Publications

2023

1. 

   ICLR DistShift Workshop

   🏆 Oral Presentation

   Implicit Stylization for Domain Adaptation

   Park, J., Barnard, F., Hossain, S., and Rambhatla, S.

   Workshop on What do we need for successful domain generalization? International Conference on Learning Representations (ICLR) 2023

   Bib PDF

   @article{park2023b, abbr = {ICLR DistShift Workshop}, img = {imsty.jpeg}, title = {Implicit Stylization for Domain Adaptation}, author = {Park, J. and Barnard, F. and Hossain, S. and Rambhatla, S.}, journal = {Workshop on What do we need for successful domain generalization? International Conference on Learning Representations (ICLR)}, pdf = {https://openreview.net/forum?id=fkFFh4fAbH}, url = {https://openreview.net/forum?id=fkFFh4fAbH}, month = feb, year = {2023}, award = {Oral Presentation} }

2. 

   KDD

   🏆 Oral and Poster Presentation

   Domain-Guided Spatio-Temporal Self-Attention for Egocentric 3D Pose Estimation

   Park, J., Kaai, K., Hossain, S., Sumi, N., Rambhatla, S., and Fieguth, P.

   ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD) 2023

   Bib PDF Code

   @article{park2023a, abbr = {KDD}, title = {Domain-Guided Spatio-Temporal Self-Attention for Egocentric 3D Pose Estimation}, img = {egostan.gif}, author = {Park, J. and Kaai, K. and Hossain, S. and Sumi, N. and Rambhatla, S. and Fieguth, P.}, journal = {ACM SIGKDD International Conference on Knowledge Discovery \& Data Mining (KDD)}, url = {https://dl.acm.org/doi/10.1145/3580305.3599312}, month = aug, code = {https://github.com/jmpark0808/Ego-STAN}, pdf = {https://dl.acm.org/doi/10.1145/3580305.359931}, year = {2023}, award = {Oral and Poster Presentation} }

3. 

   NeurIPS DistShift Workshop

   Are all classes created equal? Domain Generalization for Domain-Linked Classes

   Kaai, K., Hossain, S., and Rambhatla, S.

   Workshop on Distribution Shifts (DistShift) at Neural Information Processing Systems 2023

   Bib PDF

   @article{kaai2023, abbr = {NeurIPS DistShift Workshop}, img = {average_dl_ds_plot_workshop.pdf}, title = {Are all classes created equal? Domain Generalization for Domain-Linked Classes}, author = {Kaai, K. and Hossain, S. and Rambhatla, S.}, journal = {Workshop on Distribution Shifts (DistShift) at Neural Information Processing Systems}, pdf = {https://openreview.net/forum?id=Zz7eZoR1VN}, url = {https://openreview.net/forum?id=Zz7eZoR1VN}, month = dec, year = {2023} }

2022

1. 

   AAAI

   I-SEA: Importance Sampling and Expected Alignment-based Deep Distance Metric Learning for Time Series Analysis and Embedding

   Rambhatla, S., Che, Z., and Liu, Y.

   36th Association for the Advancement of Artificial Intelligence (AAAI) conference on Artificial Intelligence 2022

   Abs Bib PDF Code

   Learning effective embeddings for potentially irregularly sampled time-series, evolving at different times scales, is fundamental for machine learning tasks such as classification and clustering. Task-dependent embeddings rely on similarities between data samples to learn effective geometries. However, many popular time-series similarity measures are not valid distance metrics, and as a result they do not reliably capture the intricate relationships between the multi-variate time-series data samples for learning effective embeddings. One of the primary ways to formulate an accurate distance metric is by forming distance estimates via Monte-Carlo-based expectation evaluations. However, the high-dimensionality of the underlying distribution, and the inability to sample from it, pose significant challenges. To this end, we develop an Importance Sampling based distance metric – ISEA – which enjoys the properties of a metric while consistently achieving superior performance for machine learning tasks such as classification and representation learning. ISEA leverages Importance Sampling and Non-parametric Density Estimation to adaptively estimate distances, enabling implicit estimation from the underlying high-dimensional distribution, resulting in improved accuracy and reduced variance. We theoretically establish the properties of ISEA and demonstrate its capabilities via experimental evaluations on real-world healthcare datasets.

   @article{Rambhatla2022isea, abbr = {AAAI}, title = {I-SEA: Importance Sampling and Expected Alignment-based Deep Distance Metric Learning for Time Series Analysis and Embedding}, author = {Rambhatla, S. and Che, Z. and Liu, Y.}, journal = {36th Association for the Advancement of Artificial Intelligence (AAAI) conference on Artificial Intelligence}, img = {IS_visualization.png}, pdf = {papers/ISEA_AAAI2022.pdf}, code = {https://github.com/srambhatla/ISEA}, month = feb, year = {2022} }

2. 

   UCNA

   Artificial Intelligence Applications in Urology: Reporting Standards to Achieve Fluency for Urologists

   Chen, A. B., Haque, T., Roberts, S., Rambhatla, S., Cacciamani, G., Dasgupta, P., and Hung, A. J.

   Urology Clinics North America 2022

   Bib

   @article{Chen2022a, abbr = {UCNA}, title = {Artificial Intelligence Applications in Urology: Reporting Standards to Achieve Fluency for Urologists}, author = {Chen, A. B. and Haque, T. and Roberts, S. and Rambhatla, S. and Cacciamani, G. and Dasgupta, P. and Hung, A. J.}, img = {RAS.gif}, journal = {Urology Clinics North America}, month = feb, url = {https://pubmed.ncbi.nlm.nih.gov/34776055/}, year = {2022} }

3. 

