Information Sciences Group
Computational & Statistical Sciences Division
Los Alamos National Laboratory
Los Alamos, NM 87545

I am a computer scientist and project leader at Los Alamos National Laboratory, working in the areas of robustness and reliability in modern machine learning. In addition, I also have extensive research experience in semi-supervised learning, high-performance computing, graph algorithms, distributed computing and large-scale netwok modeling. I obtained my PhD in machine learning from the University of Washington.


  • Xiaoying Pang, Sunil Thulasidasan, Larry Rybarcyk. Autonomous Control of a Particle Accelerator using Deep Reinforcement Learning, NeurIPS 2020 Workshop on Machine Learning for Engineering Modeling, Simulation, and Design, December 2020 arXiv
  • Sayera Dhaubhadel, Jamaludin Mohd-Yusof, Kumkum Ganguly, Gopinath Chennupati, Sunil Thulasidasan, Nicolas Hengartner, Brent J Mumphrey, Eric B Durban, Jennifer A Doherty, Mireille Lemieux, Noah Schaefferkoetter, Georgia Tourassi, Linda Coyle, Lynne Penberthy, Benjamin McMahon, Tanmoy Bhattacharya Why I’m not Answering: Understanding Determinants of Classification of an Abstaining Classifier for Cancer Pathology Reports, In the Sixth Computational Approaches for Cancer Workshop (CAFCW20), November 2020 arXiv
  • Sunil Thulasidasan, Gopinath Chennupati, Jeff Bilmes, Tanmoy Bhattacharya, Sarah Michalak On Mixup Training: Improved Calibration and Predictive Uncertainty for Deep Neural Networks, Proceedings of the Thirty-third Conference on Neural Information Processing Systems (NeurIPS 2019), Vancouver, Canada, December 2019 arXiv
  • Sunil Thulasidasan, Robust Training and Predictive Uncertainty in Deep Learning, talk given at Workshop on AI and Tensor Factorization for Physical, Chemical, and Biological Systems, Santa Fe, 2019 Slides
  • Sunil Thulasidasan, Tanmoy Bhattacharya, Jeff Bilmes, Gopinath Chennupati, Jamaludin Mohd-Yusof Combating Label Noise in Deep Learning Using Abstention, Proceedings of the 36th International Conference on Machine Learning (ICML 2019), Long Beach, California, June 2019 arXiv code
  • Sunil Thulasidasan, Gopinath Chennupati, Jeff Bilmes, Tanmoy Bhattacharya, Sarah Michalak On Mixup Training: Improved Calibration and Predictive Uncertainty for Deep Neural Networks, ICML 2019 Workshop on Uncertainty and Robustness in Deep Learning (this is a shorter version of the NeurIPS 2019 version) paper
  • Gopinath Chennupati, Kumkum Ganguly, Benjamin McMahon, Sunil Thulasidasan, Jessica Boten, Valentina Petkov, Lynne Penberthy, Xiao- Cheng Wu, Paul Fearn and Tanmoy Bhattacharya Extracting Breast Cancer Genetic Markers in Pathology Reports using Natural Language Processing, Annual conference of North American Association of Central Cancer Registries (NAACR), June, 2018 pdf
  • Sunil Thulasidasan, Jeff Bilmes, Acoustic Classification Using Semi-Supervised Deep Neural Networks and Stochastic Entropy-Regularization over Nearest-Neighbor Graphs, Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), New Orleans, March 2017
  • Sunil Thulasidasan, Jeff Bilmes, Garrett Kenyon, Efficient Distributed Semi-Supervised Learning using Stochastic Regularization over Affinity Graphs, In the 2016 NIPS Workshop on Machine Learning Systems, Barcelona, Spain, December 2016 arXiv
  • Sunil Thulasidasan, Jeff Bilmes, Semi-Supervised Phone Classification using Deep Neural Networks and Stochastic Graph-Based Entropic Regularization, In the 2016 Workshop on Machine Learning in Speech and Language Processing, San Francisco, CA, September 2016. arXiv
  • Chandrashekhar Lavania, Sunil Thulasidasan, Anthony Lamarca, Jeff Scofield, and Jeff Bilmes, A Weakly Supervised Online Activity Recognition Framework for Real-time Synthetic Biology Laboratory Assistance, In 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp 2016), Heidelberg, Germany, September 2016.
  • Sunil Thulasidasan, Jeff Bilmes, Semi-Supervised Deep Learning with Graphs, 2016 Workshop on Data Science and Optimal Learning for Material Discovery and Design, Santa Fe, NM, May 2016 (poster) LA-UR-16-24600
  • Jiangzhuo Chen, Anil Kumar V. S., Madhav Marathe, M. V, Ravi Sundaram, Mayur Thakur, Sunil Thulasidasan, A Study of the Structure and Vulnerabilities of Metropolitan Area Networks, In Proceedings of the 8th International Conference on Communication Systems & Networks (COMSNETS), Bangalore, India, January 2016
  • Hristo Djidjev, Guillaume Chapuis, Rumen Andonov, Sunil Thulasidasan, Dominique Lavenier, All-Pairs Shortest Path Algorithms for Planar Graph for GPU-accelerated Clusters, In the Journal of Parallel and Distributed Computing, Volume 85, Issue C, November 2015
  • Sunil Thulasidasan, Lukas Kroc, Stephan Eidenbenz, Developing Parallel Discrete Event Simulations in Python: First Results and User Experiences with the SimX Library, In the 4th International Conference Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2014), Vienna, Austria, August 2014. LA-UR-12-26739 [PDF]
