignment strategies such as adversarial training. In this paper, we show that, conventional OPA, when coupled with powerful deep feature extractors and a novel bi-level optimization formulation, is indeed an effective choice for handling challenging distribution shifts. When compared to existing UDA methods, our approach offers the following benefits: computational efficiency : Through the isolation of alignment and classifier training steps during adaptation, and the use of deep OPA, our approach is computationally very effective (typically requiring only 700 K parameters more than the base feature extractor as compared to millions of extra parameters required by state-of-the-art UDA baselines); (ii) data efficiency : Our approach does not require updating our feature extractor during adaptation and hence can be effective even with limited target data; (iii) improved generalization : The resulting models are intrinsically well-regularized and demonstrate effective generalization even in the challenging partial DA setting, i.e ., target domain contains only a subset of the classes observed in the source domain.; and (iv) incremental training : Our approach allows progressive adaptation of models to novel domains (unseen during training) without requiring retraining of the model from scratch.

20 July 2025
Verified
Conference Paper
Open access
ADArunavo Dey
Arunavo Dey
NANeil Antony
Neil Antony
ADAakash Raj Dhakal
Aakash Raj Dhakal
KTKowshik Thopalli
Kowshik Thopalli
JTJayaraman J. Thiagarajan
Jayaraman J. Thiagarajan
TITanzima Islam
Tanzima Islam

2 October 2023
Verified
Conference Paper
Open access
KTKowshik Thopalli
Kowshik Thopalli
DSDevi S
Devi S
JTJayaraman J. Thiagarajan
Jayaraman J. Thiagarajan
The recent progress in developing pre-trained models, trained on large-scale datasets, has highlighted the need for robust protocols to effectively adapt them to domain-specific data, especially when there is a limited amount of available data. Data augmentations can play a critical role in enabling data-efficient fine-tuning of pre-trained object detection models. Choosing the right augmentation policy for a given dataset is challenging and relies on knowledge about task-relevant invariances. In this work, we focus on an understudied aspect of this problem – can bounding box annotations be used to design more effective augmentation policies? Through InterAug, we make a critical finding that, we can leverage the annotations to infer the effective context for each object in a scene, as opposed to manipulating the entire scene or only within the pre-specified bounding boxes. Using a rigorous empirical study with multiple benchmarks and architectures, we demonstrate the efficacy of InterAug in improving robustness and handling data scarcity. Finally, InterAug can be used with any off-the-shelf policy, does not require any modification to the architecture, and significantly outperforms existing protocols. Our codes can be found at https://github.com/kowshikthopalli/InterAug.

30 September 2020 in arXiv
Verified
Preprint
Open access
PDF file
BVBindya Venkatesh
Bindya Venkatesh
JTJayaraman J. Thiagarajan
Jayaraman J. Thiagarajan
KTKowshik Thopalli
Kowshik Thopalli
PSPrasanna Sattigeri
Prasanna Sattigeri
The hypothesis that sub-network initializations (lottery) exist within the initializations of over-parameterized networks, which when trained in isolation produce highly generalizable models, has led to crucial insights into network initialization and has enabled efficient inferencing. Supervised models with uncalibrated confidences tend to be overconfident even when making wrong prediction. In this paper, for the first time, we study how explicit confidence calibration in the over-parameterized network impacts the quality of the resulting lottery tickets. More specifically, we incorporate a suite of calibration strategies, ranging from mixup regularization, variance-weighted confidence calibration to the newly proposed likelihood-based calibration and normalized bin assignment strategies. Furthermore, we explore different combinations of architectures and datasets, and make a number of key findings about the role of confidence calibration. Our empirical studies reveal that including calibration mechanisms consistently lead to more effective lottery tickets, in terms of accuracy as well as empirical calibration metrics, even when retrained using data with challenging distribution shifts with respect to the source dataset.

27 June 2021 in bioRxiv
Verified
Preprint
Open access
PDF file
TATamim Ahmed
Tamim Ahmed
KTKowshik Thopalli
Kowshik Thopalli
TRThanassis Rikakis
Thanassis Rikakis
PTPavan Turaga
Pavan Turaga
AKAisling Kelliher
Aisling Kelliher
SWSteve Wolf
Steve Wolf
We are developing a system for long term Semi-Automated Rehabilitation At the Home (SARAH) that relies on low-cost and unobtrusive video-based sensing. We present a cyber-human methodology used by the SARAH system for automated assessment of upper extremity stroke rehabilitation at the home. We propose a hierarchical model for automatically segmenting stroke survivor’s movements and generating training task performance assessment scores during rehabilitation. The hierarchical model fuses expert therapist knowledge-based approaches with data-driven techniques. The expert knowledge is more observable in the higher layers of the hierarchy (task and segment) and therefore more accessible to algorithms incorporating high level constraints relating to activity structure (i.e. type and order of segments per task). We utilize an HMM and a Decision Tree model to connect these high level priors to data driven analysis. The lower layers (RGB images and raw kinematics) need to be addressed primarily through data driven techniques. We use a transformer based architecture operating on low-level action features (tracking of individual body joints and objects) and a Multi-Stage Temporal Convolutional Network(MS-TCN) operating on raw RGB images. We develop a sequence combining these complimentary algorithms effectively, thus encoding the information from different layers of the movement hierarchy. Through this combination, we produce a robust segmentation and task assessment results on noisy, variable and limited data, which is characteristic of low cost video capture of rehabilitation at the home. Our proposed approach achieves 85% accuracy in per-frame labeling, 99% accuracy in segment classification and 93% accuracy in task completion assessment. Although the methodology proposed in this paper applies to upper extremity rehabilitation using the SARAH system, it can potentially be used, with minor alterations, to assist automation in many other movement rehabilitation contexts (i.e. lower extremity training for neurological accidents).