   TSAS

   Towards Accurate Spatiotemporal COVID-19 Risk Scores using High Resolution Real-World Mobility Data.

   Rambhatla*, S., Zeighami*, S., Shahabi, K., Shahabi, C., and Liu., Y.

   ACM Transactions on Spatial Algorithms and Systems (TSAS) 2022

   Abs Bib arXiv PDF

   As countries look towards re-opening of economic activities amidst the ongoing COVID-19 pandemic, ensuring public health has been challenging. While contact tracing only aims to track past activities of infected users, one path to safe reopening is to develop reliable spatiotemporal risk scores to indicate the propensity of the disease. Existing works which aim to develop risk scores either rely on compartmental model-based reproduction numbers (which assume uniform population mixing) or develop coarse-grain spatial scores based on reproduction number (R0) and macro-level density-based mobility statistics. Instead, in this paper, we develop a Hawkes process-based technique to assign relatively fine-grain spatial and temporal risk scores by leveraging high-resolution mobility data based on cell-phone originated location signals. While COVID-19 risk scores also depend on a number of factors specific to an individual, including demography and existing medical conditions, the primary mode of disease transmission is via physical proximity and contact. Therefore, we focus on developing risk scores based on location density and mobility behaviour. We demonstrate the efficacy of the developed risk scores via simulation based on real-world mobility data. Our results show that fine-grain spatiotemporal risk scores based on high-resolution mobility data can provide useful insights and facilitate safe re-opening.

   @article{Rambhatla2022highres, abbr = {TSAS}, title = {Towards Accurate Spatiotemporal COVID-19 Risk Scores using High Resolution Real-World Mobility Data.}, author = {Rambhatla*, S. and Zeighami*, S. and Shahabi, K. and Shahabi, C. and Liu., Y.}, arxiv = {2012.07283v1}, img = {spread_sim.gif}, url = {https://arxiv.org/abs/2012.07283v1}, pdf = {https://arxiv.org/pdf/2012.07283v1.pdf}, journal = {ACM Transactions on Spatial Algorithms and Systems (TSAS)}, month = feb, year = {2022} }

4. 

   ILTS

   🏆 Oral Presentation

   Predicting Future Trajectories of the Waitlisted NASH patients using Deep Learning

   Punchhi, G., Sun, Y., Rambhatla, S., and Bhat, M.

   Abstract at International Liver Transplantation Society 2022

   Bib

   @article{Punchhi2022a, abbr = {ILTS}, title = {Predicting Future Trajectories of the Waitlisted NASH patients using Deep Learning}, author = {Punchhi, G. and Sun, Y. and Rambhatla, S. and Bhat, M.}, img = {liver.png}, journal = {Abstract at International Liver Transplantation Society}, url = {}, month = mar, year = {2022}, award = {Oral Presentation} }

5. 

   EPIC and Ego4D @ CVPR

   🏆 Oral Presentation

   Building Spatio-temporal Transformers for Egocentric 3D Pose Estimation

   Park, J., Kaai, K., Hossain, S., Sumi, N., Rambhatla, S., and Fieguth, P.

   Joint International Workshop on Egocentric Perception, Interaction and Computing (EPIC) and Ego4D, IEEE/CVF Computer Vision and Pattern Recognition Conference (CVPR) 2022

   Bib

   @article{Park2022a, abbr = {EPIC and Ego4D @ CVPR}, title = {Building Spatio-temporal Transformers for Egocentric 3D Pose Estimation}, author = {Park, J. and Kaai, K. and Hossain, S. and Sumi, N. and Rambhatla, S. and Fieguth, P.}, img = {egostan.gif}, journal = {Joint International Workshop on Egocentric Perception, Interaction and Computing (EPIC) and Ego4D, IEEE/CVF Computer Vision and Pattern Recognition Conference (CVPR)}, url = {http://arxiv.org/abs/2206.04785}, month = jun, year = {2022}, award = {Oral Presentation} }

2021

1. 

   Surgery

   Road to Automating Robotic Suturing Skills Assessment: Battling Mislabeling of the Ground Truth

   Hung, A. J., Rambhatla, S., Sanford, D. I., Pachauri, N., Vanstrum, E., Nguyen, J. H., and Liu, Y.

   Surgery 2021

   Bib

   @article{Hung2021b, abbr = {Surgery}, title = {Road to Automating Robotic Suturing Skills Assessment: Battling Mislabeling of the Ground Truth}, author = {Hung, A. J. and Rambhatla, S. and Sanford, D. I. and Pachauri, N. and Vanstrum, E. and Nguyen, J. H. and Liu, Y.}, journal = {Surgery}, img = {NoiseRank.png}, url = {https://authors.elsevier.com/c/1dlnr90L~q2da}, year = {2021} }

2. 

   Plastic Surgery

   Predicting burn surgical candidacy using deep learning on photographic images.

   Huang*, S., Rambhatla*, S., Trinh, L., Zhang, M., Dong, M., Unadkat, V., Lin, J., Sheth, M. K., Dang, J., Yenikomshian, H. A., Liu, Y., and Gillenwater, J.