  • Hristo Djidjev, Sunil Thulasidasan, Rumen Andonov, Guillaume Chapuis, Dominique Lavenier, Efficient Multi-GPU Computation of All-Pairs Shortest Paths, In proceedings of 2014 IEEE International Parallel and Distributed Processing Symposium (IPDPS), Phoenix, May 2014. LA-UR-13-28111 [PDF]
  • Sunil Thulasidasan, The Graph Laplacian and the Dynamics of Complex Networks, (talk), given at the Department of Applied Mathematics, University of Washington. LANL Technical Report LA-UR-12-22066. [PDF]
  • Sunil Thulasidasan, A Simulation Based Analysis of Distributed Coverage and Dynamic Relocation for Mobile Sensor Networks using Potential Fields, LANL Technical Report LA-UR-12-25669 [PDF]
  • Stephan Eidenbenz, Kei Davis, Art Voter, Hristo Djidjev, Leonid Gurvits, Christoph Junghans, Susan Mniszewski, Danny Perez, Nandakishore Santhi, Sunil Thulasidasan Optimization Principles for Codesign Applied to Molecular Dynamics: Design Space Exploration, Performance Prediction, and Optimization Strategies, In Proceedings of the DOE ASCR Exascale Research Conference, Portland, OR, April 2012. LA-UR 12-20070 [PDF]
  • Sunil Thulasidasan, Shiva Kasivishwanathan, Stephan Eidenbenz, Philip Romero, Explicit Spatial Scattering for Load Balancing in Conservatively Synchronized Distributed Discrete Event Simulations, In Proceedings of the 24th ACM /IEEE /SCS Workshop on the Principles of Advanced and Distributed Simulations (PADS), Atlanta, May 2010. [PDF]
  • Sunil Thulasidasan, Shiva Kasiviswanathan, Stephan Eidenbenz, Emmanuelle Galli, Susan Mniszewski, Philip Romero, Designing Systems for Large-Scale, Discrete-Event Simulations: Experiences with the FastTrans Parallel Microsimulator, In Proceedings of the IEEE International Conference on High Performance Computing (HiPC), Kochi, India, December 2009. [PDF]
  • Guanhua Yan, Stephan Eidenbenz, Sunil Thulasidasan, Venkatesh Ramaswamy, Pallab Datta, Criticality Analysis of Internet Infrastructure, Computer Networks Journal, Special Issue, 2009. [PDF]
  • Russell Bent, Stephan Eidenbenz, Sunil Thulasidasan, Large-scale Telephone Network Simulation: Discrete Event vs. Steady State, In Proceedings of the 2009 Spring Simulation Multiconference, March 2009.
  • Venkatesh Ramaswamy, Sunil Thulasidasan, Stephan Eidenbenz, Philip Romero, Leticia Cuellar, Simulating the National Telephone Network: A Sociotechnical Approach to Assessing Infrastructure Criticality, In Proceedings of the Military Communications Conference (MILCOM), 2007.
  • Gabriel Istrate, Anders Hansson, Sunil Thulasidasan, Madhav Marathe, Christopher Barrett, Semantic Compression of TCP Traces, IFIP Networking Conference Coimbra, Portugal, 2006. [PDF]
  • V.S. Anil Kumar, Madhav Marathe, Ravi Sundaram, Mayur Thakur, Sunil Thulasidasan, Scaling laws for the Internet over Urban Regions, CAIDA ISMA Workshop on Internet Topology, San Diego, 2006. [PDF]
  • Christopher Barrett, Gabriel Istrate, V.S Anil Kumar, Madhav Marathe, Shripad Thite, Sunil Thulasidasan, Strong Edge Coloring for Channel Assignment in Wireless RadioNetworks, IEEE International Workshop on Foundations and Algorithms for Wireless Networks (FAWN), Pisa, Italy, 2006. [PDF]
  • Mark Gardner, Sunil Thulasidasan, Wu-chun Feng, User-Space Auto-Tuning for TCP Flow Control inComputational Grids, Journal of Computer Communications, 2004, Special Issue. [PDF]
  • Sunil Thulasidasan, Mark Gardner, Wu-chun Feng, Optimizing GridFTP Through Dynamic Right-Sizing, 2003 IEEE Conference on High Performance and Distributed Computing (HPDC), Seattle, June 2003. [PDF]
  • Wu-chun Feng, Apu Kapadia, Sunil Thulasidasan, GREEN: Proactive Queue Management over a Best Effort Network, 2002 IEEE Globecom, Taipei, Taiwan, November 2002.


  • Deep Abstaining Classifier: Label denoising for deep learning. PyTorch implementation of the deep abstaining classifier (DAC) from our ICML 2019 paper is on github
  • SimX: SimX is a library for developing parallel, distributed-memory simulations in Python. SimX is written primarily in C++ and provides the simulation modeler with the core functionality needed in a parallel simulation library such as event queueing, time advancement, domain partitioning, synchronization and message passing. SimX APIs are exposed to Python, enabling rapid development and protyping of a parallel simulation entirely in Python. SimX is free software, available under the GNU LGPL license, and hosted on github.
  • PolyClip: Polyclip is a library for fast clipping (intersection) of 2-D polygons, written in C++. It supports arbitrarily shaped polygons, including multi-part, self-intersecting and holed polygons. Degenerate cases such as touching and overlapping polygons are also handled. Polyclip is free software, available under the GNU LGPL license. Visit the Polyclip SourceForge web page here.