Vijayaraghavan, Aravindan

代表成果

• Computing High-Dimensional Confidence Sets for Arbitrary Distributions, COLT 2025 (With Chao Gao, Liren Shan, Vaidehi Srinivas)
• Agnostic Learning of Arbitrary ReLU Activation under Gaussian Marginals, COLT 2025 (With Anxin Guo)
• Volume Optimality in Conformal Prediction with Structured Prediction Sets, ICML 2025 (With Chao Gao, Liren Shan, Vaidehi Srinivas)
• Theoretical Analysis of Weak-to-Strong Generalization NeurIPS 2024. (With Hunter Lang, David Sontag).
• Efficient Certificates of Anti-Concentration Beyond Gaussians FOCS 2024. (With Ainesh Bakshi, Pravesh Kothari, Goutham Rajendran and Madhur Tulsiani).
• New tools for smoothed analysis: least singular values for random matrices with dependent entries STOC 2024. (With Aditya Bhaskara, Eric Evert and Vaidehi Srinivas).
• Error tolerant e-Discovery protocols CSLAW 2024. (With Jinshuo Dong, Jason Hartline and Liren Shan).
• Higher-order Cheeger Inequality for Partitioning with Buffers SODA 2024. (With Konstantin Makarychev, Yury Makarychev and Liren Shan).
• Computing linear sections of varieties: quantum entanglement, tensor decompositions and beyond FOCS 2023. (With Nathaniel Johnston and Benjamin Lovitz).
• Training subset selection for Weak Supervision NeurIPS 2022. (Hunter Lang, Aravindan Vijayaraghavan and David Sontag.)