   Plastic Surgery: The Meeting 2021

   Bib

   @article{Huang2021a, abbr = {Plastic Surgery}, title = {Predicting burn surgical candidacy using deep learning on photographic images.}, author = {Huang*, S. and Rambhatla*, S. and Trinh, L. and Zhang, M. and Dong, M. and Unadkat, V. and Lin, J. and Sheth, M. K. and Dang, J. and Yenikomshian, H. A. and Liu, Y. and Gillenwater, J.}, img = {burn_cnn.png}, journal = {Plastic Surgery: The Meeting}, month = jan, year = {2021} }

3. 

   KDD

   Cross-Node Federated Graph Neural Network for Spatio-Temporal Data Modeling

   Meng, C., Rambhatla, S., and Liu, Y.

   ACM SIGKDD International Conference on Knowledge Discovery & Data Mining 2021

   Abs Bib

   Vast amount of data generated from networks of sensors, wearables, and the Internet of Things (IoT) devices underscores the need for advanced modeling techniques that leverage the spatio-temporal structure of decentralized data due to the need for edge computation and licensing (data access) issues. While federated learning (FL) has emerged as a framework for model training without requiring direct data sharing and exchange, effectively modeling the complex spatio-temporal dependencies to improve forecasting capabilities still remains an open problem. On the other hand, state-of-the-art spatio-temporal forecasting models assume unfettered access to the data, neglecting constraints on data sharing. To bridge this gap, we propose a federated spatio-temporal model – Cross-Node Federated Graph Neural Network (CNFGNN) – which explicitly encodes the underlying graph structure using graph neural network (GNN)-based architecture under the constraint of cross-node federated learning, which requires that data in a network of nodes is generated locally on each node and remains decentralized. CN-FGNN operates by disentangling the temporal dynamics modeling on devices and spatial dynamics on the server, utilizing alternating optimization to reduce the communication cost, facilitating computations on the edge devices. Experiments on the traffic flow forecasting task show that CNFGNN achieves the best forecasting performance in both transductive and inductive learning settings with no extra computation cost on edge devices, while incurring modest communication cost.

   @article{Meng2021a, abbr = {KDD}, title = {Cross-Node Federated Graph Neural Network for Spatio-Temporal Data Modeling}, author = {Meng, C. and Rambhatla, S. and Liu, Y.}, img = {cross-node.png}, journal = {ACM SIGKDD International Conference on Knowledge Discovery \& Data Mining}, month = jan, year = {2021} }

4. 

   AUA

   🏆 Podium Talk

   Automating suturing skills assessment with a limited surgeon dataset: Meta learning

   Hung, A. J., Rambhatla, S., Sanford, D. I., Pachauri, N., Nguyen, J. H., and Liu, Y.

   American Urology Association 2021

   Bib

   @article{Hung2021a, abbr = {AUA}, title = {Automating suturing skills assessment with a limited surgeon dataset: Meta learning}, author = {Hung, A. J. and Rambhatla, S. and Sanford, D. I. and Pachauri, N. and Nguyen, J. H. and Liu, Y.}, img = {RAS.gif}, journal = {American Urology Association}, award = {Podium Talk}, month = jan, year = {2021} }

5. 

   AMIA

   🏆 Oral Presentation

   DL4Burn: Burn surgical candidacy using multimodal deep learning.

   Huang*, S., Rambhatla*, S., Trinh, L., Zhang, M., Dong, M., Unadkat, V., Yenikomshian, H. A., Gillenwater, J., and Liu, Y.

   American Medical Informatics Association (AMIA) Annual Symposium 2021

   Bib

   @article{Huang2021b, abbr = {AMIA}, award = {Oral Presentation}, title = {DL4Burn: Burn surgical candidacy using multimodal deep learning.}, author = {Huang*, S. and Rambhatla*, S. and Trinh, L. and Zhang, M. and Dong, M. and Unadkat, V. and Yenikomshian, H. A. and Gillenwater, J. and Liu, Y.}, img = {DL4Burn.png}, journal = {American Medical Informatics Association (AMIA) Annual Symposium}, month = jan, year = {2021} }

6. 

   ICLR Workshop

   Simulating continuous-time human mobility trajectories.

   Xu*, N., Trinh*, L., Rambhatla, S., Assefa, S., Chen, J., Zeng, Z., and Liu, Y.

   Deep Learning for Simulation Workshop, International Conference on Learning Representations (ICLR) 2021

   Abs Bib

   Recent pandemic events have greatly emphasized the need to understand how hu-mans navigate in modern day cities for both urban planning and recommendingeffective public health policy. In this paper, we propose a two-stage generativemodel,DeltaGAN, to simulate realistic human mobility trajectories. Comparedwith existing work where time was discretized,DeltaGANgenerates continuousvisitation time which better captures temporal irregularity in human mobility be-haviors. Conditioned on the generated time,DeltaGANproduces accurate loca-tions via limiting the range of realizable candidates. Experimental results demon-strate our model achieves competitive distributional similarity to real-world GPStrajectories across 6 individual trajectory and geographical metrics. We furthervalidateDeltaGAN’s utility on COVID-19 spread simulation and observe the dif-fusion process under generated trajectories is consistent with that under real data.

   @article{Xu2021a, abbr = {ICLR Workshop}, title = {Simulating continuous-time human mobility trajectories.}, img = {DeltaGAN.png}, author = {Xu*, N. and Trinh*, L. and Rambhatla, S. and Assefa, S. and Chen, J. and Zeng, Z. and Liu, Y.}, journal = {Deep Learning for Simulation Workshop, International Conference on Learning Representations (ICLR)}, month = jan, year = {2021} }

7. 

