Publications
* Equal contribution
Robustness Certificates for Neural Networks against Data Poisoning and Evasion Attacks [arxiv]
Sara Taheri, Mahalakshmi Sabanayagam, Debarghya Ghoshdastidar, Majid Zamani
IEEE Open Journal of Control Systems, Special Section: Intersection of Machine Learning with Control, 2026Exact Certification of Neural Networks and Partition Aggregation Ensembles against Label Poisoning [arxiv] [poster]
Ajinkya Mohgaonkar, Lukas Gosch, Mahalakshmi Sabanayagam, Debarghya Ghoshdastidar, Stephan Günnemann
The Trustworthy AI Workshop, ICLR 2026 Spotlight (top 6.4%)Exact Certification of (Graph) Neural Networks Against Label Poisoning [arxiv] [poster] [code]
Mahalakshmi Sabanayagam*, Lukas Gosch*, Stephan Günnemann, Debarghya Ghoshdastidar
International Conference on Learning Representations (ICLR) 2025 Spotlight (top 5.1%)
The VerifAI Workshop, ICLR 2025 OralProvable Robustness of (Graph) Neural Networks Against Data Poisoning and Backdoors [arxiv] [video] [code]
Lukas Gosch*, Mahalakshmi Sabanayagam*, Debarghya Ghoshdastidar, Stephan Günnemann
Transactions on Machine Learning Research (TMLR) 2025
New Frontiers of Adversarial Machine Learning Workshop, NeurIPS (AdvML-Frontiers NeurIPS) 2024 Oral and Best Paper AwardKernels, Data & Physics [arxiv]
Francesco Cagnetta, Deborah Oliveira, Mahalakshmi Sabanayagam, Nikolaos Tsilivis, Julia Kempe
Journal of Statistical Mechanics: Theory and Experiment (JSTAT Lecture Notes) 2024Robust Features Inference: A Test-time Defense using Spectral Projections [arxiv] [poster] [code]
Anurag Singh*, Mahalakshmi Sabanayagam*, Krikamol Muandet, Debarghya Ghoshdastidar
Transactions on Machine Learning Research (TMLR) 2024Unveiling the Hessian’s Connection to the Decision Boundary [arxiv] [poster] [code]
Mahalakshmi Sabanayagam, Freya Behrens, Urte Adomaityte, Anna Dawid
Mathematics of Modern Machine Learning Workshop, NeurIPS (M3L NeurIPS) 2023Analysis of Convolutions, Non-linearity and Depth in Graph Neural Networks using Neural Tangent Kernel [arxiv] [poster] [code]
Mahalakshmi Sabanayagam, Pascal Esser, Debarghya Ghoshdastidar
Transactions on Machine Learning Research (TMLR) 2023
Learning on Graphs (LoG) Meetup at Munich 2023Improved Representation Learning Through Tensorized Autoencoders [arxiv] [poster] [code]
Pascal Esser*, Satyaki Mukherjee*, Mahalakshmi Sabanayagam*, Debarghya Ghoshdastidar
International Conference on Artificial Intelligence and Statistics (AISTATS) 2023Analysis of Graph Convolutional Networks using Neural Tangent Kernels [arxiv] [code]
Mahalakshmi Sabanayagam, Pascal Esser, Debarghya Ghoshdastidar
MLG workshop at European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2022)Graphon based Clustering and Testing of Networks: Algorithms and Theory [arxiv] [poster] [code]
Mahalakshmi Sabanayagam, Leena Chennuru Vankadara, Debarghya Ghoshdastidar
International Conference on Learning Representations (ICLR) 2022Rough Set-based Feature Selection for Credit Risk Prediction using Weight Adjusted Boosting Ensemble Method [springer]
Sivasankar Elango, Selvi Chandran, Mahalakshmi Sabanayagam
Journal of Soft Computing 2019Cross Domain Sentiment Analysis Using Different Machine Learning Techniques [springer]
Mahalakshmi Sabanayagam, Sivasankar Elango
Fifth International Conference on Fuzzy and Neuro Computing (FANCCO) 2015
Poster in Grace Hopper Celebration India (GHCI) 2016
Preprints
Exact Generalisation Error Exposes Benchmarks Skew Graph Neural Networks Success (or Failure) [arxiv]
Nil Ayday, Mahalakshmi Sabanayagam, Debarghya GhoshdastidarGeneralization Certificates for Adversarially Robust Bayesian Linear Regression [arxiv]
Mahalakshmi Sabanayagam, Russell Tsuchida, Cheng Soon Ong, Debarghya GhoshdastidarCluster Specific Representation Learning [arxiv]
Mahalakshmi Sabanayagam, Omar Al-Dabooni, Pascal EsserMachine learning-based image detection for lensless microscopy in life science [link]
Mahalakshmi Sabanayagam, Jan Brunckhorst, Andreas Pirchner, Nikhitha Radhakrishna Naik