Skip to content

Author

Souvik Chakraborty

We have 3 of 19 papers

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

#artificial intelligence Preprint Oct 2026

Training-Loss Guarantees for Muon with Finite-Step Newton--Schulz Orthogonalization

Existing convergence analyses of Muon either assume exact orthogonalization or analyze classical Newton--Schulz polynomials, and guarantee only stationarity, so it is unresolved what Muon's five tuned Newton--Schulz steps preserve and whether that suffices to reach a prescribed neural-network training loss. We establis...

Amartya Roy, Souvik Chakraborty · 0 citations
#machine learning Preprint Aug 2026

Beyond Pairwise Graphs in Science: Hypergraph Adaptive Wavelet Operators for Parametric PDEs

The Hypergraph Adaptive waveLet Operator (HALO), which lifts the domain to a hypergraph and learns in its spectral wavelet domain, achieves best or near-best accuracy among frequency-, transformer-, DeepONet-, state-space-, and graph-based baselines and sustains stable multi-step rollouts.

R. Sarkar, Venkataramana Runkana, Souvik Chakraborty · 0 citations
#artificial intelligence Preprint Aug 2026

See, Hypothesize, Validate: Multimodal Agentic Framework for Discovering Governing PDEs

MAGE is introduced, an agentic framework that organizes PDE discovery as a confidence governed hypothesis validation loop inspired by the scientific cycle of observation, hypothesis, and falsification, and supports further study of structured agentic reasoning for library-free governing-law discovery.

Sarang Manoj Pekhale, Amartya Roy, R. Sarkar et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.