   JHIR

   PolSIRD: Modeling Epidemic Spread under Intervention Policies and an Application to the Spread of COVID-19

   Kamra, N., Zhang, Y., Rambhatla, S., Meng, C., and Liu, Y.

   Journal of Healthcare Informatics Research 2021

   Abs Bib arXiv PDF

   Epidemic spread in a population is traditionally modeled via compartmentalized models which represent the free evolution of disease in absence of any intervention policies. In addition, these models assume full observability and do not account for under-reporting of cases. We present a mathematical model, namely PolSIRD, which accounts for the under-reporting by introducing an observation mechanism. It also captures the effects of intervention policies on the disease spread parameters by leveraging intervention policy data along with the reported disease cases. Furthermore, we allow our recurrent model to learn the initial hidden state of all compartments end-to-end along with other parameters via gradient-based training. We apply our model to spread of the recent global outbreak of COVID-19 where our model outperforms the current methods employed by the CDC in most of the metrics. We also provide actionable guidance on the lifting of intervention policies via counterfactual simulations from our model.

   @article{kamra2020pred, abbr = {JHIR}, title = {{PolSIRD}: Modeling Epidemic Spread under Intervention Policies and an Application to the Spread of COVID-19}, author = {Kamra, N. and Zhang, Y. and Rambhatla, S. and Meng, C. and Liu, Y.}, img = {polsird.png}, arxiv = {2009.01894}, url = {https://arxiv.org/abs/2009.01894}, pdf = {https://arxiv.org/pdf/2009.01894.pdf}, journal = {Journal of Healthcare Informatics Research}, month = jan, year = {2021} }

8. 

   WACV

   Interpretable and Trustworthy Deepfake Detection via Dynamic Prototypes

   Trinh, L., Tsang, M., Rambhatla, S., and Liu, Y.

   IEEE Winter Conference on Applications of Computer Vision (WACV) 2021

   Abs Bib arXiv PDF

   In this paper we propose a novel human-centered ap- proach for detecting forgery in face images, using dynamic prototypes as a form of visual explanations. Currently, most state-of-the-art deepfake detections are based on black-box models that process videos frame-by-frame for inference, and few closely examine their temporal inconsistencies. However, the existence of such temporal artifacts within deepfake videos is key in detecting and explaining deep- fakes to a supervising human. To this end, we propose Dynamic Prototype Network (DPNet) – an interpretable and effective solution that utilizes dynamic representations (i.e., prototypes) to explain deepfake temporal artifacts. Extensive experimental results show that DPNet achieves competitive predictive performance, even on unseen test- ing datasets such as Google’s DeepFakeDetection, Deep- erForensics, and Celeb-DF, while providing easy referential explanations of deepfake dynamics. On top of DPNet’s pro- totypical framework, we further formulate temporal logic specifications based on these dynamics to check our model’s compliance to desired temporal behaviors, hence providing trustworthiness for such critical detection systems.

   @article{trinh2020interpretable, abbr = {WACV}, title = {Interpretable and Trustworthy Deepfake Detection via Dynamic Prototypes}, author = {Trinh, L. and Tsang, M. and Rambhatla, S. and Liu, Y.}, journal = {IEEE Winter Conference on Applications of Computer Vision (WACV)}, pdf = {https://openaccess.thecvf.com/content/WACV2021/papers/Trinh_Interpretable_and_Trustworthy_Deepfake_Detection_via_Dynamic_Prototypes_WACV_2021_paper.pdf}, arxiv = {2006.15473}, url = {https://arxiv.org/abs/2006.15473}, img = {DPNet_resized.gif}, month = jan, year = {2021} }

9. 

   IJCAI

   Physics-aware Spatiotemporal Modules with Auxiliary Tasks for Meta-Learning

   Seo, S., Meng, C., Rambhatla, S., and Liu, Y.

   International Joint Conferences on Artificial Intelligence (IJCAI) 2021

   Abs Bib arXiv PDF

   Modeling the dynamics of real-world physical systems is critical for spatiotemporal prediction tasks, but challenging when data is limited. The scarcity of real-world data and the difficulty in reproducing the data distribution hinder directly applying meta-learning techniques. Although the knowledge of governing partial differential equations (PDEs) of the data can be helpful for the fast adaptation to few observations, it is difficult to generalize to different or unknown dynamics. In this paper, we propose a framework, physics-aware modular meta-learning with auxiliary tasks (PiMetaL) whose spatial modules incorporate PDE-independent knowledge and temporal modules are rapidly adaptable to the limited data, respectively. The framework does not require the exact form of governing equations to model the observed spatiotemporal data. Furthermore, it mitigates the need for a large number of real-world tasks for meta-learning by leveraging simulated data. We apply the proposed framework to both synthetic and real-world spatiotemporal prediction tasks and demonstrate its superior performance with limited observations.

   @article{seo2021physics, abbr = {IJCAI}, title = {Physics-aware Spatiotemporal Modules with Auxiliary Tasks for Meta-Learning}, author = {Seo, S. and Meng, C. and Rambhatla, S. and Liu, Y.}, journal = {International Joint Conferences on Artificial Intelligence (IJCAI)}, url = {https://arxiv.org/abs/2006.08831}, pdf = {https://arxiv.org/pdf/2006.08831.pdf}, arxiv = {2006.08831}, img = {PiMetal_zoom.png}, month = jan, year = {2021} }

2020

1. 

   NeurIPS

   Provable Online CP/PARAFAC Decomposition of a Structured Tensor via Dictionary Learning

   Rambhatla, S., Li, X., and Haupt, J.

   Advances in Neural Information Processing Systems (NeurIPS) 2020

   Abs Bib arXiv PDF Code Slides

   We consider the problem of factorizing a structured 3-way tensor into itsconstituent Canonical Polyadic (CP) factors. This decomposition, whichcan be viewed as a generalization of singular value decomposition (SVD)for tensors, reveals how the tensor dimensions (features) interact with eachother. However, since the factors area prioriunknown, the correspondingoptimization problems are inherently non-convex. The existing guaranteedalgorithms which handle this non-convexity incur an irreducible error (bias),and only apply to cases where all factors have the same structure. To thisend, we develop a provable algorithm for online structured tensor factor-ization, wherein one of the factors obeys some incoherence conditions, andthe others are sparse. Specifically we show that, under some relatively mildconditions on initialization, rank, and sparsity, our algorithm recovers thefactorsexactly(up to scaling and permutation) at a linear rate. Comple-mentary to our theoretical results, our synthetic and real-world data eval-uations showcase superior performance compared to related techniques.

   @article{Rambhatla19Tensor, abbr = {NeurIPS}, title = {Provable Online {CP/PARAFAC} Decomposition of a Structured Tensor via Dictionary Learning }, author = {Rambhatla, S. and Li, X. and Haupt, J.}, journal = {Advances in Neural Information Processing Systems (NeurIPS)}, slides = {/posters/Poster_TensorNOODL.pdf}, img = {TensorNOODL.png}, url = {https://papers.nips.cc/paper/2020/hash/85b42dd8aae56e01379be5736db5b496-Abstract.html}, pdf = {https://papers.nips.cc/paper/2020/file/85b42dd8aae56e01379be5736db5b496-Paper.pdf}, arxiv = {2006.16442}, code = {https://github.com/srambhatla/TensorNOODL}, selected = {true}, year = {2020} }

2. 

   NeurIPS

   How does this interaction affect me? Interpretable attribution for feature interactions

   Tsang, M., Rambhatla, S., and Liu, Y.

   Advances in Neural Information Processing Systems (NeurIPS) 2020

   Abs Bib arXiv PDF Code

   Machine learning transparency calls for interpretable explanations of how inputs relate to predictions. Feature attribution is a way to analyze the impact of features on predictions. Feature interactions are the contextual dependence between features that jointly impact predictions. There are a number of methods that extract feature interactions in prediction models; however, the methods that assign attributions to interactions are either uninterpretable, model-specific, or non-axiomatic. We propose an interaction attribution and detection framework called Archipelago which addresses these problems and is also scalable in real-world settings. Our experiments on standard annotation labels indicate our approach provides significantly more interpretable explanations than comparable methods, which is important for analyzing the impact of interactions on predictions. We also provide accompanying visualizations of our approach that give new insights into deep neural networks.

   @article{tsang2020does, abbr = {NeurIPS}, title = {How does this interaction affect me? Interpretable attribution for feature interactions}, author = {Tsang, M. and Rambhatla, S. and Liu, Y.}, journal = {Advances in Neural Information Processing Systems (NeurIPS)}, arxiv = {2006.10965}, pdf = {https://proceedings.neurips.cc/paper/2020/file/443dec3062d0286986e21dc0631734c9-Paper.pdf}, url = {https://proceedings.neurips.cc/paper/2020/hash/443dec3062d0286986e21dc0631734c9-Abstract.html}, code = {https://github.com/mtsang/archipelago}, img = {archipelago.png}, selected = {true}, year = {2020} }

3. 

   NeurIPS Workshop

   Physics-aware Spatiotemporal Modules with Auxiliary Tasks for Meta-Learning

   Seo, S., Meng, C., Rambhatla, S., and Liu, Y.

   Workshop on Machine Learning and the Physical Sciences, NeurIPS 2020

   Abs Bib arXiv PDF

   Modeling the dynamics of real-world physical systems is critical for spatiotemporal prediction tasks, but challenging when data is limited. The scarcity of real-world data and the difficulty in reproducing the data distribution hinder directly applying meta-learning techniques. Although the knowledge of governing partial differential equations (PDEs) of the data can be helpful for the fast adaptation to few observations, it is difficult to generalize to different or unknown dynamics. In this paper, we propose a framework, physics-aware modular meta-learning with auxiliary tasks (PiMetaL) whose spatial modules incorporate PDE-independent knowledge and temporal modules are rapidly adaptable to the limited data, respectively. The framework does not require the exact form of governing equations to model the observed spatiotemporal data. Furthermore, it mitigates the need for a large number of real-world tasks for meta-learning by leveraging simulated data. We apply the proposed framework to both synthetic and real-world spatiotemporal prediction tasks and demonstrate its superior performance with limited observations.

   @article{seo2020physics, abbr = {NeurIPS Workshop}, title = {Physics-aware Spatiotemporal Modules with Auxiliary Tasks for Meta-Learning}, author = {Seo, S. and Meng, C. and Rambhatla, S. and Liu, Y.}, journal = {Workshop on Machine Learning and the Physical Sciences, NeurIPS}, url = {https://arxiv.org/abs/2006.08831}, pdf = {https://arxiv.org/pdf/2006.08831.pdf}, arxiv = {2006.08831}, img = {PiMetal_zoom.png}, year = {2020} }

4. 

   Under Review

   Coronavirus on social media: Analyzing misinformation in Twitter conversations

   Sharma, K., Seo, S., Meng, C., Rambhatla, S., and Liu, Y.

   arXiv preprint arXiv:2003.12309 2020

   Abs Bib arXiv PDF

   The ongoing Coronavirus (COVID-19) pandemic highlights the inter-connectedness of our present- day globalized world. With social distancing policies in place, virtual communication has become an important source of (mis)information. As increasing number of people rely on social media platforms for news, identifying misinformation and uncovering the nature of online discourse around COVID-19 has emerged as a critical task. To this end, we collected streaming data related to COVID-19 using the Twitter API, starting March 1, 2020. We identified unreliable and misleading contents based on fact-checking sources, and examined the narratives promoted in misinformation tweets, along with the distribution of engagements with these tweets. In addition, we provide examples of the spreading patterns of prominent misinformation tweets. The analysis is presented and updated on a publically accessible dashboard1 to track the nature of online discourse and misinformation about COVID-19 on Twitter from March 1 - June 5, 2020. The dashboard provides a daily list of identified misinformation tweets, along with topics, sentiments, and emerging trends in the COVID-19 Twitter discourse. The dashboard is provided to improve visibility into the nature and quality of information shared online, and provide real-time access to insights and information extracted from the dataset.

   @article{sharma2020coronavirus, abbr = {Under Review}, title = {Coronavirus on social media: Analyzing misinformation in Twitter conversations}, author = {Sharma, K. and Seo, S. and Meng, C. and Rambhatla, S. and Liu, Y.}, url = {https://arxiv.org/abs/2003.12309}, pdf = {https://arxiv.org/pdf/2003.12309.pdf}, arxiv = {2003.12309}, img = {spread_misinfo.gif}, journal = {arXiv preprint arXiv:2003.12309}, year = {2020} }

5. 

   TSP

   A Dictionary-Based Generalization of Robust PCA With Applications to Target Localization in Hyperspectral Imaging

   Rambhatla, S., Li, X., Ren, J., and Haupt, J.

   IEEE Tran. on Signal Processing 2020

   Abs Bib arXiv PDF Blog Code

   We consider the decomposition of a data matrix assumed to be a superposition of a low-rank matrix and a component which is sparse in a known dictionary, using a convex demixing method. We consider two sparsity structures for the sparse factor of the dictionary sparse component, namely entry-wise and column-wise sparsity, and provide a unified analysis, encompassing both undercomplete and the overcomplete dictionary cases, to show that the constituent matrices can be successfully recovered under some relatively mild conditions on incoherence, sparsity, and rank. We leverage these results to localize targets of interest in a hyperspectral (HS) image based on their spectral signature(s) using the a priori known characteristic spectral responses of the target. We corroborate our theoretical results and analyze target localization performance of our approach via experimental evaluations and comparisons to related techniques.

   @article{Rambhatla20LrTheoApp, abbr = {TSP}, author = {{Rambhatla}, S. and {Li}, X. and {Ren}, J. and {Haupt}, J.}, journal = {IEEE Tran. on Signal Processing}, title = {A Dictionary-Based Generalization of Robust PCA With Applications to Target Localization in Hyperspectral Imaging}, year = {2020}, volume = {68}, code = {https://github.com/srambhatla/Dictionary-based-Robust-PCA}, number = {}, pdf = {https://arxiv.org/pdf/1902.08304v3.pdf}, arxiv = {1902.08304v3}, img = {LplusDA_hyper_spectral.gif}, url = {https://arxiv.org/abs/1902.08304}, pages = {1760-75}, selected = {true}, blog = {2017/HyperSpectral/} }

2019

1. 

   ICLR

   🏆 Travel Award

   NOODL: Provable Online Dictionary Learning and Sparse Coding

   Rambhatla, S., Li, X., and Haupt, J.

   International Conference on Learning Representations (ICLR) 2019

   Abs Bib arXiv PDF Code Slides

   We consider the dictionary learning problem, where the aim is to model the givendata as a linear combination of a few columns of a matrix known as adictionary,where the sparse weights forming the linear combination are known ascoeffi-cients. Since the dictionary and coefficients, parameterizing the linear model areunknown, the corresponding optimization is inherently non-convex. This was amajor challenge until recently, when provable algorithms for dictionary learningwere proposed. Yet, these provide guarantees only on the recovery of the dic-tionary, without explicit recovery guarantees on the coefficients. Moreover, anyestimation error in the dictionary adversely impacts the ability to successfully lo-calize and estimate the coefficients. This potentially limits the utility of existingprovable dictionary learning methods in applications where coefficient recoveryis of interest. To this end, we develop NOODL: a simple Neurally plausible alter-nating Optimization-based Online Dictionary Learning algorithm, which recoversboththe dictionary and coefficientsexactlyat a geometric rate, when initializedappropriately. Our algorithm, NOODL, is also scalable and amenable for largescale distributed implementations in neural architectures, by which we mean thatit only involves simple linear and non-linear operations. Finally, we corroboratethese theoretical results via experimental evaluation of the proposed algorithmwith the current state-of-the-art techniques.

   @article{Rambhatla2019, abbr = {ICLR}, title = {{NOODL}: Provable Online Dictionary Learning and Sparse Coding}, author = {Rambhatla, S. and Li, X. and Haupt, J.}, journal = {International Conference on Learning Representations (ICLR)}, year = {2019}, url = {https://openreview.net/forum?id=HJeu43ActQ}, selected = {true}, arxiv = {1902.11261}, slides = {/posters/NOODL_poster_ICLR.pdf}, pdf = {https://openreview.net/pdf?id=HJeu43ActQ}, code = {https://github.com/srambhatla/NOODL}, img = {NOODL.png}, award = {Travel Award} }

2. 

   Doctoral Thesis

   Provably Learning from Data: New Algorithms and Models for Matrix and Tensor Decompositions

   Rambhatla, S.

   University of Minnesota – Twin Cities, Minneapolis, MN 2019

   Abs Bib PDF

   Learning and leveraging patterns in data has fueled the recent advances in data driven services. As these solutions become more ubiquitous, and get incorporated into critical applications in healthcare and transportation, there is an increasing need to understand the limits of these learning algorithms and to develop algorithms with guar- antees. Moreover, with data being generated at unprecedented rates, these algorithms need to be fast, learn on-the-fly (online), handle large volumes of data (scalable), and be computationally efficient, while possessing guarantees on their behavior. Further- more, to make the learning-based products widely applicable there is also a need to make their reasoning and decision making process transparent (interpretable). These challenges inspire and motivate this dissertation. Specifically, we focus on analyzing various matrix/tensor demixing and factorization tasks, where we leverage the inherent interpretability endowed by the structure of problem (such as sparsity and low-rankness) to characterize the (theoretical) conditions for successful recovery, and analyze their performance in real-world settings. To this end, we make contributions on three fronts. First, we develop algorithm- aware theoretical guarantees for sparse matrix and tensor factorization tasks. Second, we establish algorithm-agnostic theoretical results for matrix demixing models and demonstrate their applications on real-world datasets. Lastly, we develop application-specific techniques for navigation and source separation. Bringing together Algorithms, Theory, and Applications, the techniques and theoretical results developed as part of this dissertation facilitate and motivate future explorations into the inner workings of learning algorithms for their safe use in critical applications.

   @article{Rambhatla2019thesis, abbr = {Doctoral Thesis}, author = {Rambhatla, S.}, title = {Provably Learning from Data: New Algorithms and Models for Matrix and Tensor Decompositions}, journal = {{U}niversity of {M}innesota -- {T}win {C}ities, {M}inneapolis, {MN}}, year = {2019}, img = {phd_thesis.png}, pdf = {PhD_thesis_Sirisha_Rambhatla.pdf} }

2018

1. 

   ICASSP

   Robust PCA via Dictionary Based Outlier Pursuit

   Li, X., Ren, J., Rambhatla, S., Xu, Y., and Haupt, J.

   2018 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP) 2018

   Abs Bib PDF Code

   In this paper, we examine the problem of locating vector out- liers from a large number of inliers, with a particular focus on the case where the outliers are represented in a known basis or dictionary. Using a convex demixing formulation, we provide provable guarantees for exact recovery of the space spanned by the inliers and the supports of the outlier columns, even when the rank of inliers is high and the number of outliers is a constant proportion of total observations. Comprehensive numerical experiments on both synthetic and hyper-spectral imaging real datasets demonstrate the efficiency of our pro- posed method.

   @article{Li2018icassp, abbr = {ICASSP}, title = {Robust PCA via Dictionary Based Outlier Pursuit}, author = {Li, X. and Ren, J. and Rambhatla, S. and Xu, Y. and Haupt, J.}, journal = {2018 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP)}, year = {2018}, code = {https://github.com/srambhatla/Dictionary-based-Robust-PCA}, pdf = {Robust_PCA_via_Dictionary_Based_Outlier_Pursuit.pdf}, img = {DRPCA.png}, organization = {IEEE} }

2. 

   GlobalSIP

   TensorMap: Lidar-based Topological Mapping and Localization via Tensor Decompositions

   Rambhatla, S., Sidiropoulos, N., and Haupt, J.

   IEEE Global Conference on Signal and Information Processing (GlobalSIP) 2018

   Abs Bib arXiv PDF Code

   We propose a technique to develop (and localize in) topological maps from light detection and ranging (Lidar) data. Localizing an autonomous vehicle with respect to a reference map in real-time is crucial for its safe operation. Owing to the rich information provided by Lidar sensors, these are emerging as a promising choice for this task. However, since a Lidar outputs a large amount of data every fraction of a second, it is progressively harder to process the information in real-time. Consequently, current systems have migrated towards faster alternatives at the expense of accuracy. To overcome this inherent trade-off between latency and accuracy, we propose a technique to develop topological maps from Lidar data using the orthogonal Tucker3 tensor decomposition. Our experimental evaluations demonstrate that in addition to achieving a high compression ratio as compared to full data, the proposed technique, TensorMap, also accurately detects the position of the vehicle in a graph-based representation of a map. We also analyze the robustness of the proposed technique to Gaussian and translational noise, thus initiating explorations into potential applications of tensor decompositions in Lidar data analysis.

   @article{Rambhatla18TensorMap, abbr = {GlobalSIP}, author = {Rambhatla, S. and Sidiropoulos, N. and Haupt, J.}, title = {TensorMap: Lidar-based Topological Mapping and Localization via Tensor Decompositions}, journal = {IEEE Global Conference on Signal and Information Processing (GlobalSIP)}, year = {2018}, organization = {IEEE}, arxiv = {1902.10226}, img = {TensorMap.png}, code = {https://github.com/srambhatla/TensorMap}, url = {https://arxiv.org/abs/1902.10226}, pdf = {https://arxiv.org/pdf/1902.10226.pdf} }

2017

1. 

   Asilomar

   🏆 Best Paper Finalist

   Target-based hyperspectral demixing via generalized robust PCA

   Rambhatla, S., Li, X., and Haupt, J.

   51st Asilomar Conference on Signals, Systems, and Computers, 2017 2017

   Bib arXiv PDF Blog Code

   @article{Rambhatla17DRPCAApp, abbr = {Asilomar}, author = {Rambhatla, S. and Li, X. and Haupt, J.}, title = {Target-based hyperspectral demixing via generalized robust {PCA}}, journal = {51st Asilomar Conference on Signals, Systems, and Computers, 2017}, img = {LplusDA_hyper_spectral.gif}, pages = {420--424}, year = {2017}, crossref = {DBLP:conf/acssc/2017}, doi = {10.1109/ACSSC.2017.8335372}, timestamp = {Fri, 20 Apr 2018 22:49:39 +0200}, biburl = {https://dblp.org/rec/bib/conf/acssc/RambhatlaLH17}, bibsource = {dblp computer science bibliography, https://dblp.org}, arxiv = {1902.11111}, code = {https://github.com/srambhatla/Dictionary-based-Robust-PCA}, pdf = {https://arxiv.org/pdf/1902.11111.pdf}, url = {https://arxiv.org/abs/1902.11111}, award = {Best Paper Finalist}, blog = {2017/HyperSpectral/} }

2016

1. 

   GlobalSIP

   🏆 NSF Travel Award

   A Dictionary Based Generalization of Robust PCA

   Rambhatla, S., Li, X., and Haupt, J.

   IEEE Global Conference on Signal and Information Processing (GlobalSIP) 2016

   Bib arXiv PDF Blog Code Slides

   @article{Rambhatla2016DRPCA, abbr = {GlobalSIP}, title = {A Dictionary Based Generalization of Robust {PCA}}, author = {Rambhatla, S. and Li, X. and Haupt, J.}, journal = {IEEE Global Conference on Signal and Information Processing (GlobalSIP)}, year = {2016}, slides = {/slides/Dictionary_based_generalization_of_robust_PCA_Sirisha_R.pdf}, organization = {IEEE}, arxiv = {1902.08171}, img = {DRPCA.png}, code = {https://github.com/srambhatla/Dictionary-based-Robust-PCA}, pdf = {https://arxiv.org/pdf/1902.08171.pdf}, url = {https://arxiv.org/abs/1902.08171}, award = {NSF Travel Award}, blog = {2017/HyperSpectral/} }

2013

1. 

   Asilomar

   Semi-blind source separation via sparse representations and online dictionary learning

   Rambhatla, S., and Haupt, J.

   Asilomar Conference on Signals, Systems and Computers 2013

   Abs Bib arXiv PDF Blog Slides

   This work examines a semi-blind single-channel source separation problem. Our specific aim is to separate one source whose local structure is approximately known, from another a priori unspecified background source, given only a single linear combination of the two sources. We propose a separation technique based on local sparse approximations along the lines of recent efforts in sparse representations and dictionary learning. A key feature of our procedure is the online learning of dictionaries (using only the data itself) to sparsely model the background source, which facilitates its separation from the partially-known source. Our approach is applicable to source separation problems in various application domains; here, we demonstrate the performance of our proposed approach via simulation on a stylized audio source separation task.

   @article{Rambhatla2013sbmca, abbr = {Asilomar}, title = {Semi-blind source separation via sparse representations and online dictionary learning}, author = {Rambhatla, S. and Haupt, J.}, journal = {Asilomar Conference on Signals, Systems and Computers}, pages = {1687--1691}, year = {2013}, organization = {IEEE}, arxiv = {1212.0451}, pdf = {https://arxiv.org/pdf/1212.0451.pdf}, img = {SBMCA.png}, blog = {2016/SBMCA/}, slides = {PresentationThesisPDF.pdf} }

2012

1. 

   Masters Thesis

   Semi-blind source separation via sparse representations and online dictionary learning

   Rambhatla, S.

   University of Minnesota – Twin Cities, Minneapolis, MN 2012

   Abs Bib PDF Blog

   This work examines a semi-blind source separation problem having applications in audio, image, and video processing. The essential aim is to separate one source whose local structure is partially or approximately known from another a priori unspecified but structured source, given only a single linear combination of the two sources. We propose a novel separation technique based on local sparse approximations. A key feature of our procedure is the online learning of dictionaries (using only the data itself) to sparsely model the unknown structured background source. The performance of the proposed approach is demonstrated via simulation in a stylized application of audio source separation.

   @article{Rambhatla2012mthesis, abbr = {Masters Thesis}, author = {Rambhatla, S.}, title = {Semi-blind source separation via sparse representations and online dictionary learning}, journal = {{U}niversity of {M}innesota -- {T}win {C}ities, {M}inneapolis, {MN}}, year = {2012}, img = {SBMCA_thesis.png}, blog = {2016/SBMCA/}, pdf = {Sirisha_Rambhatla_Masters_Thesis.pdf}, organization = {{U}niversity of {M}innesota -- {T}win {C}ities, {M}inneapolis, {MN}} }